The Insulin:Glucagon Ratio and the Choice of Glucose-Lowering Drugs.

The influence of alpha and beta cells, through glucagon and insulin, on energy metabolism is well known. The insulin:glucagon ratio (IGR) is a frequently discussed entity in the medical literature. However, in recent years, focus has shifted to other pathways and markers of health and disease. This communication revisits the insulin:glucagon bipolar axis and describes the significance of the IGR. It reviews the effects of various glucose-lowering drugs on this ratio, and hypothesizes that the ratio can be used to predict the appropriate choice of drugs for managing diabetes. Drugs which increase the IGR may be beneficial in insulinopenic conditions, while those which decrease IGR may be of help in the setting of hyperinsulinemia or insulin resistance.

The Insulin:Glucagon Bipolar Axis

Glucagon and insulin are two opposing hormones which work in tandem to maintain a normal fuel balance. By modulating the relative concentrations of glucagon and insulin, the alpha and beta cells of the pancreas control endogenous glucose production, triacylglycerol deposition, and protein synthesis [1]. In simple terms, glucagon acts as a catabolic hormone, while insulin exerts anabolic effects on the body. A low insulin:glucagon ratio (IGR) stimulates mobilization of stored nutrients, increases glycogenolysis and gluconeogenesis, and promotes the breakdown of adipose tissue into free fatty acids and glycerol. A high IGR encourages biosynthesis of proteins, inhibits the production of glucose, and reduces free fatty acid release [2]. The effects of insulin and glucagon on various physiological processes are listed in Table 1.

Table 1

Biochemical effects of insulin and glucagon

Process	Insulin	Glucagon
Glycogenesis	Increased (glycogen synthase activation)	Inhibited (glycogen synthase deactivation)
Gluconeogenesis	Inhibited (inhibition of pyruvate carboxylase, phosphoenolpyruvate carboxykinase, glucose 6 phosphatase)	Increased (inhibition of pyruvate kinase, phosphofructokinase; activation of fructose 1,6-bisphosphatase)
Glycolysis	Increased (activation of glucokinase, phosphofructokinase, pyruvate kinase)
Glycogenolysis	Inhibited (inhibition of glycogen phosphorylase)	Increased (activation of glycogen phosphorylase)
Ketogenesis	Inhibited (less substrate acetyl coenzyme A (CoA); inhibition of 3-hydroxy-3-methylglutaryl-coenzyme A (HMG CoA) synthetase)
Lipogenesis	Increased (more substrate glycerol 3 phosphate and nicotinamide adenine dinucleotide phosphate (NADPH); activation of acetyl CoA carboxylase)
Protein synthesis	Decreased protein degradation	Increased amino acid uptake by liver; decreased amino acids in plasma

This aspect of physiology is not, however, highlighted adequately in diabetology circles and texts. Though Defronzo [3] included the alpha cell as part of the ominous octet, and Schwartz et al. [4] list it as one of the egregious eleven, less attention has been paid to this pathophysiologic player. Some reasons for this may be the multitude of factors which affect insulin and glucagon secretion, their highly variable concentrations, the existence of a significant, and variable, porto-systemic gradient, and the difficulty involved in estimating them. In spite of extensive research, there is still controversy surrounding the absolute concentrations of insulin and glucagon in persons with diabetes. Yet another explanation may be the current interest in nonpancreatic pathophysiological pathways of diabetes. This article does not contain any new studies with human or animal subjects performed by any of the authors.

Insulin

Ambient insulin concentrations present a spectrum of values, ranging from absolute deficiency in type 1 diabetes to hyperinsulinemia in those with type 2 diabetes. Beta-cell secretory capacity is known to fall with increasing duration of diabetes, so insulin levels may vary depending upon diabetes stage. Insulin levels are also reported to be high in obese persons, irrespective of glycemic status. This is a reflection of the insulin resistance that occurs with obesity. Insulin levels vary with meals, with specific nutrients, and with comorbid conditions. Insulinogenic stimuli, for example, include glucose as well as certain amino acids (Table 2). Alterations in growth hormone and thyroid hormone secretion also affect insulin secretion. In man, a first phase and a second phase of insulin can easily be identified, along with distinct basal and prandial secretion [5]. In some primates, insulin is released in oscillations which occur at intervals of roughly 9 min [6]. This phenomenon is not seen in humans.

Table 2

Physiological determinants of insulin and glucagon secretion

	Insulin	Glucagon
Stimulants	Glucose; amino acids—arginine, leucine; gastrointestinal hormones—secretin, gastrin, pancreozymin	Hypoglycemia, amino acids, low adrenaline
Inhibitors	Adrenaline	Hyperglycemia

Glucagon

Glucagon secretion varies from the fasting to the postprandial state according to the composition of the meal taken, and is influenced by body weight as well as comorbid illnesses [7, 8]. There are varying opinions about the concentration of glucagon in obesity and—just like diabetes—current research suggests that glucagon, or glucagon response to arginine, increases in obesity. Glucagon levels are higher in diabetes, increase on stimulation with arginine, and are not suppressed after glucose administration. Workers also suggest a resistance to glucagon in obese persons with diabetes [9].

The issue is also complicated by issues associated with the assessment of insulin and glucagon. Various researchers estimate a variety of species, such as C-peptide, insulin, pancreatic glucagon, enteroglucagon, and immunoreactive glucagon. Altered tissue response to insulin and glucagon is also reported in obesity, raising the question of whether the determination of absolute glucagon values is worth the effort. Yet another controversy relates to the systemic and portal concentrations of these antipodal hormones. As the portosystemic gradients of insulin and glucagon are dissimilar, the relevance of testing peripheral IGR is also doubtful [10, 11].

Insulin:Glucagon Ratio

The insulin:glucagon bipolar axis serves to regulate carbohydrate and lipid metabolism. The therapeutic value of glucagon in diabetes management is gradually becoming clear [12]. As these hormones work in tandem with each other, albeit in opposing directions, it may make sense to interpret the IGR instead of assessing absolute values. This concept was popularized by Unger [1], who proposed the bihormonal hypothesis. The IGR has been found to vary with the presence of, or the need for, anabolism and catabolism. The IGR varies inversely with the need for endogenous glucose production. It is lowest in total starvation and highest during loading with exogenous carbohydrate. The molecular mechanisms which control this energy homeostasis are listed in Table 1. In brief, the IGR acts as a physiological fulcrum, balancing two opposite ends of the metabolic spectrum to provide energy when needed and conserve it if possible.

Glucose-Lowering Drugs

Recent years have witnessed the development of various classes of glucose-lowering drugs. These molecules are studied for their effect on a wide range of glucotropic hormones and metabolites, apart from their direct effect on glucose lowering. The effect of various glucose-lowering drugs on insulin levels is well known, but their effect on glucagon levels, and specifically on the IGR, is rarely discussed. This is unfortunate, as detailed insight into this aspect of pharmacodynamics may aid the selection of the appropriate therapy for each patient based upon their insulin-glucagon homeostasis. Drugs which increase the IGR may be beneficial in insulinopenic conditions, while those which decrease the IGR may be of help in the setting of hyperinsulinemia or insulin resistance.

Recent studies have shown that metformin therapy does not lead to any change in glucagon secretion [12]. However, the insulin-sensitizing effect of metformin causes a reduction in hyperinsulinemia, and thus reduces the IGR. Similar effects are reported with pioglitazone [13].

Sulfonylureas such as glimepiride are insulinotropic agents which increase insulin secretion but have no impact whatsoever on alpha cell function or glucagon secretion. Thus, glimepiride certainly increases the IGR. In another study, glimepiride was found to increase both prandial insulin as well as prandial glucagon [area under curve (AUC), 0–2 h] [14]. The opposing effects of insulin sensitizers and insulin secretagogues are visible at the molecular level as well. While metformin activates carnitine palmitoyltransferase 1 (CPT1), glibenclamide inhibits this enzyme [15, 16]. This molecular difference may explain the effects of these drugs on weight: metformin promotes β-oxidation of free fatty acids, while glibenclamide facilities triacylglycerol production.

The incretin-based therapies are, in general, thought to reduce plasma glucagon levels. The evidence reveals variable effects. Recently, liraglutide, a glucagon-like peptide-1 receptor agonist (GLP1RA), has been shown to increase post-challenge glucagon levels after 12 weeks of therapy, though no such effects were noted after shorter periods of treatment. This effect persisted for 48 weeks (the duration of the trial), but was not noted in the fasting state [17]. Tachyphylaxis has been suggested as a possible mechanism to explain this phenomenon [18]. Other studies have demonstrated no effect of liraglutide on fasting glucagon, but have noted its ability to reduce 24-h glucagon levels (by reducing glucagon secretion after a protein-rich meal) [19]. The authors reported no change in the IGR with liraglutide therapy. However, in a Japanese study, liraglutide increased the IGR. While the drug reduced the incremental release of glucagon between 15 and 60 min after a meal, it did not change overall glucagon secretion. Coupled with an increase in insulin secretion, this led to an increase in the IGR with liraglutide [20].

A 28-day-long pharmacodynamics study compared the effect of lixisenatide and liraglutide on plasma insulin and glucagon concentrations, but did not statistically analyze the IGR [21]. Lixisenatide provided a significantly greater decrease in postprandial glucagon as compared to liraglutide. Postprandial insulin and C-peptide levels were also significantly reduced with lixisenatide versus liraglutide [21], while decreases in proinsulin were comparable between groups. The trend was for a higher IGR with liraglutide and a lower one with lixisenatide. In individuals without diabetes, exenatide increased the IGR at rest, but this trend reversed when exercise was performed [22]. A lowering of plasma glucagon was noted in individuals with type 2 diabetes with the same drug [23]. Dulaglutide, a once-weekly GLP1RA reduces fasting glucagon concentrations while increasing homeostatic model assessment 2 beta (HOMA2B). This implies that dulaglutide administration increases the IGR [12].

Relative to the dipeptidyl peptidase inhibitor sitagliptin, exenatide has been shown to increase the IGR [24]. However, in absolute terms, sitagliptin also increases insulin secretion and reduces glucagon concentrations, thus causing a rise in the IGR. Vildagliptin, another dipeptidyl peptidase inhibitor, decreases prandial glucagon AUC (0–2 h). This effect has been reported to persist for at least 2 years, and reflects the improved alpha cell sensitivity caused by vildagliptin. Vildagliptin also increases insulin secretion, though not to the same extent as glimepiride. In total, vildagliptin increases the IGR [13]. Similar effects are noted with linagliptin and saxagliptin [25, 26].

Empagliflozin, a sodium glucose co-transporter-2 (SGLT2) inhibitor, has recently been shown to reduce the IGR. This effect is pronounced after acute administration of the drug, but persists—to an attenuated degree—after 28-day exposure. This reduction in the IGR reflects an increased need for—and an actual increase in—endogenous glucose production, to balance the increase in glucosuria. Similar results have been reported with dapagliflozin (Table 3) [27, 28]. The possible reasons for this phenomenon have been hypothesized in detail [29].

Table 3

Effects of antidiabetic drugs on the insulin:glucagon bipolar axis

Antidiabetic drugs	Insulin	Glucagon	IGR
Glimepiride	↑	No change	↑
DPP4i	↑	↓	↑
Liraglutide	↑	↓/No change	No change/↑
Dulaglutide	↑	↓	↑
Exenatide	↑	↓	↑
Lixisenatide	↓	↓	↓
Dapagliflozin	↑	↑	↓
Metformin	↓	No change	↓
Pioglitazone	↓	–	↓
Acarbose	↓	No change	↓

Based upon [11–14, 16, 18–22, 24–27]

↓ = reduction; ↑ = increase

DPP4i dipeptidyl peptidase-4 inhibitor, IGR insulin:glucagon ratio

No significant differences in serum insulin and plasma glucagon levels were documented after 1 week of colestimide administration [30]. Acarbose, an alpha-glucosidase inhibitor, reduces serum insulin levels without influencing glucagon secretion, and thus reduces the IGR [31].

Impact on Weight Gain and Hypoglycemia

Weight gain and hypoglycemia are common but unwanted corollaries of conventional glucose-lowering therapy. However, newer classes of drugs offer the advantage of weight loss along with glycemic control while also avoiding hypoglycemia. In general, drugs which increase the IGR will be associated with an anabolic effect or tendency for weight gain and an increased risk of hypoglycemia. On the other hand, drugs which reduce the IGR will facilitate weight loss and prevent hypoglycemia. Our IGR-based construct is concordant with the clinical pharmacology of directly acting insulin secretagogues such as sulfonylureas (which increase the IGR, may cause weight gain, and are associated with hypoglycemia) and novel drugs like GLP1RA and SGLT2 inhibitors (which reduce the IGR, aid weight loss, and are associated with a minimal risk of hypoglycemia).

Therapeutic Implications

In spite of the increasing number of molecules available for diabetes management, glycemic control remains suboptimal. A large number of people with diabetes continue to be classified as “refractory” or unresponsive to treatment. Even in the best randomized controlled trials, a significant proportion of subjects remain nonresponders to allocated therapy. There is also a dearth of studies that are able to differentiate responders and nonresponders based upon their demographic characteristics, physical attributes, or biochemical phenotype. Thus, there is still an unmet need to explore and identify biochemical markers which may serve as predictors of response to chosen therapeutic interventions.

The IGR, therefore, may be used to create a matrix for choosing an antidiabetic therapy. Though data are limited and may be conflicting, one can create a model, based upon absolute and relative values of insulin and glucagon, to facilitate the selection of the appropriate therapy (Fig. 1). Persons with high insulin levels(hyperinsulinemia, secondary to insulin resistance) will benefit from drugs which lower insulin secretion (insulin sensitizers, SGLT2 inhibitors, alpha-glucosidase inhibitors). Within this group of patients, those with low glucagon values may respond better to SGLT2 inhibitors. Incretin-based therapy has been shown to be more effective in persons with high glucagon levels [32]. As these modes of treatment cause hyperinsulinemia, incretin-based therapy may prove more suitable for persons with a normal IGR. Glimepiride may be a better choice for persons with low insulin and a low IGR. The effects of metformin and SGLT2 inhibitors on the IGR are similar to those seen with calorie restriction. Thus, clinical situations requiring calorie restriction mimicry (e.g., obese diabetes; need for geroprotection) [33] may benefit from these drug classes.

Fig. 1

Choice of glucose-lowering therapy based upon the IGR. X-axis: IGR; y-axis: insulin concentration. AGI alpha-glucosidase inhibitors, DPP4i dipeptidyl peptidase-4 inhibitor, GLP1RA glucagon-like peptide-1 receptor agonist, IGR insulin:glucagon ratio, SGLT2i sodium glucose transporter-2 inhibitor

Rational combinations may also be created using knowledge of the effects of drugs on insulin and glucagon. For example, it may be beneficial to combine incretin-based therapy, which reduces glucagon concentrations, with SGLT2 inhibitors, which are known to increase glucagon.

It is not our suggestion that IGR estimation becomes an integral part of routine diabetes care; it is, and will remain, a research tool. However, clinical endocrinologists should understand this physiological concept and be able to translate it into an appropriate choice of pharmacotherapy. Preclinical and clinical development programs for modern drugs should include detailed study of their effects on the insulin:glucagon axis: this will help to unravel new facets of human biochemistry and physiology.