The prevalence of type 2 diabetes is continuously increasing, and the mean age of affected patients is decreasing. Although several antihyperglycaemic agents are available, there is an urgent need for novel agents with different mechanisms of action. Existing agents are often ineffective (even after initial effectiveness) and/or are associated with considerable side effects. Although antihyperglycaemic agents differ in their side effects, two kinds of adverse events are particularly common, perturbing and potentially harmful: weight gain and hypoglycaemia.
The prevalence of type 2 diabetes is continuously increasing, and the mean age of affected patients is decreasing. Although several antihyperglycaemic agents are available, there is an urgent need for novel agents with different mechanisms of action. Existing agents are often ineffective (even after initial effectiveness) and/or are associated with considerable side effects. Although antihyperglycaemic agents differ in their side effects, two kinds of adverse events are particularly common, perturbing and potentially harmful: weight gain and hypoglycaemia. One or the other, or both, are typical for all antihyperglycaemic agents, with the exception of metformin and acarbose. These latter agents, however, are typically used in early or less severe forms of type 2 diabetes, have a limited effectiveness and exert frequent gastrointestinal side effects. Therefore, drug development is directed towards effective antihyperglycaemic medications, without the risk of weight gain and hypoglycaemia. Incretin-based therapy is one such new, promising approach and incretin-enhancing agents and incretin mimetics are now available.
The observation that glucose taken orally results in a more pronounced insulin secretion than intravenously given glucose is the so-called incretin effect.1,2 Glucose-dependent insulinotropic polypeptide (GIP) and glucagon-like peptide 1 (GLP-1) are the most important mediators of this effect.1 Furthermore, in addition to stimulating insulin secretion in a glucose-dependent manner, GLP-1 exerts other physiological actions, including suppression of glucagon,3,4 reduction of food intake and bodyweight5 and delayed gastric emptying.6,7 Importantly, while it is well known that type 2 diabetes is characterised by a combination of insulin resistance and impaired insulin secretion, elevated fasting plasma glucagon and the failure to normally suppress glucagon following a meal also contribute to dysregulation of endogenous glucose production in patients with type 2 diabetes.8 Patients with type 2 diabetes usually lack the insulinotropic response to GIP, while GLP-1 levels may be reduced;9 however, the response is preserved. GLP-1 was therefore developed as a therapeutic target.10 Indeed, subcutaneously injected GLP-1 induces an increase in insulin secretion, improvement in insulin sensitivity and a reduction of glycated haemoglobin (HbA1c) and bodyweight.11 Moreover, the glucose dependence of the effect of GLP-1 on both insulin and glucagon secretion may diminish its potential to produce hypoglyacaemia: the insulinotropic properties of GLP-1 are abolished during euglycaemia and hypoglycaemia.3,12 However, due to its short half-life – a consequence of rapid degradation by depeptidyl peptidase 4 (DPP-4) – its clinical usefulness is limited. Therefore, longacting GLP-1 mimetics and DPP-4 inhibitors were developed (see Table 1). In this short review we evaluate incretin-based therapies with respect to the risk of hypoglycaemia.
How Does Glucagon-like Peptide-1 Result in Insulin Secretion?
In healthy individuals, pancreatic β cells respond to elevated glucose by an increase in insulin secretion. Glucose enters β cells through glucose transporters (GLUT-2), resulting in mitochondrial adenosine triphosphate (ATP) production, with subsequent closure of ATPsensitive K+ channels. This results in membrane depolarisation, opening calcium (Ca2+) channels, which then leads to exocytosis of the insulin-containing granules. Sulfonylureas directly close K+ channels and result in insulin secretion independent of the glucose concentration. Therefore, hypoglycaemia is a typical side effect of sulfonylurea therapy.
Insulin secretion stimulates GLP-1 in vitro with clear glucose dependency.13 Several studies in animal models and in human β cells have demonstrated that GLP-1 inhibits K+ channels in a glucosedependent manner.14 Binding of GLP-1 to its receptor on the β cell increases intracellular cyclic adenosine-3’ and 5’-monophosphate (cAMP)15 and thus acts synergistically with glucose to close K+ channels, which facilitates membrane depolarisation and induces electrical activity. Furthermore, in the presence of GLP-1, Ca2+ channels are active at lower action potentials and are more slowly inactivated, resulting in greater Ca2+ influx. It is also believed that Ca2+ mobilisation from intracellular stores promotes mitochondrial ATP synthesis and produces further membrane depolarisation, and thus enhances insulin secretion. Although the binding of GLP-1 to its receptor affects insulin secretion at several levels, all of these effects are glucose-dependent.16 As a result of this glucose dependency, it is rather unlikely to induce hypoglycaemia in normal subjects by exogenous GLP-1, regardless of dose. In this context, another observation of interest was that endocrine tumours producing GLP-1 did not show symptoms of hypoglycaemia.17 On the other hand, an exaggerated GLP-1 response to nutrients was observed in patients with accelerated gastric emptying and reactive hypoglycaemia, indicating that – at least in certain clinical situations – GLP-1 may be associated with hypoglycaemia.18 It should also be noted that not only the insulin secretion but also the glucagonostatic effects of GLP-1 are glucose-dependent.8 Thus, GLP-1 results in a much more physiological regulation of β cells and β-cell function, which minimises the risk of hypoglycaemia.
Rates of Hypoglycaemia in Clinical Studies Using Incretins to Lower Glucose Concentration
Native Glucagon-like Peptide 1
Since native GLP-1 is degraded rapidly by DPP-4, it must be given as a continuous infusion. A six-week continuous infusion of GLP-1 in type 2 diabetes subjects decreased HbA1c by 1.3%.11 No hypoglycaemias were reported. Similarly, in another study in which a continuous infusion of GLP-1 for 12 weeks was compared with standard treatment, no difference in diabetic control was noted, but a markedly reduced number of hypoglycaemic events was reported (0.5 events/patient-year during GLP-1 versus 43.5 events/patient-year during standard treatment).19
Glucagon-like Peptide-1 Mimetics and Analogues
Exanatide (Byetta®), the synthetic form of exendin-4, was developed as a GLP-1 mimetic. It has a 53% amino acid overlap with human GLP-1 and resembles a super-agonist at the GLP-1 β-cell receptor.20 In contrast to GLP-1, it has a much longer half-life, because it is resistant to degradation by DPP-4.21
Exenatide was used in several placebo-controlled studies as monotherapy in combination with metformin, with sulfonylurea or with both. Table 2 shows that this therapy resulted in a significant change of HbA1c associated with a wide range in the rate of hypoglycaemia. However, in studies where exenatide is used in combination with oral antihyperglycaemic medications that do not induce hypoglycaemia, such as metformin, the rates of hypoglycaemia are not increased.25 On the other hand, if exenatide is used together with potentially hypoglycaemiainducing agents, hypoglycaemia is observed in up to 36%.24 This indicates that exenatide itself has no, or only very little, potential for inducing hypoglycaemia. Rather, if glucose metabolism is improved by exenatide, other potentially hypoglycaemia-inducing agents can cause hypoglycaemia. It should also be noted that the rates of hypoglycaemia vary considerably between studies; this is mostly related to varying definitions of hypoglycaemia and different patient populations.
Other Glucagon-like Peptide-1 Analogues
Several other GLP-1 analogues are currently under development. Liraglutide has a long half-life (13 hours) and shows similar antidiabetic effects to exenatide. In monotherapy, no hypoglycemias were observed despite significant HbA1c reduction.9
Dipeptidyl Peptidase-4 Inhibitors
Another approach to using the effect of GLP-1 as an antidiabetic therapy is by inhibiting DPP-4, the enzyme responsible for the short half-life of GLP-1.24 DPP-4 can be found ubiquitously and can deactivate many hormones, neuropeptides, cytokines and chemokines. Despite this broad function in metabolism, it has very few side effects.
Although DPP-4 inhibitors also result in an elevated concentration of GLP-1, there are some clinically relevant differences compared with GLP- 1 analogues. DPP-4 inhibitors can be applied orally. Furthermore, during DPP-4 therapy GLP-1 concentration is not continuously elevated, but only following meals (when endogenous GLP-1 is secreted). On the other hand, the effect on glucose concentration and gastric motility seems to be less pronounced, since GLP-1 concentrations are much lower during DPP-4 inhibitor therapy compared with GLP analogues.
A number of DPP-4 inhibitors are on the market or in development.23,24 For sitagliptin (Januvia®) and vildagliptin (Galvus®) there are a number of clinical studies in which the rate of hypoglycaemia was also examined (see Tables 3 and 4).
Table 3 indicates that sitagliptin results in an average reduction of HbA1c of approximately 0.7% (with a baseline HbA1c of approximately 8%). The rates of hypoglycaemia vary between 0 and 16.4%, depending on the clinical setting and the concomitant therapy. It should also be noted that in the study with the lowest baseline HbA1c35 the rate of hypoglycaemia was highest. However, when glipicide was used to achieve an identical HbA1c reduction, the rate of hypoglycaemia was almost seven times higher (32 versus 4.9%). Furthermore, there seems to be no dose-dependency in the rate of hypoglycaemia. These data again indicate that sitagliptin itself at 100mg/day or 200mg/day does not induce hypoglycaemia. However, similar to GLP-1 analogues, this form of therapy can decrease glucose concentration to a point at which concomitant therapy (such as sulfonylureas or insulin) may induce hypoglycaemia.
Very similar data are available for vildagliptin (see Table 4). Data from these studies suggest that vildagliptin possesses comparable glucoselowering efficacy to established antidiabetic medications, including metformin, rosiglitazone and pioglitazone, while exhibiting advantages with respect to tolerability, lipid profile and weight regulation. With respect to the effect on HbA1c and the rates of hypoglycaemia, no clinically relevant difference compared with sitagliptin can be observed, although, notably, vildagliptin was generally investigated in patients with higher HbA1c baseline values. The rates of hypoglycaemia for the vildagliptin studies vary between 0 and 0.7%, depending on the clinical setting and the concomitant therapy.
Clinical data indicate that incretin-based therapy results in a very low rate of hypoglycaemia. Hypoglycaemia observed in studies using native GLP-1, GLP-1 analogues or DPP-4 inhibitors is usually related to concomitant therapy such as sulfonylureas or insulin. This makes this form of therapy a very valuable tool for diabetes management. This is particularly true because incretin-based therapy is also not associated with the other major complications of many glucose-lowering therapies (weight gain), since DPP-4 inhibitors are weight-neutral and GLP-1 analogues even result in weight reduction.