New treatments in type 2 diabetes: a focus on the incretin-based therapies


A. H. Barnett, Undergraduate Centre, Birmingham Heartlands Hospital, Bordesley Green East, Birmingham, B9 5SS, UK. Tel.: +44 121 4243587; Fax: +44 121 4240593; E-mail:


The demonstration that the incretin hormone glucagon-like peptide 1 can improve glycaemic control in patients with type 2 diabetes has led to the rapid development during the last decade of promising new classes of agent for the management of type 2 diabetes. These agents possess a range of physiological effects that are associated with improved glycaemic control in diabetes including stimulation of glucose-dependent insulin secretion, suppression of glucagon secretion, slowing of gastric emptying, and reduction of food intake. In addition, preclinical studies suggest that incretin-based therapies may improve β-cell function via enhancement of β-cell mass and induction of genes important for differentiated β-cell function. Exenatide, and the dipeptidyl peptidase-4 inhibitors, sitagliptin and vildagliptin are already approved, and liraglutide is currently completing Phase 3 trials. As these agents and standard oral therapies for type 2 diabetes lower glucose levels through different, but potentially complementary mechanisms, their use in combination should provide effective, potentially additive, glycaemic control. The incretin-based therapies also offer other advantages such as weight loss with exenatide and liraglutide, a reduced risk of hypoglycaemia, and as suggested by preclinical studies, a potential β-cell preserving effect. Long-term outcome and safety data are not available for these agents, but they appear generally well-tolerated in comparison with existing therapies for type 2 diabetes. The multiple underlying glucose-lowering actions of the incretin-based therapies, as well as a lack of weight gain or even weight loss, make these important new additions to available antidiabetic agents expanding the treatment options available for patients.


Diabetes mellitus is a major contributor to the global disease burden and is currently experiencing a dramatic rise in prevalence. In 1985 an estimated 30 million people worldwide had diabetes, and in 2000, the figure had risen to over 150 million. This figure is expected to rise to almost 350 million by 2025.1 Type 2 diabetes will account for most of the projected increase, which reflects not only the demographics of an aging population, but also increasing numbers of overweight and obese people.2 Randomized, controlled trials have provided compelling evidence that achieving strict glycaemic control can reduce the long-term complications of diabetes,3–7 highlighting the importance of increasing the proportion of patients achieving current glycaemic targets. Despite this knowledge, achievement of glycaemic control in clinical practice is far from satisfactory, many patients fail to achieve long-term glycaemic control even when receiving optimal treatment with current antidiabetic medications. Several reasons may account for this failure, including difficulty in implementing and maintaining adequate lifestyle modifications, poor compliance with therapy, limited efficacy and/or side-effects of current antidiabetic therapy, and the underlying disease pathophysiology.

While strategies that favour prevention of diabetes and adherence to healthy behaviours among people with existing diabetes are essential to address the current epidemic, there is also a need for the continued development of new treatments in an attempt to gain greater efficacy and an improved tolerability profile. The pathophysiology of diabetes is an important consideration in the development of any new treatments. Insulin resistance and β-cell dysfunction both play important roles in the development of type 2 diabetes8,9 and both defects remain closely linked with the progression of the disease: declining β-cell function associated with deteriorating glycaemic control6,10 and insulin resistance associated with numerous risk factors for cardiovascular disease.11–13 Effective glycaemic control will therefore require therapies that address both underlying defects. In the healthy pancreas, the coordinated actions of insulin and glucagon maintain glucose homeostasis, preventing both hyper- and hypoglycaemia. However, it is now recognized that glucose homeostasis is governed by the interplay between a number of hormones including insulin, glucagon, and the incretin hormones. This enhanced understanding of glucose homeostasis is proving central to the design of new pharmacological agents to promote better clinical outcomes and quality of life for people with diabetes. This review will focus on the endocrine effects of new diabetes treatments, highlighting their strengths and weaknesses and discussing their likely place in therapy.

Why do we need new drugs for the treatment of type 2 diabetes?

Diabetes is a chronic disease that causes substantial premature morbidity and mortality. Both the degree and duration of exposure to elevated plasma glucose levels are important determinants of the risk of developing complications.3 Certain ethnic minorities, for example, individuals originating from the Indian subcontinent, are particularly prone to developing type 2 diabetes and its complications at a younger age and with a shorter duration of disease.14 Achieving glycaemic control as close to normal as possible is essential to reduce the risk of complications. While available antidiabetic agents may initially be effective at achieving recommended levels of glycaemic control, type 2 diabetes is characterized by a progressive decline in β-cell function and prolonged efficacy is difficult to achieve requiring regular adjustment and intensification of treatment. One reason for this is the differing ability of the different drugs to target the underlying causes of the disease, rather than just treating its symptoms. Agents such as metformin and the thiazolidinediones (TZDs) attempt to sensitize specific tissues to the effects of insulin, whereas the main role of the insulin secretagogues (sulphonylureas and meglitinides) is to enhance the insulin producing activity of the pancreas. However, the true underlying cause of disease progression, the declining function of pancreatic β-cells, is not sufficiently targeted.

There are also difficulties in balancing the reduction in hyperglycaemia with the risk of hypoglycaemia. The increased risk of hypoglycaemia and the weight gain associated with several therapies both represent major barriers to optimal glycaemic control. Given that a large proportion of individuals presenting with type 2 diabetes are also overweight, therapies that do not promote additional weight gain are important. Metformin is usually the treatment of choice in these patients, but tolerability issues and certain contraindications may preclude its use in some individuals. New therapies that do not increase the already significant barriers to patient compliance in a chronic disease such as diabetes may improve the proportion of patients achieving current targets.

While both fasting and postprandial glucose contribute to HbA1c, postprandial glucose contributes the most to overall glycaemic exposure in individuals with HbA1c < 8·5%.15 Therefore, to achieve an HbA1c target of < 7·0%, control of postprandial glucose is key. Furthermore, postprandial glucose excursions have been linked to increased risk of cardiovascular disease,16 both in people with diabetes and in those without diabetes (although there is presently no evidence that reducing postprandial glucose will reduce cardiovascular risk per se). While many existing diabetes medications lower both postprandial glucose and fasting glucose levels, new agents with a greater postprandial glucose lowering capacity may provide additional benefit.

Normal glucose regulation

The glucoregulatory hormones of the body are designed to maintain circulating glucose concentrations in a relatively narrow range. They include insulin, glucagon, amylin, glucagon-like peptide 1 (GLP-1), glucose-dependent insulinotropic peptide (previously termed gastric inhibitory polypeptide, GIP), adrenaline, cortisol and GH. Of these, insulin and amylin are derived from the β-cells, glucagon from the α-cells of the pancreas, and GLP-1 and GIP from the L-cells and K-cells of the intestine, respectively. Insulin is secreted in response to increased blood glucose following ingestion of a meal and is carefully regulated, being secreted in increasing amounts if glucose concentrations increase beyond a threshold of approximately 3·3 mmol/l. In the postprandial state, the secretion of insulin occurs in two phases: an initial rapid release of preformed insulin, followed by increased insulin synthesis and release in response to blood glucose. Long-term release of insulin occurs if glucose concentrations remain high. Insulin exerts its actions through binding to insulin receptors present on many cells of the body, including fat, liver and muscle cells. The main ways by which insulin controls postprandial glucose are by initiating glucose uptake in insulin-sensitive peripheral tissues and promoting glycogenesis in the liver. In addition, insulin inhibits glucagon secretion in a negative feedback loop thus signalling the liver to stop producing glucose via glycogenolysis and gluconeogenesis.

The physiology and biology of the incretin hormones

It was recognized that the incretin hormones play a role in helping regulate glucose appearance and in enhancing insulin secretion when it was demonstrated that ingested food caused a more potent release of insulin than glucose infused intravenously.17 This effect, termed the ‘incretin effect’, suggested that signals from the gut are important in the hormonal regulation of glucose disappearance.

The incretin hormones comprise GIP and GLP-1. Both molecules are rapidly released by cells lining the gastrointestinal tract after a meal. GIP is secreted in a single bioactive form by K-cells and released from the upper small intestine. GLP-1 is secreted from L-cells of the distal intestine in two bioactive forms, GLP-1(7–37) and the predominant circulating active form GLP-1(7–36).18 Both GLP-1 isoforms are equipotent, with the same plasma half-life and identical activity through the same receptor.

GIP stimulates insulin secretion and regulates fat metabolism, but unlike GLP-1 does not appear to have significant effects on glucagon secretion, gastric emptying or body weight.19 GIP levels are normal or slightly elevated in people with type 2 diabetes.20 While GIP is a more potent incretin hormone, GLP-1 is secreted in greater concentrations and is more physiologically relevant in humans.21 Although predominantly expressed in mucosal L-cells, GLP-1 is also expressed in pancreatic α- and β-cells, lung, kidneys, heart as well as several brain areas (hypothalamus, pituitary).22

GLP-1 exerts its effect on postprandial glucose concentrations through several mechanisms (Fig. 1), including enhancing insulin secretion and suppressing postprandial glucagon secretion in a glucose-dependent manner.23,24 Circulating levels of GLP-1 are very low in the fasting state, and rapidly increase following food ingestion. Although this increase is reduced in patients with type 2 diabetes or impaired glucose tolerance, the GLP-1 insulin and glucagon response remains intact if exogenous GLP-1 is administered. In the pancreas, GLP-1 binds to a specific receptor on the β-cell that is coupled to glucose entry and metabolism.25,26 The effect of GLP-1 on insulin secretion is strictly glucose-dependent and there is no effect of GLP-1 on insulin secretion for glucose concentrations below a certain threshold (approximately 4·5 mmol/l).27,28 In addition to its effects on insulin secretion, GLP-1 has been shown to stimulate insulin gene transcription and all steps of insulin biosynthesis, thereby providing continual supplies of insulin for secretion.29

Figure 1.

Anti-diabetes effects of GLP-1.108 *The effect on insulin and glucagon secretion is glucose dependent. †The effect on β-cell mass has only been demonstrated in preclinical studies.

GLP-1 is also associated with a number of other glucoregulatory actions besides enhancement of glucose-dependent insulin secretion. The inhibitory effects of GLP-1 on glucagon secretion are an important mechanism for regulating elevated levels of blood glucose. The exact mechanism remains to be determined, but is probably indirectly mediated via insulin release and via somatostatin secretion. A direct effect of GLP-1 is not completely excluded, however, as GLP-1 receptors are also expressed on pancreatic α-cells.30

Gastric emptying also plays a physiological role in postprandial glucose regulation as delivery of nutrients from the stomach to the small intestine is an important contributor to postprandial glucose excursions.31 In nondiabetic subjects, gastric emptying is normally restrained as plasma glucose levels increase. In patients with type 2 diabetes, the failure to synchronize gastric emptying and insulin release contributes to postprandial hyperglycaemia. GLP-1 exerts inhibitory effects on gastrointestinal secretion and motility, particularly on gastric emptying thereby reducing meal-associated increases in glycaemic excursions.27,30 Administration of GLP-1 at physiological doses in healthy volunteers results in a dose-dependent slowing of gastric emptying and of glucose absorption, with a subsequent reduction in postprandial plasma glucose concentrations.27

GLP-1 has been shown to reduce caloric intake and to enhance satiety. In normal subjects, intravenous administration of GLP-1 above physiological levels induced increased feelings of satiety as well as a reduction of food intake.32 Similar effects have been observed in obese subjects and in patients with type 2 diabetes.33,34 In patients with type 2 diabetes treated with a subcutaneous infusion of GLP-1 for up to 6 weeks, the reduction of food intake was sustained and associated with a reduction of body weight.23 The exact mechanism by which GLP-1 mediates satiety effects is not known. It may bind to receptors in brain centres controlling food intake and energy expenditure thus suppressing food intake, or act on vagal afferent fibres to modulate GLP-1 neuronal transmission in the central nervous system.35 It is also likely that gastric emptying mediated by GLP-1 increases the sensation of fullness and leads to the termination of meal ingestion, thereby participating in the regulation of food intake.

Animal data have shown that GLP-1 has a trophic effect on β-cells, preserving or enhancing β-cell function as a result of β-cell proliferation and neogenesis and inhibition of apoptosis.28 Whether GLP-1 also expresses pancreatic trophic properties in vivo in human remains to be confirmed. However, these findings are of considerable interest as GLP-1 and GLP-1 analogues could be potentially useful in preserving functional β-cell mass in patients with type 2 diabetes.

GLP-1 has a very short half-life of about 2 min following exogenous administration as it is rapidly cleaved and inactivated by the protease dipeptidyl peptidase-4 (DPP-4). Therefore, while GLP-1 itself cannot be used as a practical therapy, a number of therapeutic interventions have been developed that exploit the GLP-1 pathway.

Incretin-based therapies

The incretin-based therapies are a new class of antidiabetic drug with a mechanism of action distinct from any existing class of oral glucose-lowering agent. They enhance the body's own ability to control blood glucose by increasing active levels of GLP-1. There are two current approaches to enhancing endogenous GLP-1 action in vivo. The first approach involves incretin mimetics, which are GLP-1 analogues that mimic the effect of GLP-1 but are resistant to degradation by DPP-4. Compounds in this class include exenatide and liraglutide. The second approach is to produce substances that increase the half life of endogenous GLP-1. Agents in this class include the DPP-4 inhibitors, sitagliptin and vildagliptin, which potentiate the incretin hormones by inhibiting the enzyme responsible for their degradation.


Exenatide was the first incretin mimetic to reach the market. It is a partial structural analogue of human GLP-1, sharing 53% amino acid sequence identity, but otherwise displaying effects similar to those of endogenous GLP-1 in terms of glycaemic control.25 In contrast to GLP-1, which contains an alanine at position 2, exenatide has a position 2 glycine, it is therefore not a substrate for DPP-4 and has a much longer plasma half-life than GLP-1.36 Following a single subcutaneous injection, the peak plasma exenatide concentration is reached within 2 h, and it can be measured in plasma for up to 10 h.37,38 Exenatide improves glycaemic control primarily by reducing postprandial hyperglycaemia, with a modest reduction in fasting plasma glucose levels.39

A prominent feature of type 2 diabetes is a dramatic reduction in first-phase insulin secretion, the insulin normally secreted by pancreatic β-cells within 10 min after a sudden rise in plasma glucose concentrations. Exenatide has been shown to restore both first phase and second phase insulin secretion in response to a glucose bolus in subjects with type 2 diabetes.40 The pattern of insulin secretion observed with exenatide was comparable with the responses in healthy control subjects given saline (Fig. 2).

Figure 2.

Plasma insulin secretion rates in 13 subjects with type 2 diabetes given an intravenous infusion of exenatide or saline, and in 12 healthy volunteers given an intravenous infusion of saline. Repeated-measures anova indicated significant differences in insulin secretion as measured by area-under-the-curve (AUC) analysis of the first (0–10 min) and second (10–120 min) phase insulin release between subjects with type 2 diabetes treated with exenatide and saline. Exenatide-treated subjects with type 2 diabetes also had significantly increased second phase insulin release compared to healthy volunteers. Reproduced with permission from Fehse et al.40

Given the known action of glucagon to maintain hepatic glucose output and the well-documented elevated fasting and postprandial glucagon concentrations in patients with type 2 diabetes,41 it was expected that glucagon suppression by exenatide would contribute to the overall effect of lowering glucose concentrations in both the fasting and postprandial periods and this has been demonstrated in Phase 2 studies.42,43 Similar to insulin concentrations, glucagon concentrations also returned toward baseline beyond 3 h postinjection during the fasting state, coincident with reaching nadir glucose concentrations. This suggests that suppression of glucagon secretion is not just a result of a slowing of nutrient presentation to the small intestine (gastric emptying).

The effect of exenatide on gastric emptying has been studied in several Phase 2 studies using the rate of appearance of acetaminophen (paracetamol) as a prototype for oral drugs that are absorbed in the small intestine, but not the stomach to indicate the rate of gastric emptying. These studies have demonstrated a consistent, dose-dependent slowing of gastric emptying by subcutaneous exenatide administration.42–44 The inhibitory effects of incretin-based therapies such as exenatide on gastric emptying are of particular interest because they have the potential to reduce postprandial glucose excursions. Like GLP-1, exenatide also influences postprandial satiety through neurohormonal networks that signal the brain to suppress appetite and food intake thereby reducing the rate at which meal-derived glucose reaches the circulation.44

Preclinical studies provide strong evidence that exenatide plays an important role in the maintenance of β-cell mass and function by increasing the expression of key β-cell genes and stimulating islet cell proliferation and neogenesis, and inhibition of islet cell apoptosis.45–49 In clinical studies, exenatide, has shown improvement in parameters associated with β-cell function as measured by several markers, including HOMA-B, pro-insulin : insulin ratios, and the restoration of first-phase insulin secretion.40,50,51

The efficacy and safety of exenatide have been assessed in placebo-controlled trials and open-label comparator studies. All trials enrolled patients with HbA1c levels of 7·5–11% and BMI > 25 kg/m2. In placebo-controlled trials, exenatide resulted in improved glycaemic control which was dose related. The highest 10 µg exenatide twice-daily dose reduced HbA1c from baseline by 0·8–1% at 30 weeks in patients inadequately controlled on metformin and/or sulphonylureas (patients continued to use their existing oral glucose-lowering medication throughout the trial).51–53 Across these three studies, approximately 40% of patients on the 10 µg exenatide dose achieved HbA1c measurements of ≤ 7%. There was also a significant reduction in mean body weight of approximately 2 kg.

However, whereas exenatide has a profound effect on postprandial glucose levels, its action on fasting blood sugar levels is modest.51–53 In the 30-week trials, patients receiving exenatide 10 µg twice daily had a 0·6 mmol/l (10 mg/dl) decline in fasting glucose compared to baseline and a 0·9–1·4 mmol/l (16–25 mg/dl) decline in fasting glucose levels compared to placebo-treated patients.51–53 The smaller effect on fasting blood glucose might be related to its relatively short duration of action. Studies with a once-weekly preparation, exenatide LAR, showed a 2·8 mmol/l (50 mg/dl) fall in fasting blood glucose and a 2% drop in HbA1c after 15 weeks of therapy.54

Sustained improvements in glycaemic control with exenatide have now been demonstrated for up to 3 years with progressive weight reduction from baseline.55 In an open-label, randomized, two-period, crossover noninferiority trial in patients with type 2 diabetes treated with a single oral antidiabetic agent, exenatide 10 µg twice daily was compared with insulin glargine once daily (titrated to target a fasting serum glucose level ≤ 5·6 mmol/l).56 Both exenatide and titrated insulin glargine were associated with similar significant changes from baseline in HbA1c (both, –1·36%; P < 0·001) (Fig. 3); the difference between groups was not statistically significant. However, there was a significant difference in body weight in favour of exenatide (Fig. 3).

Figure 3.

Efficacy of exenatide and titrated insulin glargine in adult patients with type 2 diabetes previously uncontrolled with metformin or a sulphonylruea. The study included two 16-week treatment periods. Patients were randomized to receive exenatide followed by insulin glargine or the reverse sequence for 16 weeks each. A, Mean (SE) HbA1c over time, and B, Mean (SE) change in body weight over time in the intent-to-treat population (n = 138). Reproduced with permission from Barnett et al., 2007.56

The most common adverse effects of exenatide are nausea and vomiting. Studies have shown that the incidence is significantly reduced by progressive dose escalation of exenatide.57 A lengthy dose-escalation period may not, however, be practical for patients outside of a study environment. There was no evidence from the Phase 3 studies for a correlation between the presence of nausea and weight loss achieved.

As stimulation of insulin secretion occurs only in the presence of elevated blood–glucose concentrations, the risk of hypoglycaemia should be greatly reduced with exenatide. Mild to moderate hypoglycaemia has been reported in Phase 3 studies with exenatide.51,53 However, it appears that it is only when in combination with a sulphonylurea that the risk of hypoglycaemia is increased.51,53

In the UK, exenatide is indicated for the treatment of type 2 diabetes mellitus in combination with metformin and/or sulphonylureas in patients who have not achieved adequate glycaemic control on maximally tolerated doses of these oral therapies.58 It is usually injected twice a day within 60 min before the morning and evening meals although a long-acting release form is in development with the objective of providing once weekly subcutaneous administration of exenatide.54 Exenatide should not be injected after meals. Currently, the UK licence states that limited experience exists concerning the combination of exenatide with TZDs58 and it is not currently licensed for use with insulin.


Another strategy to produce a longer acting form of GLP-1 that is resistant to DPP-4 degradation is to modify the GLP-1 molecule to promote albumin binding. Liraglutide is the most studied agent in this group and is currently in Phase 3 trials. When injected subcutaneously, the compound is slowly released from the injection site. As liraglutide is extensively bound to albumin it is protected from degradation by DPP-4 while at the same time reducing renal clearance. These characteristics combine to give the compound a plasma half-life of 10–12 h in humans,59–61 endowing it with pharmacokinetic properties that may be suitable for once daily administration.62 Liraglutide has 97% homology to GLP-1 and the in vitro potency of this agent is equal to that of native GLP-1.62 However, as liraglutide binds strongly to albumin, the plasma concentration of liraglutide cannot be directly compared with GLP-1 concentrations in studies involving administration of the native peptide. Preclinical studies have demonstrated that liraglutide decreases plasma glucose levels, increases insulin secretion, reduces glucagon secretion, and inhibits gastric emptying and appetite.63,64 The onset of action of liraglutide is expected to be relatively slow, however, because of its slowly increasing concentration. As the concentration increases and glucose slowly decreases, the stimulation of insulin secretion falls as a result of the glucose-dependent nature of liraglutide's action.

In a substudy of a larger Phase 2 trial, subjects underwent standard tests to assess first-phase insulin secretion and maximal β-cell insulin secretory capacity at baseline and 14 weeks.65 The larger study was a double-blind, placebo-controlled, randomized trial conducted over 14 weeks and included 165 patients with type 2 diabetes who were previously treated with diet or a single oral antidiabetic agent.66 After an initial 4-week washout period, patients were randomized to one of three once-daily doses of liraglutide (0·65, 1·25 and 1·9 mg) or placebo. Of 39 participants who began the substudy, 28 completed the 14 weeks of treatment. The two higher doses of liraglutide (1·25 and 1·9 mg) significantly increased maximal β-cell insulin secretory capacity compared to placebo by 114% and 97%, respectively (P < 0·05 for both doses), and first-phase insulin secretion by 124% and 107%, respectively (P < 0·05). Dose-ranging studies in patients with type 2 diabetes have demonstrated that liraglutide is effective and well-tolerated at doses up to 1·9 mg/day.60,66,67

The clinical effectiveness of liraglutide is being evaluated in the Liraglutide Effect and Action in Diabetes, or LEAD programme, the results of which have been used for the recent regulatory filing in Europe and the USA. The LEAD programme has compared liraglutide with widely used classes of antidiabetic drugs in a series of randomized, double-blind, controlled, 26-week studies in approximately 3800 patients with type 2 diabetes with blood glucose inadequately controlled with standard oral therapies. The data from these studies have now been presented in abstract form only and indicate that the addition of liraglutide to ongoing oral antidiabetic drugs can significantly improve glycaemic control as well as weight loss in previously uncontrolled patients with type 2 diabetes.68–71 Liraglutide reduces fasting and postprandial glucose and levels of HbA1c by up to 1·74% for the 1·9 mg dose vs. placebo.66

Like exenatide, liraglutide produces weight loss and has a very low risk of hypoglycaemia when used alone or with agents that themselves do not cause hypoglycaemia. Differentiating it from exenatide is liraglutide's once-daily usage; plus, it can be taken at any time of the day, as long as administration is at the same time day to day. The profile of action of liraglutide can be likened to basal insulin, with no peak, and while it shares similar gastrointestinal side-effects to exenatide they appear to be lessened.

DPP-4 inhibitors

Another class of agent that has recently entered the market is the DPP-4 inhibitors. These agents target the same biochemical pathway as GLP-1 agonists, but instead of supplementing the natural pool of GLP-1, they inhibit the enzyme responsible for its breakdown. As a result, the endogenous levels of GLP-1 are much lower than that observed after pharmacological administration of GLP-1 receptor agonists. The inhibitors belong to two general classes: nonpeptide heterocyclic compounds with rapid onset and duration of action, for example, sitagliptin, and cyanopyrrolidines, covalent modified ‘irreversible’ inhibitors with slow onset and more prolonged actions, for example, vildagliptin. Sitagliptin and vildagliptin are the first two agents in this class to be approved, but a number of other DPP-4 inhibitors such as alogliptin and saxagliptin are in Phase 3 trials.

DPP-4 is widely distributed in numerous tissues and exists as both membrane-spanning and soluble forms.72 The enzymatic activity of DPP-4 is exhibited by both the membrane-spanning and soluble forms of the molecule. DPP-4 preferentially cleaves substrates with a proline residue at the P1 position (first position N-terminal to the cleavage site) and accepts most residues at P2 and in prime side positions.73,74

DPP-4 inhibitors depend on endogenous GLP-1. They improve glycaemic control by preventing the rapid degradation of incretin hormones, thereby resulting in postprandial increases in levels of biologically active intact GLP-1, reducing glucose production from the liver by inhibition of glucagon from the α-cells of the pancreas and increasing insulin production. The DPP-4 inhibitors amplify the early insulin effect to glucose, and although fasting glucagon is only slightly altered, postprandial glucagon secretion is abolished.75

Selective inhibition of DPP-4 is required for an acceptable safety and tolerability profile. At least two human postproline dipeptidyl–peptidases, DPP-8 and DPP-9, whose functions are still unknown, are structurally closely related to DPP-4·76,77 Acute toxicity in animal models was reported for at least one compound with strong DPP-8/DPP-9 inhibitory potency.77 IC50 and Ki values for DPP-4 for the leading DPP-4 inhibitors show a high selectivity for DPP-4, whereas affinity for DPP-8 and DPP-9 is low.78–82 It is not known whether in vivo DPP-8 and DPP-9 are inhibited on administration of therapeutic doses of the DPP-4 inhibitors. However, to date no in vivo effects have been reported that point to a pharmacological impact of the inhibition of other peptidases.

DPP-4 inhibition has been shown to increase circulating levels of intact GLP-1 and improve glucose tolerance in a number of animal models of insulin resistance.78,81,83–85 This effect has been attributed to the higher concentrations of intact biologically active GLP-1 arising as a consequence of DPP-4 inhibition rather than increasing secretion because the ratio of active to total GLP-1 is increased and total GLP-1 levels reduced.85,86 Preclinical studies on DPP-4 inhibitors have demonstrated that glucose tolerance is improved by enhancing many of the biological actions characteristic of GLP-1 receptor agonists, including glucose-dependent stimulation of insulin and inhibition of glucagon secretion, enhancement of insulin sensitivity, and promotion of β-cell proliferation, islet cell function, neogenesis, and survival.83–85,87,88

Single-dose administration of DPP-4 inhibitors produces long-lasting DPP-4 inhibition in both healthy volunteers89 and patients with type 2 diabetes.75,86 In patients with type 2 diabetes, sitagliptin produced approximately twofold elevations in weighted average active GLP-1 levels during an oral glucose tolerance test (OGTT) 2 h postdose. At 24 h postdose, the twofold increase in active GLP-1 was still present for the 200 mg dose during either an OGTT or meal tolerance test.86 In this study, both insulin and C-peptide levels were significantly increased and glucagon levels were significantly reduced with sitagliptin at 2 h postdose.86 These changes in glucoregulatory hormones were associated with a 22–26% reduction in glucose excursion with sitagliptin treatment following an OGTT at 2 h postdose. In a similar study, vildagliptin 100 mg twice daily increased the mean GLP-1 over the 13·5-h period following dosing by approximately twofold relative to placebo after 1 day of treatment90 and demonstrated similar effects on postprandial glucose, insulin, and/or glucagon concentrations following a meal.75,90–92

Long-term DPP-4 inhibitor treatment has been shown to both preserve and increase β-cell number through an apparent stimulation of islet neogenesis, and β-cell regeneration (differentiation from precursor cells) and/or enhanced insulin biosynthesis in the rat.87 Histological examination of the pancreas following DPP-4 inhibitor administration in this study demonstrated increased numbers of islets and β-cells.87


Randomized controlled trials of sitagliptin in combination with metformin, a TZD, or a sulphonylurea have shown that it reduces HbA1c levels to a greater extent than placebo.93–95 Only one published study was designed to compare sitagliptin with an active comparator: it was found to be noninferior to glipizide when added to ongoing metformin therapy.96

Adverse events reported in the trials included nasopharyngitis, urinary tract infection, abdominal pain and headache. Sitagliptin was associated with a higher incidence of hypoglycaemia when added to glimepiride (with or without metformin),97 but not when added to metformin or pioglitazone.93,95,98,99

Sitagliptin was the first DPP-4 inhibitor to enter the market. It is indicated to improve glycaemic control in combination with metformin and/or a sulphonylurea when diet and exercise plus metformin/and or a sulphonylurea alone do not provide adequate glycaemic control.100 It is also indicated in combination with a TZD when diet and exercise plus the TZD alone do not provide adequate glycaemic control. The dose is 100 mg once daily, which can be taken with or without food.


Randomized controlled studies have shown that the addition of vildagliptin 50 mg once daily or 50 mg twice daily to metformin, glimepiride, or a TZD produces statistically significant improvements in glycaemic end-points vs. placebo at 24-weeks.101–104 When added to ongoing metformin therapy, vildagliptin was deemed to be noninferior to pioglitazone in HbA1c reduction at 24 weeks.104 The incidence of hypoglycaemia was similar to placebo when vildagliptin was added to metformin or pioglitazone therapy, although when added to ongoing glimepiride, confirmed hypoglycaemia occurred in 0·6% and 1·2% of placebo and vildagliptin 50 mg/day recipients, respectively.102

For vildagliptin, adverse events reported in at least 5% of patients included upper respiratory tract infections, nasopharyngitis, dizziness, influenza and headache.101,103–105 Skin lesions, including blistering and ulceration, may occur with both DPP-4 inhibitors and routine monitoring is recommended. Vildagliptin has been associated with rare cases of hepatic dysfunction (including hepatitis). Liver function tests should be performed prior to the initiation of treatment with vildagliptin to determine the patient's baseline value, and should be monitored at 3-month intervals during the first year and periodically thereafter.106 In contrast to the results obtained with the GLP-1 analogues, no significant decrease in body weight has been observed with the DPP-4 inhibitors either as monotherapy or as combination therapy and they are generally regarded as weight neutral.

In Europe, vildagliptin is approved as a 50-mg dose taken either once or twice daily in combination with metformin, a TZD or a sulphonylurea. In the USA, the FDA has requested more data on vildagliptin before it can be approved.


Despite the presence of well-established therapies for type 2 diabetes, there are still important unmet needs in the treatment of this chronic disease including prolonged efficacy and the potential to act on the underlying cause of the disease rather than on its symptoms. This and the rising prevalence of type 2 diabetes continue to drive growth in diabetes drug spending. The current focus is on the incretin-based therapies, with exenatide and two DPP-4 inhibitors already available, and exenatide LAR and liraglutide likely to enter the market by 2010.

Until recently patients with type 2 diabetes not adequately controlled with diet plus metformin and/or a sulphonylurea usually received additional oral medications, to which insulin was sometimes added, or insulin therapy alone. The incretin-based therapies offer an alternative approach for this patient population. They control postprandial glucose levels by reducing postprandial glucose production by the liver and augmenting insulin secretion in response to a meal. Of particular importance for patients is the finding that incretin-based therapies depend absolutely on glucose for their actions, a major distinction between them and other insulin secretagogues such as sulphonylureas. This glucose dependency provides a low risk for hypoglycaemia. Exenatide, the first agent to reach the market, is also able to promote satiety so that the patient eats less, and to slow gastric emptying so that less glucose reaches the intestine and bloodstream after a meal. Although exenatide has to be injected, it may represent a useful alternative to insulin in patients with type 2 diabetes, as, unlike insulin, it is not associated with weight gain but with significant weight loss, and it does not require as frequent blood glucose monitoring or continual dose titration.

Theoretically, exenatide may be useful in slowing the progression of type 2 diabetes due to its effect on β-cell mass and function. However, until this has been confirmed in long-term clinical trials the target population for exenatide is probably patients poorly controlled with diet plus metformin and/or a sulphonylurea or TZD but who are reluctant to move to insulin therapy out of fear of hypoglycaemia, weight gain and/or frequent glucose monitoring. In the future exenatide LAR, a long-acting, sustained release injection that can be given once a week may become available, which is likely to be more appealing to patients as it also appears to be associated with less nausea.

The two most recent agents to be approved for the treatment of type 2 diabetes are the DPP-4 inhibitors sitagliptin and vildagliptin. Like exenatide, these agents produce clinically significant reductions in HbA1c as monotherapy and in combination with metformin, sulphonylureas, and TZDs. To date they appear to be well-tolerated, not associated with weight gain and with a low potential to provoke hypoglycaemia. The main advantage of the DPP-4 inhibitors is their oral mode of administration. As a result, DPP-4 inhibitors will be able to avoid comparison with insulin and will be most likely compared to other oral medications such as TZDs. This makes them more attractive for many patients and physicians and is likely to help with uptake and increased patient compliance. However, concern remains over the long-term safety of the DPP-4 inhibitors due to specificity issues. Besides having a role in eliminating GLP-1 from the body, the DPP-4 inhibitors play an important role in the regulation of the immune system. Reports of serious allergic reactions are already reported in the package insert for these drugs. The launch of vildagliptin was delayed in Europe following safety concerns over elevated liver enzymes. The license has now been approved for a 100-mg dose when used in combination with metformin or a TZD (taken as one 50 mg dose in the morning and one 50 mg dose in the evening). When used in dual combination with a sulphonylurea, the recommended dose of vildagliptin is 50 mg once daily administered in the morning. A single dose of vildagliptin 100 mg has not been approved because of reports of more frequent liver enzyme elevations in patients taking this dosage. In the USA, the FDA has requested more data on vildagliptin before it can be approved. At this stage of development, morbidity and mortality end-points are not available for the DPP-4 inhibitors.

Although there is not yet enough experimental or clinical information available to make firm comparisons between GLP-1 mimetics and DPP-4 inhibitors, several major differences are apparent (Table 1). DPP-4 inhibitors can be given orally and appear to have minimal side-effects. However, they depend on endogenous GLP-1, and possibly other peptides, whose levels are increased only into the upper physiological range. GLP-1 mimetics such as exenatide and liraglutide, on the other hand, must be given by injection. The plasma levels of these drugs can be boosted to higher levels than those of natural GLP-1 during therapy with DPP-4 inhibitors, with potentially greater potency, but also more adverse events, particularly nausea.

Table 1.  Incretin enhancers vs. incretin mimetics. Adapted from Drucker and Nauck107
 DPP-4 inhibitorExenatide
↑Glucose-dependent insulin secretion
Restores first phase insulin secretionnot confirmed
↓Glucagon secretion
↓Hepatic glucose output
Regulates gastric emptying
↓Food intake
Weight loss
Effect on β-cell
 neogenesis/apoptosis(animal data only)
Oral administration

While the development of new classes of glucose-lowering agents has expanded the treatment options for type 2 diabetes, it has also introduced more uncertainty regarding which treatment option is the most appropriate. As several of these new agents are now available, it is essential that they are incorporated into revised treatment algorithms in the near future. Despite the fact that pipeline products will have a great impact on the diabetes market in the next 5 years, there are currently no agents in late-stage clinical development that are likely to fully replace established antidiabetic therapies such as metformin. Type 2 diabetes is therefore likely to remain an add-on market rather than develop into a switch market.

Both the incretin mimetics and the DPP-4 inhibitors offer clinically meaningful reductions in HbA1c without significant risk of hypoglycaemia and without causing weight gain. They also offer the theoretical potential of improving or maintaining β-cell function and thus, favourably affecting the progressive loss of function that is characteristic of type 2 diabetes. Continued progression of the natural history of type 2 diabetes, mainly due to ongoing loss of β-cell function, means that most patients will eventually need a combination of two differently acting classes of oral agents. If observations in preclinical studies that the incretin-based therapies can slow the natural decline in insulin production or protect against further decline are confirmed in humans, then these agents could become first-line therapies in type 2 diabetes as to maintain glycaemic control in the long-term, interventions that preserve β-cell function should probably be initiated early in the disease process.