In addition, GLP-1 has a spectrum of effects thought to be desirable in an antidiabetic agent, including trophic effects on β-cells, inhibition of glucagon secretion, and suppression of food intake and appetite.2,3 The native peptides cannot be used therapeutically because they are rapidly degraded and inactivated by the enzyme, dipeptidyl peptidase-4 (DPP-4). Two strategies have been proposed to overcome this drawback:4 the use of
In addition, GLP-1 has a spectrum of effects thought to be desirable in an antidiabetic agent, including trophic effects on β-cells, inhibition of glucagon secretion, and suppression of food intake and appetite.2,3 The native peptides cannot be used therapeutically because they are rapidly degraded and inactivated by the enzyme, dipeptidyl peptidase-4 (DPP-4). Two strategies have been proposed to overcome this drawback:4 the use of
- long-acting stable analogs of GLP-1, the so-called incretin mimetics; and
- inhibitors of DPP-4 to potentiate levels of the endogenously released active forms of the incretin hormones, the so-called incretin enhancers.
Compounds from both classes have recently been approved for therapy of T2DM.
Incretins and T2DM
In T2DM, pancreatic islet hormone secretion is inappropriate for the prevailing glycemia, while the accompanying insulin resistance contributes to an impairment of insulin action that may feed back to a further progressive deterioration in islet function.5 The UK Prospective Diabetes Study (UKPDS) demonstrated that β-cell function is already reduced by ~50% at the time of diagnosis, and continues to decline regardless of treatment.6 Additionally, recent studies indicate that β-cell volume in subjects with T2DM and impaired glucose tolerance (IGT) is reduced compared with matched individuals without T2DM, suggesting that this defect occurs early and is important in the pathogenesis of the disease, rather than simply arising as a consequence of hyperglycemia.7
There is also evidence that β-cell dysfunction occurs early in disease development. Subjects with IGT have reduced β-cell responsiveness to glucose.8 First-phase insulin secretion is reduced in non-diabetic offspring or first-degree relatives of patients with T2DM,9,10 and this is a powerful predictor for the progression to overt diabetes.5GLP-1 Actions
GLP-1 stimulates insulin secretion, but its effect depends on normal or elevated glucose levels; the risk of hypoglycemia due to elevated GLP-1 concentrations is thereby practically eliminated. In addition, insulin stores are maintained because insulin gene expression and all steps in insulin biosynthesis are stimulated.11 Studies in isolated human islets and rodents have demonstrated that GLP-1 has trophic effects on the β-cell, enhancing proliferation and differentiation12-14 and reducing apoptosis.15 Although these actions have still to be demonstrated clinically, the possibility arises that GLP-1 may be able to halt, or even reverse, the progressive loss of functional β-cell mass that occurs in T2DM.
GLP-1 also counteracts the hyperglucagonemia that occurs in T2DM by strongly inhibiting α-cell secretion.16 Again, this effect is glucose-dependent, and it has been demonstrated that that elevated GLP-1 concentrations do not impair the glucagon counterregulatory response to hypoglycemia.17
GLP-1 inhibits gastrointestinal motility and secretion,18 thereby delaying the supply of nutrients for absorption in the small intestine, and helping to reduce mealinduced glucose excursions. In healthy subjects, this appears to be more important than the insulinotropic effect in lowering postprandial glucose excursions.19
In humans, peripherally administered GLP-1 has a satiating effect,20,21 and when given over a prolonged period (six weeks) by continuous subcutaneous infusion results in significant weight loss.22
The Incretin Effect
The insulin response to oral glucose greatly outweighs its response to an isoglycemic intravenous glucose infusion23 due to the release of GLP-1 and GIP, which enhance β-cell responses to glucose.This phenomenon, the incretin effect,24 can account for up to 70% of oral glucose-induced insulin secretion in healthy subjects. In patients with T2DM, the incretin effect is severely impaired or absent,25 and is characterized by:
- impaired secretion of GLP-1 (whereas GIP secretion is relatively normal);26,27
- markedly impaired β-cell sensitivity to GLP-1, although its efficacy is at least partially preserved;28 and
- completely abolished effect of GIP on second phase insulin secretion.29
These defects do not precede the development of diabetes, and have been suggested to be a consequence, rather than a cause, of the condition.30,31 However, once glucose control deteriorates, incretin defects are likely to contribute to the impaired insulin secretion that characterizes T2DM.30
Although β-cell responsiveness to glucose is reduced in patients with T2DM,28 this defect is completely corrected by GLP-1, even though β-cell sensitivity to GLP-1 itself is reduced.28 Biology and Actions of DPP-4
Enzymatic Action
DPP-4 is a membrane-bound serine peptidase, found in numerous sites, including the kidney and intestinal brush-border membranes, hepatocytes, T-cells, and vascular endothelial cells, as well as in a soluble form in plasma.32 It has strict substrate requirements, but in vitro kinetic studies revealed that both incretins were good substrates.33 The truncated metabolites comprise the majority of endogenous GLP-1 and GIP immunoreactivity in humans, suggesting that DPP-4 may be important in incretin hormone metabolism.34,35 This was confirmed by the demonstration that DPP-4 inhibition prevented the in vivo N-terminal degradation of exogenous and endogenous GLP-1 and GIP.36-38 GLP-1 in particular is degraded extremely rapidly.39,40 In vitro kinetic studies have indicated that a number of peptides can be degraded by DPP-4 (see Lambeir 200332 for review), but the physiological relevance (if any) has yet to be established for most of them.41
DPP-4 is identical to the T-cell antigen, CD26, and contributes to T-cell activation and proliferation.42 However, it appears that its catalytic activity is probably not involved, and that its presence is not obligatory for normal immune function, since animal models lacking DPP-4 activity are viable and seem to suffer no ill effects.43,44 In vitro studies comparing highly selective DPP-4 inhibitors and inhibitors for other related enzymes like DPP-8 and DPP-9 have demonstrated that the catalytic activity of these other enzymes, but not DPP-4, is involved in T-cell activation and proliferation.45 Furthermore, in long-term studies the DPP-4 inhibitors in clinical development have, to date, proved to be remarkably safe and well tolerated and have not been associated with immune side effects.
DPP-4 and T2DM
Several studies have found that concentrations of intact biologically active GLP-1 are reduced in patients with T2DM,26,27 raising speculation that a more extensive or rapid degradation of the incretin hormones may contribute to the reduced incretin effect observed. However, elimination of intact GLP-1 and GIP is identical in healthy subjects and patients with T2DM.40,46 There is some evidence that plasma DPP-4 activity may be positively correlated with glycated hemoglobin (HbA1c),47,48 but this seems not to correlate with active incretin hormone concentrations.48 Nevertheless, since up to 70–80% of endogenous GLP- 1 and ~50% of endogenous GIP circulates in the degraded form in humans,34,35 inhibition of this process has the potential to raise active incretin hormone concentrations into the range shown to be therapeutically useful.49DPP-4 Inhibitors and T2DM
DPP-4 inhibitors are low molecular weight compounds that are orally available.3 Vildagliptin is suitable for both once and twice daily dosing. It has low protein binding and 85% of the dose is renally excreted as a pharmacologically inactive metabolite.50 Sitagliptin is longer acting than vildagliptin,51 and differs by being excreted in the urine primarily as the unchanged parent drug.52 Saxagliptin53 and denagliptin54 (both currently in phase III trials) are also long-acting inhibitors with an expected once daily dosing regimen. Inhibitor PSN9301, formerly developed under the name P93/01, is currently in phase II trials, and has a rapid onset and short duration of action with dosing expected to be meal-related.55
Sitagliptin was approved as both monotherapy or in combination with metformin or thiazolidinediones in October 2006. Vildagliptin was submitted for regulatory approval in the US in spring 2006 (decision expected in early 2007) and in Europe in August 2006.
Pre-clinical Experience
Initial acute studies in normal and glucose-intolerant animal models confirming the anti-hyperglycemic effectiveness of DPP-4 inhibition were rapidly followed by longer-term studies with a variety of inhibitors in different animal models, all showing improvements in glucose tolerance.3 Rather fewer studies have investigated whether DPP-4 inhibition affects β-cell mass, although the number of β-cells and small islets were increased by DPP-4 inhibition in diabetic (streptozotocin-treated) rats and insulin-resistant high fat fed mice.56,57 Vildagliptin increased β-cell mass via effects on replication and apoptosis in neonatal rats.58 In a mouse model of impaired insulin sensitivity and secretion, a sitagliptin analog increased insulin-positive β-cell number, and restored the β-cell:α-cell ratio and distribution pattern.59 These studies, therefore, suggest that DPP-4 inhibitors can have beneficial effects on the pancreatic islets, although whether this is also the case in humans remains unknown.
Clinical Proof of Concept
The first clinical study with a DPP-4 inhibitor used the short-acting predecessor to vildagliptin, NVPDPP728.60 Patients with relatively mild T2DM (mean HbA1c, 7.4%) received the inhibitor twice or three times daily as monotherapy for four weeks, resulting in reduced fasting and postprandial glucose levels and a fall in HbA1c levels to 6.9%. Similar results were obtained with once-daily vildagliptin monotherapy.61 Intact GLP-1 concentrations increased two-fold, glucagon secretion was suppressed, and insulin levels were maintained despite the lower glucose concentrations (see Figure 1).
Efficacy as Monotherapy
Three months of monotherapy with once or twice daily vildagliptin reduced fasting and postprandial glucose levels, and lowered HbA1c from baseline (7.7–8%) by an average of 0.6%, with those patients with more poorly controlled T2DM showing the greatest improvements.62,63 These reductions are sustainable, as shown in another study with vildagliptin (50mg bid) administered for 52 weeks (see Figure 2).64 Twice-daily sitagliptin gives similar results, dose dependently reducing fasting plasma glucose by up to 1.0mmol/L and HbA1c levels by up to 0.8% from baseline (7.9%), compared with placebo,65 while with once-daily administration, the average HbA1c reduction was 0.6% relative to placebo (baseline 7.8%).66 Much less is known about the other inhibitors in clinical development, although all are said to have sustained anti-hyperglycemic efficacy.54,55,67 All these compounds appear well tolerated and seem to have neutral effects on body weight.
Efficacy as Combination Therapy
In a combination study with metformin, patients on placebo (metformin alone) showed a gradual deterioration in HbA1c levels (baseline, 7.9% rising to 8.3% by one year), whereas those also receiving twicedaily vildagliptin showed sustainable improvements, giving between-group differences of -0.7% at three months and -1.1% at one year. After one year, ~40% of patients receiving vildagliptin and metformin achieved HbA1c levels of <7%, compared with 10% of patients on metformin alone.68 With once-daily sitagliptin combined with metformin, HbA1c levels fell by 0.7% (baseline 8.0%), with 47% and 17% of patients achieving target levels of <7% and <6.5%, respectively, after 24 weeks, compared with no significant changes with metformin alone.69 In both studies,68,69 metformin treatment gave modest weight reductions, but this was unaffected by either inhibitor.
In a 28-day study, once daily vildagliptin was reported to give additional glycemic control when given in combination with pioglitazone, with the combination being well tolerated without need for dose adjustment of either compound.70 Similarly, sitagliptin, given once daily with pioglitazone for 24 weeks, reduced fasting glucose (by 1.0mmol/L) and HbA1c levels (by 0.9%; baseline, 8.1%), while there were no significant changes with pioglitazone alone.71 In this study, 45% and 24% of patients reached HbA1c levels <7% and 6.5%, respectively (compared with 23% and 5% on pioglitazone alone). The combination was generally well tolerated, but with a slightly higher incidence of gastrointestinal side effects. While pioglitazone caused a small weight gain, this was not affected by concomitant sitagliptin.71
A single study with vildagliptin (twice daily) as addon therapy to glimepiride gave additional reductions in HbA1c of 0.6% (baseline 8.5%) after 24 weeks, compared to a 0.1% increase for glimepiride alone.50 When vildagliptin was given twice daily together with existing insulin therapy, HbA1c levels decreased by a further 0.7% (baseline 8.5%) after 24 weeks, whereas they remained stable with insulin alone. Despite improved glycemic control, there were significantly fewer hypoglycemic events with the combination therapy.72Comparison with Existing Therapies
It is of considerable interest to know how effective DPP-4 inhibitors are compared with existing therapies. Treatment for one year with twice-daily vildagliptin or metformin monotherapy gave sustainable reductions in HbA1c levels (-1% and -1.4%, from baseline 8.7%), although the between-group difference failed to establish non-inferiority of vildagliptin to metformin (see Figure 2). There was no weight gain with vildagliptin (in contrast to modest weight loss with metformin), and while the overall incidence of side effects was similar, there were fewer gastrointestinal side effects with vildagliptin.64
Non-inferiority of vildagliptin to rosiglitazone was established in a 24-week study, with both treatments reducing mean HbA1c levels by 1.1% from baseline (8.7%). In a subset of patients (baseline HbA1c >9%), both drugs were equally effective (reductions of 1.8% (vildagliptin) and 1.9% (rosiglitazone)). Lipid profiles were improved more with vildagliptin, and in contrast to rosiglitazone, where patients gained weight, vildagliptin was weight neutral. The overall incidence of side effects was similar, although the incidence of edema with rosiglitazone was higher (4.9% versus 2.5%).73
In another study, sitagliptin was equally effective as glipizide when either agent was added to on-going metformin (mean HbA1c reductions of 0.7% from baseline, 7.5%), with greater red
ctions at higher baseline (~-0.2% from baseline <7%; ~-1.7% from baseline >9% in both groups). After one year, ~60% of patients in each group achieved target HbA1c levels of <7%, but the incidence of hypoglycemia was higher with glipizide (32%) than sitagliptin (5%).74
DPP-4 inhibitors have not yet been directly compared with the incretin mimetics. Exenatide entered the US market in June 2005 as add-on to existing therapy (metformin, sulfonylurea, or both), while liraglutide is in phase III trials. Combination therapy with exenatide lowers HbA1c levels by ~1.0% from baseline (8.4%) with 44% reaching 7% or below. Body weight is reduced, and both weight loss and glycemic improvements are sustained.75 Liraglutide monotherapy over 14 weeks also reduces HbA1c by up to 1.7% relative to placebo (baseline 8.1–8.5%), with ~50% achieving targets levels ≤7% (versus 8% on placebo). Body weight is also reduced.76Mechanism of Action
It appears that effects on both β- and α-cells contribute to the antidiabetic effects of DPP-4 inhibitors. Vildagliptin and sitagliptin treatment decreases proinsulin/insulin ratios reflecting improved β-cell function.77,78 The insulinogenic index, insulin sensitivity, and insulin secretory tone also all improve with vildagliptin (see Figure 3).68,79,80
Copyright © 2005 American Diabetes Association, from Diabetes Care,Vol. 28, 2005;1936-40.77 Reprinted with permission from the American Diabetes Association.
Similar results are obtained with sitagliptin, demonstrating improvements in β-cell responsiveness to glucose both under fasting and post-prandial conditions.81 Glucagon levels are significantly reduced by vildagliptin,61,80,82 suggesting that glucagon suppression, leading to reduced hepatic glucose production, is an important mediator of the improvements in glucose tolerance. Indeed, when glucose metabolism was directly assessed in patients with T2DM, postprandial and fasting (overnight) endogenous glucose production was suppressed by vildagliptin, and was positively correlated with reductions in fasting plasma glucose.82Conclusions
DPP-4 inhibitors are a new class of oral anti-diabetic agents, which, in contrast to most existing antidiabetic agents, take advantage of several mechanisms of action. Both insulin secretion and biosynthesis are enhanced, while over-secretion of glucagon is suppressed, restoring the normal glucagon:insulin relationship, which is important for the regulation of hepatic glucose metabolism. As yet, β-cell trophic effects have only been demonstrated in animal and in vitro studies, but clinical studies have shown that β- cell function improves. However, it is still unknown how durable these β-cell effects are or whether they will overcome the gradual loss of functional β-cell mass, which is part of the natural progression of T2DM, to reduce the incidence or delay progression to secondary failure relative to other secretagogues. They target fasting and post-prandial glucose concentrations, both of which are believed to contribute to the development of many of the diabetic complications.
Although there may be subtle differences between the DPP-4 inhibitors in clinical development, at present there is apparently little to distinguish them in the clinic. With the exception of PSN9301, all are suitable for once-daily dosing.They seem likely to be as efficacious as currently available oral anti-diabetic agents, giving sustainable and clinically meaningful reductions in HbA1c levels, both as monotherapy and in combination with other anti-diabetic agents, and apparently can be used across a broad spectrum of patient groups (elderly; obese, poorly controlled diabetes; hepatic or renal failure).There are apparently no pharmacokinetic drug-drug interactions with other commonly prescribed agents, meaning that dose adjustments are unlikely to be needed. As a class, DPP-4 inhibitors appear to have an excellent safety profile, with little or no risk for hypoglycemia, no weight gain, and the potential benefit of addressing the islet dysfunction that characterizes T2DM.