GLP-1: target for a new class of antidiabetic agents?

In 1998 a large highly powered trial, the UK Prospective Diabetes Study (UKPDS), demonstrated that in type 2 diabetes several clinical endpoints were improved by better diabetic control. A 0.9% decrease in glycosylated haemoglobin (HbA1c) over ten years was associated with a 25% reduction in microvascular complications (P=0.0099) and a less impressive 16% reduction in myocardial infarction (P=0.052).¹ With this evidence base, the pressure is on for physicians to improve the HbA1c of people with diabetes.

CURRENT ANTIDIABETIC DRUGS

Therapies currently available for type 2 diabetes are limited, and all have drawbacks. Metformin, which reduced mortality and morbidity in obese people with type 2 diabetes in the UKPDS, is widely used as therapy for such patients.² Contraindications are renal failure, heart failure and liver dysfunction but these are relative: in one study a quarter of patients treated with it had contraindications,³ and many clinicians use metformin despite moderate renal dysfunction (e.g. creatinine less than 150 μmol/L), chronic stable heart failure or mild liver dysfunction when transaminases are less than three times normal. A rare side-effect is lactic acidosis, probably avoidable if the drug is not used in severe renal or liver failure.⁴ In addition, many patients cannot tolerate the gastrointestinal side-effects.⁵

The other compounds commonly used to treat type 2 diabetes are sulphonylureas. These drugs have likewise been shown to improve morbidity (at least microvascular).¹ However, sulphonylureas carry a substantial risk of hypoglycaemia,⁶ particularly in elderly patients and those with poor renal function. The rate of major hypoglycaemic events in the UKPDS was 1-1.4% per year. In addition, sulphonylureas caused weight gain of 1.7-2.6 kg.¹ Metformin and sulphonylureas are commonly prescribed in combination; however, in the UKPDS, the group in whom metformin was added to a sulphonylurea had a 96% higher rate of diabetes-related death than those treated with sulphonylureas alone.² Although the metformin-sulphonylurea group was small and this statistical analysis was not a primary endpoint of the study, concern remains regarding the combination.⁷ Two new sulphonylurea-like drugs, nateglinide and repaglinide, bind to the sulphonylurea receptor and stimulate insulin secretion. Their action is short-lasting and hypoglycaemic episodes are less trouble-some than with established sulphonylureas; however, their usefulness alone appears limited to early type 2 diabetes and, even then, they may be less effective than established agents in reducing HbA1c.^8,9

The alpha-glucosidase inhibitor acarbose delays intestinal carbohydrate absorption. It appears less efficacious than other antidiabetic drugs¹⁰ and has not proved as successful—not least because of its gastrointestinal side-effects. In contrast, the thiazolidinediones are increasingly prescribed. Rosiglitazone and pioglitazone are peroxisome-proliferator-activated receptor gamma (PPARγ) agonists which alter transcription of several genes involved in carbohydrate and lipid metabolism. These agents decrease insulin resistance¹¹ and seem to be as potent as sulphonylureas or metformin.^12,13 Unfortunately, thiazolidinediones induce weight gain and can cause fluid retention and are thus contraindicated in heart failure. In recent NICE guidelines thiazolidinediones are recommended only in combination with metformin or sulphonylureas, in patients who either cannot tolerate a combination of the latter two drugs through side-effects or have a contraindication to one of them. In reality, clinicians are starting to use thiazolidinediones outside the NICE guidelines and even beyond the terms of the UK drug licence, particularly as triple therapy with sulphonylureas and metformin. There is very little published information on the safety or efficacy of triple therapy; glycaemic control does seem to improve, though at the expense of more hypoglycaemic events, weight gain and oedema.¹⁴

GLP-1

All the current treatments for type 2 diabetes have important limitations, so the search is on for alternatives. Glucagon-like peptide-1 (GLP-1) analogues seem an attractive possibility.

With oral ingestion of glucose, plasma concentrations of insulin are about twice those induced by intravenous infusion of an equivalent dose of glucose.¹⁵ GLP-1 and GIP (glucose-dependent insulinotropic peptide) are responsible for most of the differences between these two values, known as the incretin effect.¹⁶ GLP-1 has a physiological role in the incretin effect in some species though not all.^17-19 When infused in people with type 2 diabetes, GIP appears ineffective.²⁰ By contrast, GLP-1 decreases glucose levels,^20,21 stimulates insulin secretion, decreases glucagon, delays gastric emptying, reduces food intake, stimulates beta-cell neogenesis, may enhance insulin sensitivity,²² and may inhibit beta-cell apoptosis.²³ Its effect is glucose-dependent, lessening though probably not removing the risk of hypoglycaemia.^24,25 Thus, GLP-1 has several potential advantages over current treatments for type 2 diabetes. ‘Proof of principle’ for GLP-1 as a therapeutic agent was demonstrated by regular subcutaneous injections²⁶ and by intravenous²⁷ or subcutaneous infusion.^28,29 The circulating half-life of GLP-1 is about one minute,³⁰ making it an unlikely diabetic agent, but several strategies have been explored to utilize the principle.

GLP-1 is broken down by the enzyme dipeptidyl peptidase IV (DPP IV).^31,32 Mice lacking this enzyme (DPP IV knockout) show better glucose tolerance, higher GLP-1 levels and greater insulin sensitivity than their non-knockout equivalents,³³ and less obesity and insulin resistance when fed a high fat diet.³⁴ Various DPP IV antagonists and DPP IV resistant analogues of GLP-1 are under investigation.

DPP IV ANTAGONISTS

P32/98, NVP-DPP728 and FE 999011 are DPP IV antagonists. Treatment of Zucker fatty rats (a model of type 2 diabetes) with P32/98 for three months caused sustained improvement in glucose tolerance,³⁵ and mice fed a standard or high-fat diet had better glycaemic control after eight weeks of NVP-DPP728.³⁶ P32/98 stimulated islet neogenesis and beta-cell survival in rats with streptozotocin-induced diabetes, suggesting possible usefulness in type 1 or late type 2 diabetes.³⁷ Administration of FE 999011 to Zucker rats for seven days delayed the onset of diabetes.³⁸

Published work with DPP IV in man is limited. Twice or three times daily oral treatment with NVP-DPP728 for four weeks reduced HbA1c by 0.5%.³⁹ Fasting, postprandial and mean 24-hour glucoses were all reduced, but body weight was unchanged. The medication was generally well tolerated in this patient group, although one out of a group of sixty-five developed transient nephrotic syndrome and was withdrawn from the study.³⁹ Pharmacokinetic assessment of NVP-DPP728 and its daughter compound NVP-LAF237 in monkeys indicates that NVP-LAF237 is suitable for once daily administration and this seems a better therapeutic option, though both products are currently in phase II clinical testing.⁴⁰

DPP IV is not specific to GLP-1 and breaks down several other peptides including neuropeptide Y, peptide YY and GIP as well as chemokines such as macrophage-derived chemokine and eotaxin.⁴¹ Whether increases in the half-lives of some or all of these compounds will cause side-effects awaits further evaluation.

GLP-1 ANALOGUES

Long-acting GLP-1 receptor agonists such as exendin-4 and liraglutide (NN2211) are also under therapeutic investigation. Exendin-4, isolated from the salivary gland of the Gila monster (Heloderma suspectum [Figure 1]), has 53% sequence homology to GLP-1 and is a high-affinity GLP-1 receptor agonist.⁴² Exendin-4 has a longer duration of action than GLP-1: it improved glycaemic control in diabetic mice, rats and baboons and decreased food intake and body weight in Zucker rats.^43,44 Exendin-4 stimulated beta-cell replication and neogenesis, improving glucose tolerance,⁴⁵ and stimulated non-insulin-secreting pancreatic cells into producing insulin.⁴⁶ The beta-cell neogenesis may occur via increased expression of the homeodomain protein IDX-1,⁴⁷ the lack of which results in failure of pancreas development.⁴⁸ Interestingly, transgenic mice processing excess exendin-4 exhibited improved glucose tolerance and ate less food in the short term but had normal beta-cell mass and islet histology.⁴⁹ Injection of exogenous exendin-4 to streptozotocin treated newborn rats increased beta-cell mass though glucose-stimulated insulin secretion was unaltered.⁵⁰ Exendin-4 increased beta-cell mass and delayed the onset of diabetes when administered in the prediabetic state to two rodent models of diabetes.^51,52

Figure 1.

The Gila monster (Heloderma suspectum) (courtesy of Dr Mark Seward, www.drseward.com)

Intravenous infusion of exendin-4 in healthy volunteers was well tolerated at one dose but doubling of the dose caused postprandial nausea in some and trebling caused vomiting in most.⁵³ The half-life of intravenous exendin-4 was about 30 minutes. It decreased fasting and postprandial glucose and reduced food intake by 19%.⁵³ Exendin-4 seems not to affect insulin sensitivity in healthy volunteers.⁵⁴

More data are available for exendin-4 (exenatide is the synthetic peptide) than for the DPP IV antagonists. Exenatide is insulinotropic in healthy volunteers and people with type 2 diabetes.⁵⁵ Subcutaneous injection of exenatide prevented any postprandial rise in glucose for 300 minutes in people with diabetes whether on diet, oral antidiabetic agents or insulin.⁵⁶ This effect seemed partly due to delayed gastric emptying and glucagon suppression. Exenatide decreased fasting glucose with a maximum effect 3-4 hours after subcutaneous injection, seemingly via insulin stimulation. Though no individuals withdrew from these studies there were slightly more gastrointestinal side-effects in the treatment groups.⁵⁶

Subcutaneous injection of exenatide, in patients on no other antidiabetic therapy⁵⁷ or as an additive treatment to metformin or sulphonylureas,⁵⁸ caused a drop in HbA1c of 0.8% and 0.6%, respectively. However, there was no control group in the first study⁵⁷ and neither study showed a change in body weight. Serious side-effects were rare; reported nausea declined over time and hypoglycaemia occurred only in patients also taking sulphonylureas. There was no difference between twice daily and three times daily dosing,⁵⁸ but once daily was insufficient.⁵⁷ Exenatide therapy is likely to require twice daily subcutaneous injections, although a long-acting preparation is under investigation.⁵⁹

Liraglutide is an acylated derivative of GLP-1 with an aminoacid substitution at position 34 protecting it from DPP IV degradation and increasing the half-life to about 14 hours in pigs.⁶⁰ Twice daily subcutaneous injections of liraglutide for ten days reduced body weight in normal and obese rats,⁶¹ and a similar protocol in ob/ob and db/db diabetic mice for two weeks improved glycaemic control and increased beta-cell mass (though only significantly in the db/db mice).⁶² Twice daily injections of liraglutide for six weeks caused weight loss in normal rats.⁶³

Subcutaneous liraglutide has a half life of 11-15 hours in healthy volunteers.⁶⁴ Subcutaneous injection of liraglutide at night to people with type 2 diabetes decreased fasting glucose the next morning as well as postprandial glucose 12.5 hours later, seemingly at least partly via delayed gastric emptying, glucagon suppression and stimulation of insulin.⁶⁵ The two patients with the highest concentrations of liraglutide developed nausea and one was unable to eat the meal. Liraglutide was more potent during hyperglycaemia when administered before a graded glucose infusion in people with type 2 diabetes.⁶⁶ No long-term studies are published to date. Exenatide and liraglutide are both the subjects of continued clinical trials.

Another method for protection of GLP-1 from degradation by DPP IV is via drug affinity complex technology. A preparation in which CJC-1131 binds GLP-1 to albumin in vivo gives better glycaemic control in mice. No human data are available, though the product is in phase II trials.⁶⁷ DPP IV inactivates GLP-1 by removal of the N-terminal dipeptide His(7)-Ala(8).³² Several GLP-1 analogues with longer half-lives have been assessed, particularly those with substitutions or insertions of aminoacids at positions 7, 8 or 9.^68-72 To date there are no publications of the effects of these compounds in people with type 2 diabetes.

CONCLUSIONS

The range of drugs against diabetes is growing but is still inadequate. DPP IV antagonists and GLP-1 analogues have advantages over current therapies, particularly in terms of hypoglycaemia risk and potential weight loss. NVP-LAF237 appears the most advantageous DPP IV antagonist. It is an oral agent, seemingly without side-effects. However, the likely increase in other products of DPP IV inhibition, together with multiple daily dosing, reduces its potential impact. More clinical data are available for the GLP-1 analogue exenatide. Exenatide causes weight loss and may result in beta-cell restoration, but it seems to produce nausea and has to be injected.

Therapy for diabetes will probably not alter radically in the next few years unless long-term data demonstrate other advantages over metformin and insulin. However, since the number of people with diabetes is increasing rapidly, agents modulating GLP-1 are likely to be licensed, with second or third generation molecules possibly playing a major role in combating the world-wide burden of diabetes in the 21st century.