Both insulin and glucagon levels are influenced by incretin hormones: glucagon-like peptide-1 (GLP-1), which is secreted by the intestinal glucose-responsive neuroendocrine (L) cells of the intestinal mucosa after a meal, and glucose-dependent insulinotropic polypeptide (GIP), secreted by intestinal K cells that are mainly located in the jejunum and throughout the gut. Both incretins have a significant number of glucose-regulating actions, including exerting an insulinotropic effect—that is, a glucose-dependent secretion of insulin.
Both insulin and glucagon levels are influenced by incretin hormones: glucagon-like peptide-1 (GLP-1), which is secreted by the intestinal glucose-responsive neuroendocrine (L) cells of the intestinal mucosa after a meal, and glucose-dependent insulinotropic polypeptide (GIP), secreted by intestinal K cells that are mainly located in the jejunum and throughout the gut. Both incretins have a significant number of glucose-regulating actions, including exerting an insulinotropic effect—that is, a glucose-dependent secretion of insulin. GLP-1 not only stimulates insulin secretion and inhibits glucagon secretion under hyperglycemic conditions, but also slows gastric emptying and acts as a mediator of satiety in the central nervous system.2 The incretin effect is the phenomenon by which greater insulin secretion occurs after oral glucose intake than after the infusion of comparable amounts of intravenous glucose.3 In type 2 diabetes, exogenous GLP-1 can normalize blood glucose.4 Different classes of diabetes drugs act at different parts of this glucose–insulin pathway. They include agents that increase the amount of insulin secreted by the pancreas, increase the sensitivity of target organs to insulin, and decrease the rate at which glucose is absorbed in the gastrointestinal tract. The American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD) have both stated that treatment for type 2 diabetes requires a progressive pharmacological approach to cover all of these aspects.5 Most existing antidiabetic drugs were developed without a prior definition of molecular targets. The advances being made in understanding the pathogenesis of type 2 diabetes and the molecules involved provide the opportunity to develop new treatment interventions.6 This article will review each class, examine the various compounds within them, and comment on their mechanisms of action.
Sulfonylureas were the first widely used oral antidiabetic treatments, having been available in the US since 1954. They trigger insulin release by acting directly on the KATP channel of the pancreatic b-cells.7 As insulin secretion is relatively deficient in type 2 diabetes, use of insulin secretagogues is logical for patients in whom the b-cell defect is not too advanced. Treatment with sulfonylureas has been shown to reduce glyclated hemoglobin (HbA1c) by approximately 1–2%.8 Furthermore, in the UK Prospective Diabetes Study (UKPDS) it was also associated with a significant reduction in microvascular complications, with a trend toward reduction in myocardial infarction but no significant effect on diabetesrelated or all-cause mortality.9
Sulfonylureas have the advantage of being available in multiple formulations at low cost with minimal side effects, and with demonstrated efficacy in controlling hyperglycemia. Sulfonylureas are approved for use as monotherapy and in combination with insulin and all other oral agent classes except the rapid-acting secretagogs. The downsides of sulfonylureas are well-known, and include weight increase and hypoglycemia. Therefore, sulfonylureas are widely regarded as less attractive for first-line therapy in obese patients. Secondgeneration sulfonylureas, including glyburide, glipizide, and glimepiride, are more potent than the first-generation drugs (chlorpropamide, tolbutamide, acetohexamide, and tolazamide) and generally have fewer side effects and are of comparable efficacy, but cause weight gain.10 Most sulfonylureas are metabolized hepatically and cleared renally, and are therefore not recommended in patients with advanced liver or kidney disease. Given the epidemiological association between hyperinsulinemia and cardiovascular disease, there have been concerns that sulfonylureas might increase cardiovascular morbidity.11 Cardiac tissues contain KATP channels similar to those in b cells. However, there was no increase in mortality seen in the UKPDS trial.9 In addition, the newest member of the sulfonylureas class, glimepiride, binds less strongly in the myocardium and may therefore actually reduce ischemic pre-conditioning.12
Rapid- or short-acting secretagogs, also known as meglitinides, have a mode of action that is similar to that of the sulfonylureas. By closing the potassium channels of the pancreatic b cells, they open the calcium channels and enhance insulin secretion. They were developed to have a rapid onset and short metabolic half-life, resulting in preferential targeting of post-prandial hyperglycemia and decreased risk for hypoglycemia later on.13 This class consists of repaglinide and, more recently, nateglinide, which is a D-phenylalanine derivative and developed to be even shorter-acting. Meglitinides reduce HbA1c to a similar extent as sulfonylureas (about 1–2%) but require multiple daily doses. In a study involving 576 patients with type 2 diabetes, pharmacotherapy-naïve patients exhibited less weight gain with repaglinide than with sulfonylurea glyburide (2.5 versus 3.6kg, respectively), although treatment-experienced patients did not exhibit this trend.14 However, meglitinides have not been assessed for their longterm effectiveness in decreasing microvascular or macrovascular risk.
Biguanides reduce hepatic glucose output and increase uptake of glucose by the peripheral tissues, including skeletal muscle. Metformin, the only widely available biguanide, acts primarily by reducing glucose production and thus fasting hyperglycemia in the presence of a sufficient amount of insulin. Metformin only became available in the US in 1995, although it had been marketed in Europe for nearly 20 years.15 Metformin’s mechanism of action is not completely understood, but it is typically classified as an insulin sensitizer.16 As with the sulfonylureas, biguanides reduce HbA1c by approximately 1–2%.17 In contrast to sulfonylurea therapy, metformin monotherapy is associated with weight loss (or little to no weight gain) and a lower incidence of hypoglycemia. In the UKPDS, obese patients randomized to metformin gained only 1–2kg compared with gains of 5–7kg in patients receiving sulfonylurea or insulin treatment.18 The drug also has non-glycemic benefits, including reducing low-density lipoprotein cholesterol and triglycerides, and reducing the antifibrinolytic factor plasminogen activator inhibitor. In addition, its lack of b-cell stimulation may have consequent positive effects on other cardiovascular risk factors.19 The most serious complication of biguanide use is lactic acidosis, which can be fatal. Two of the three drugs in this class (phenformin and buformin) were withdrawn in the 1980s owing to this side effect. Fortunately, the incidence of lactic acidosis with metformin use is low (one case per 33,000 patient-years).20 Nevertheless, this concern does restrict metformin use to patients with sufficient renal function to avoid drug accumulation. It is contraindicated in patients with cardiac or respiratory insufficiency or other conditions associated with hypoxia or reduced perfusion, hepatic dysfunction, alcoholism, or a history of metabolic acidosis.
Introduced in 1997, thiazolidinediones (TZDs, also known as glitazones) bind to peroxisome proliferator-activated receptor gamma (PPARg), a type of nuclear regulatory protein involved in the transcription of genes that regulate glucose and fat metabolism. This class consists of rosiglitazone, pioglitazone, and troglitazone, although the latter was withdrawn in 2000 owing to a risk for hepatitis and liver damage. As with biguanides, the mechanism of action of the thiazolidinediones is not fully understood, but the two classes of drugs are known to work independently of each other.21 Neither class stimulates pancreatic islet cells to secrete more insulin. The most prominent effect of TZDs is to increase insulin-stimulated glucose uptake by skeletal muscle cells.22 This results in a reduction in insulin concentrations, often to an even greater extent than with metformin.23 Preliminary data also suggest that this drug class may actually prolong b-cell survival.24 Unlike other antidiabetic agents, TZDs have a very slow onset of action. Although effects begin to manifest within two weeks of commencing treatment, the maximal benefit is not seen for around three months. When combined with insulin or with sulfonylureas, the onset and peak effect occur more rapidly, perhaps within four weeks.25 Fluid retention is a concern with this class of drugs. A small proportion of patients develop leg edema, and in vulnerable patients there is an increased risk for heart failure, particularly for those who are also taking insulin. Weight gain is generally similar to that seen with sulfonylurea therapy (i.e. about 1–4kg, with stabilization over six to 12 months), although this may also be caused by fluid retention.26 In terms of cardiovascular disease in high-risk patients with type 2 diabetes, the Prospective Pioglitazone Clinical Trial In Macrovascular Events (PROactive) study found a statistically significant association between pioglitazone and a 16% reduction in the secondary end-point of all-cause mortality, non-fatal myocardial infarction, or stroke. However, there is also an increased frequency of heart failure and edema without heart failure compared with placebo, which detracts from the otherwise positive vascular effects.27
In Europe, TZDs are contraindicated in patients with a history of heart failure or who show current evidence of heart failure, particularly those also taking insulin. Recent US guidelines also urge a cautious approach to TZD use in patients with evidence of heart failure. They are also contraindicated in patients with active liver disease, although their effects on the liver are still under investigation; reduced levels of hepatic transaminases have been reported in several studies. Recent metaanalyses indicate concern about the safety of rosiglitazone. Among patients with impaired glucose tolerance or type 2 diabetes, rosiglitazone use for at least 12 months is associated with a significantly increased risk for myocardial infarction and heart failure, without a significantly increased risk for cardiovascular mortality.28
Incretin-based therapies are the newest additions to the diabetes armamentarium. GLP-1 and GIP are rapidly degraded by the proteolytic enzyme dipeptidyl peptidase-4 (DPP-4), and thus are only bioavailable for a very short time. Patients with type 2 diabetes usually lack the insulinotropic response to GIP, but while GLP-1 levels may also be reduced in this patient population the response is usually preserved.29 It should also be noted that not only is insulin secretion glucosedependent, but so too are the glucagonostatic effects of GLP-1.30 Thus, GLP-1 results in a much more physiological regulation of a- and b-cell function, which minimizes the risk for hypoglycemia.
Exenatide, the only currently available incretin mimetic, is the synthetic form of exendin-4, found in the saliva of the Gila monster. It has more than 50% homology with native human GLP-1 and exhibits many of the same biological effects, except that it has a much longer half-life (10 hours versus just a few minutes). Unlike native GLP-1, exenatide is resistant to degradation by DPP-4. Exenatide was approved by the US Food and Drug Administration (FDA) in April 2005 as an injection for use in combination with metformin and/or sulfonylureas.
The pharmacokinetic profile of exenatide reveals that peak plasma concentrations are reached after 2.1 hours through subcutaneous administration in the abdomen, thigh, or upper arm. Its half-life is 2.4 hours, and it is detectable in plasma for up to 10 hours postadministration. 31 It is eliminated primarily by glomerular filtration, followed by proteolytic degradation. In patients with mild or moderate renal impairment, exenatide clearance is also impaired. In patients with severe renal impairment (creatinine clearance <30ml/minute), exenatide clearance impairment is more than 10-fold, and use of the drug is not recommended.32,33 Furthermore, as exenatide affects the rate of gastric emptying, the timing of other orally available medications will need to be carefully considered.
The efficacy and safety of 5 and 10mg exenatide administered twice a day (BID) were evaluated in three multicenter, randomized, triple-blind, placebo-controlled phase III trials—the AC2993: Diabetes Management for Improving Glucose Outcomes (AMIGO) studies. All studies were 30 weeks long, and enrolled more than 1,400 patients with type 2 diabetes. Study participants had inadequate glycemic control with metformin,34 sulfonylurea,35 or a combination of both (at maximally effective doses).36 The main efficacy end-point was the change in HbA1c from baseline. Secondary measures included changes in fasting plasma glucose and bodyweight and the percentage of subjects who achieved HbA1c ²7% by the end of the trial. Nausea was the most common adverse event in the AMIGO studies, with most episodes being mild to moderate; progressive dose escalation lessened its incidence. Nausea was the reason for a drop-out rate of up to 4% in the 10mg group. Patients also taking sulfonylureas experienced a higher incidence of hypoglycemia than those taking either metformin or placebo. Exenatide has been shown to be non-inferior to insulin aspart and glargnine in terms of HbA1c reduction and has helped reduce bodyweight, while insulin therapy was associated with significant weight gain.37,38 Thirty post-marketing reports of acute pancreatitis and six cases of hemorrhagic or necrotizing pancreatitis have been reported in patients taking exenatide. No definitive causal relationship between exenatide and pancreatitis has been established; however, the FDA felt it prudent to issue a safety warning relating to these rare adverse events.39 There is some evidence that exenatide may exhibit a disease-modifying effect. Recent preliminary studies in patients who have undergone pancreatic islet transplants indicated that exenatide enhanced glycemic response and HbA1c levels.40 It has also been shown to restore the insulin secretion patterns, similar to those observed in subjects with no diabetes.41
Liraglutide, a human GLP-1 analog, is currently in late-stage clinical development. Preliminary data from clinical trials suggest that liraglutide monotherapy was more effective than glimepiride monotherapy in terms of reducing HbA1c, and was also associated with significant weight reduction in previously treated patients with type 2 diabetes.42 Additionally, liraglutide plus metformin was as effective as glimepiride plus metformin in patients with type 2 diabetes previously treated with oral antidiabetic monotherapy.43
Dipeptidyl Peptidase-4 Inhibitors
DPP-4 inhibitors are orally available agents that prevent the degradation of GLP-1 and GIP, thus increasing endogenous incretin levels. DPP-4 is rapidly and sustainably inhibited with these agents: it is seen within 30 minutes of administration and lasts for 24 hours. Sitagliptin was the first oral DPP-4 inhibitor to be approved in the US (in October 2006) for use either as monotherapy or in combination with metformin or TZDs; in Europe it is currently indicated for use only in combination with metformin or TZDs. Vildagliptin was approved for use in Europe in September 2007, but is still under review in the US. Therapeutic doses such that DPP-
is >80% inhibited are sitagliptin 200mg and vildagliptin 100mg.44 In a trial of 552 patients, sitagliptin monotherapy was shown to reduce HbA1c over a 12-week period by 0.4% (at 25mg once daily [QD]) and 0.6% (100mg QD) from a baseline of 7.7%.45 Similar effects were reported in a study involving 743 patients over 12 weeks.46 Sitagliptin does not appear to have an effect on bodyweight and it has been shown to be generally well tolerated, with no increased risk for hypoglycemia compared with placebo. Adverse events reported in at least 5% of patients—more frequently in the study group—include upper respiratory tract infections and nasopharyngitis. Vildagliptin over 12 weeks at 25mg BID was shown to reduce both fasting and prandial glucose, while HbA1c was reduced by 0.6%.47 In a 24-week parallel-group study in drug-naïve type 2 diabetes patients, vildagliptin (50mg or 100mg QD or 50mg BID) significantly reduced HbA1c compared with placebo.48
Compared with other medications, DPP-4 inhibitors generally show non-inferiority, although there as have yet been no comparisons with insulin. In a 24-week study on vildagliptin as monotherapy (50mg BID) in 459 patients, DDP-4 inhibition showed a reduction in HbA1c by 1.1% from a baseline of 8.7% and was non-inferior compared with rosiglitazone. Patients on vildagliptin did not gain weight, while those on rosiglitazone did.49 A 52-week study using vildagliptin as monotherapy at 50mg BID in 526 subjects with type 2 diabetes showed a reduction in HbA1c of 1% from a baseline of 8.7%, although non-inferiority to metformin was not reached.50 Overall, vildagliptin is generally well tolerated, with no increased risk for hypoglycemia compared with placebo. Adverse events reported in at least 5% of vildagliptin patients include upper respiratory tract infections, nasopharyngitis, dizziness, influenza, and headache. A number of DPP-4 inhibitors are currently in late-stage clinical development, including alogliptin and saxagliptin. The efficacy and safety of both agents have been evaluated as monotherapy and as combination therapy. Early data suggests that both alogliptin and saxagliptin also effectively lower glucose and HbA1c as monotherapy in treatment naïvepatients. 51,52 These agents also appear to be effective as combination therapy in type 2 diabetes patients inadequately controlled on monotherapy with currently available antidiabetic agents.53–56 In addition, currently available data suggest that they have a similar safety profile to sitagliptin and vildagliptin.
To date, the clinical outcomes of incretin-based therapies have not been directly compared in a type 2 diabetes patient population, and can be attempted only indirectly. However, comparisons based exclusively on the HbA1c effect are of little relevance for the purpose of estimating practical clinical benefit, as populations and designs differ among studies. Study participants in the exenatide studies typically had a longer diabetes duration (4.9–9.9 years) compared with the DPP-4 inhibitor study patients (1.9–6.5 years). One may conclude that DPP-4 inhibitors are a good treatment option for patients with a shorter diabetes duration as they do not cause weight gain and have low hypoglycemic risk—important advantages over sulfonylureas or thiazolidinediones. Nevertheless, the effect of exenatide in terms of glycemic control has been demonstrated in insulin non-inferiority studies. Exenatide is the only antidiabetic with an additional weight-reduction effect—a desired goal of treatment in any stage of type 2 diabetes, especially in obese patients.57
Summary and Conclusions
Type 2 diabetes is a progressive disease that is a growing health problem. Several classes of drug are available to treat the condition, many of which have a long history of use. Some of the earlier drugs have known side effects that limit their use in certain patient populations. For example, sulfonylureas can cause weight increase and hypoglycemia and are not advised in patients with advanced liver or kidney disease; rapid-acting variants of this class appear to cause fewer side effects, but are not as well studied. Metformin is part of a widely prescribed class of drugs, but may have GI side effects in a significant percentage of patients. It also may have a rare risk for lactic acidosis, and may be contraindicated in patients with cardiac or respiratory insufficiency. TZDs have a very slow onset of action and have an attendant risk for fluid retention. There is also some concern about possible cardiovascular effects such as aggrevation of congestive heart failure. Incretin-based therapies are the newest additions to the antidiabetic fold, and offer the promise of controlling blood glucose through stimulation of the pancreatic b cells. The incretin mimetic exenatide mimics human GLP-1 but is not degraded by DPP-4, thus extending its half-life. It has so far been shown to reduce HbA1c and stimulate weight loss without serious adverse events. It is also non-inferior to many other available antidiabetic drugs, and may even be diseasemodifying. DPP-4 inhibitors have also had positive results, although they are not associated with weight loss and have not yet been compared with either insulin or exenatide. Nevertheless, the addition of these new drugs brings new hope to those with type 2 diabetes, and more studies will help reveal the best patient populations for each one.