Incretin based therapies: Do they hold their promise?

This issue of the journal contains a report on glucagon-like peptide (GLP)-1 based therapies for type 2 diabetes mellitus (T2DM) authored by scientists from Amylin Pharmaceuticals (San Diego, CA, USA), which produces one of the pharmaceuticals covered by this designation. A few comments regarding the development and status of these therapies might therefore be appropriate.

With the discovery of the insulin-stimulating properties of GLP-1,¹ it was clear that a new incretin hormone had been identified, but this did not immediately suggest that a new diabetes therapy was at hand – it was already known that the other incretin hormone, glucose-dependent insulin tropic polypeptide (GIP), had no effect on insulin secretion in patients with T2DM.² When it was found that GLP-1 also has pronounced inhibitory effects on glucagon secretion and had effects on gastrointestinal motility and gastric emptying,³ it was clear that something else might be expected from the new hormone. Indeed, early studies demonstrated the powerful insulin-sparing effect of GLP-1⁴ and in 2002 it was demonstrated that a 6-week, continuous, subcutaneous infusion of slightly supraphysiological amounts of GLP-1 dramatically improved glycemic levels in patients with long-standing T2DM and an almost complete loss of glucose-induced β-cell function.⁵ In addition, glucagon levels were suppressed, Hb_A1c levels were significantly improved by 1.3%, fructosamine levels were normalized, β-cell function and insulin resistance were massively improved, and there was a significant weight loss and no side effects. The extremely rapid inactivation of GLP-1 by the ubiquitous enzyme dipeptidyl-peptidase-4 (DPP-4) necessitated development of resistant analogs for better clinical efficacy,⁶ but also inspired the development of inhibitors of DPP-4 in an attempt to prolong the survival of the endogenous molecule,⁷ and, indeed, such inhibitors are capable of completely protecting the endogenous molecules from DPP-4 mediated inactivation. GIP turned out to be a substrate for DPP-4 as well, so the use of DPP-4 inhibitors also enhances the survival of this hormone,⁸ hence the designation “incretin enhancers” for the inhibitors. The designation incretin mimetics for the GLP-1 agonists discussed below is less fortunate, since GIP mimetics are not effective in T2DM. The pharmaceutical company Novartis pioneered the development of the first DPP-4 inhibitors for human use, and the inhibitor vildagliptin was eventually approved for clinical use in large parts of the world. Other companies followed suite and Merck were able to launch their inhibitor sitagliptin in 2006 on the US market, and thus sitagliptin became the first available DPP-4 inhibitor.

Regarding the GLP-1 receptor agonists, it was soon clear that only minor modifications needed to be introduced into the molecule to make it resistant to DPP-2 (namely, conservative substitutions in position 2),⁶ but such analogs were still cleared rapidly by the kidneys (resulting in half-lives of approximately 4 min – still not good enough for clinical use). In 1993, it was discovered that a peptide, exendin-4, from the saliva of the Gila monster, a poisonous lizard from Arizona (see Holst for the history of its development)⁹ was a full agonist for the GLP-1 receptor,¹⁰ and was cleared by the kidneys by glomerular filtration only; this would give it a plasma half-life of 30 min,¹¹ corresponding to a subcutaneous half-life of 2–3 h, enough for clinical use. The peptide was subsequently developed clinically by the Amylin Corporation in San Diego, and a synthetic form, exenatide (trade name Byetta), was introduced to the market in the US in 2005. For full efficacy, exenatide has to be injected twice daily, as opposed to the DPP-4 inhibitors that can be taken orally once daily. Although a potent GLP-1 receptor agonist and 53% homologous with the “common” region of human GLP-1, exendin-4 is still a foreign peptide, with no homologs in mammals (actually the Gila monster has its own more closely-related GLP-1 molecule).¹² This also means that it is fairly antigenic in humans, and antibodies are found in about half of those treated with Byetta.¹³ However, although a fraction of patients exhibit decreased efficacy because of this, there haven’t been any major clinical problems with the antigenicity of the peptide. A “once-weekly” formulation of exenatide is being marketed in certain countries.¹⁴ NovoNordisk in Denmark selected the human GLP-1 sequence for acylation (with palmitic acid), which causes the peptide to self-associate and bind to albumin. This gives it DPP-4 stability, slow absorption from the injection site, and a long plasma half-life, that is, approximately 12 h. A single subcutaneous injection thus results in an exposure lasting much longer than 24 h, and after a few daily injections a rather stable plasma plateau is reached. Its pharmacological properties are therefore very similar to those of the “once-weekly” exenatide, except that it has to be injected once daily.¹⁵ As expected, its antigenicity is much lower than that of exenatide and “once weekly exenatide”, and antibody formation is infrequent.^13,16

In clinical praxis worldwide, the problem with both DPP-4 inhibitors and GLP-1 agonists is that they are fairly expensive, compared to standard treatment with metformin and sulfonylureas. So are they any better?

The DPP-4 inhibitors have very few side effects (large safety database reports essentially show no differences between placebo and drug);¹⁷ furthermore, they are weight neutral. They work very well with metformin (may even be used in a supra-additive manner) making combination therapy very attractive. Since metformin is also weight neutral, a rather effective therapy can be devised that does not lead to weight gain, which will be attractive to many patients.¹⁸ In addition, the DPP-4 inhibitors, as well as the combination with metformin, are unlikely to cause hypoglycemia. Because of their ease of administration, lack of side effects and effectiveness, one might suggest that DPP-4 inhibitors should be used for very early treatment, perhaps even with a preventive purpose in patients at risk of developing diabetes. Unfortunately, very little has been done to investigate this. Instead the pharmaceutical companies have been forced to study cardiovascular safety of all new T2DM therapies, including DPP-4 inhibitors and GLP-1 agonists. This requires enormous studies, including up to 10 000 patients studied for up to 5 years, and that is a big mouthful, even for the largest companies. One may fear that these cardiovascular risk studies may turn out to be harmful, since the patients enrolled need to be at a high risk for cardiovascular events, and, although there are no signs that the incretin therapies will increase cardiovascular risk (on the contrary!),¹⁹ the patients will necessarily have long-standing and severe disease, for which the incretin-based therapies, like all other therapies, are likely to be less effective. So from these studies it may be concluded erroneously that incretins are are not as effective as predicted from registration studies. It should be noted that elderly, frail patients represent another attractive target for DPP-4 inhibitor treatment, because neither the patients nor the prescriber need to worry about adverse effects.

The great virtue of the GLP-1 agonists, apart from being somewhat more effective in terms of Hb_A1c reduction compared to the inhibitors, is that they inhibit appetite and may cause weight loss. Indeed, in studies extended over 2 years, liraglutide has proven effective in reducing weight in non-diabetic individuals by up to 10% of body weight.²⁰ In these studies, about one-third of the patients were glucose intolerant and, in most of these, glucose tolerance normalized during therapy. Treatment with GLP-1 agonists may therefore be useful in preventing the development of T2DM in obese individuals at risk. The necessity of injections is unlikely to present a major problem since both exenatide and liraglutide are provided with “pen” devices with very thin needles. The problem with the injectables is the gastrointestinal side effects, including nausea and vomiting. However, using a titration scheme for treatment initiation these problems may be avoided, particularly with the long-acting preparations, liraglutide and “once weekly” exenatide.¹³

Since GLP-1 and its analogs have been demonstrated in vitro and in animal experiments to be β-cell protective,²¹ the long term clinical course during GLP-1 agonist treatment has been awaited with great interest. There is data suggesting that β-cell function is better preserved after 3 years of agonist treatment than with insulin therapy to similar targets,²² but further studies are needed to document this. For both inhibitors and agonists, there is a great need for studies such as the ADOPT study,²³ where β-cell function was followed for 4 years after diagnosis and the start of treatment. This clearly demonstrated superior durability of thiazolidinediones versus metformin, which again was superior to sulfonylureas. Today it is unknown where the inhibitors and agonists would be positioned in a similar comparison.

Where would one position the GLP-1 agonist in the treatment algorithm? In view of their effectiveness and beneficial effects on both glycemia and body weight, and mild side effect spectrum, their prime use should be as second in-line therapy after metformin failure, where they are more effective that sulfonylureas, and do not cause weight gain or hypoglycemia. In all clinical studies reported so far, the GLP-1 agonists have been equally effective to insulin therapy for type 2 diabetes in “treat-to-target” paradigms. Recently, GLP-1 agonists have been combined with basal insulin for more advanced T2DM, and, in such studies, up to 80% of patients have reached their glycemic target (<7% Hb_A1c).²⁴ This could indicate that, with the combination of a GLP-1 agonist and basal insulin, we have finally reached the ultimate treatment of T2DM, equal in efficacy to bariatric surgery, for instance.

Recently there has been considerable discussion about the ability of DPP-4 inhibitors and agonists to cause pancreatitis and cancers, particularly in the pancreas and thyroid. For the DPP-4 inhibitors, it seems highly unlikely that they should have such side effects,¹⁷ but for the agonists the number of cases of acute pancreatitis reported is somewhat larger. There is no clear mechanism for this (in fact, in animal studies, the agonists have protected against pancreatitis)²⁵ and studies of huge health claim databases with more than one million patients have not supported such claims.²⁶ GLP-1 agonists have caused C-cell tumors in rodents in toxicology studies, but no cases have been observed in clinical studies.²⁷ A much discussed report from the summer of 2011 claiming increased rates of both pancreatitis and pancreatic and thyroid cancer on the basis of the US Food and Drug Administration (FDA) adverse effects data base,²⁸ probably represents improper use of this data base, in which the frequencies of adverse effects are influenced by factors such as public awareness (note that the FDA issued a black box warning regarding pancreatitis and Byetta treatment in 2007).

Thus, based on the available evidence both DPP-4 inhibitors and GLP-1 agonists represent valuable additions to the therapeutic armamentarium appropriate for the treatment of distinct segments of the T2DM disease spectrum.