DPP-4 Inhibition and the Known Unknown.

Science sometimes moves very slowly. Observations in 1964 (1,2) that the insulin response to oral (and enteral) glucose is substantially greater than that to an isoglycemic intravenous glucose infusion (the incretin effect), followed by characterization of the two known incretin hormones, glucose-dependent insulinotropic polypeptide (GIP) (1973) and glucagon-like peptide 1 (GLP-1) (1985), have only recently led to the development of two classes of antidiabetes drugs, GLP-1 agonists and dipeptidyl peptidase 4 (DPP-4) inhibitors, now used widely in the management of type 2 diabetes. GIP and GLP-1 are predominantly released from the proximal and distal intestine, respectively, primarily in response to enteral nutrients, and stimulate insulin secretion in a glucose-dependent manner (3). Attenuation of the incretin effect is probably an early marker of b-cell dysfunction (4). In type 2 diabetes, the incretin effect is markedly reduced (5), partly because the insulinotropic effect of GIP is essentially lost, whereas pharmacological doses of GLP-1 still stimulate insulin secretion substantially (6). GLP-1 also suppresses glucagon and slows gastric emptying (3,7). Following their release, GLP-1 and GIP are rapidly degraded by the ubiquitous enzyme, DPP-4. The concept that inhibition of DPP-4 may represent a therapy for type 2 diabetes was expressed in a seminal article by Deacon et al. (8) in 1995. DPP-4 inhibition markedly increases circulating intact (active) GLP-1 and GIP concentrations. In 2006, the first DPP-4 inhibitor, sitagliptin, was approved by the U.S. Food and Drug Administration, and some 11 different DPP-4 inhibitors are now available worldwide. Their efficacy in reducing HbA1c is comparable to other oral hypoglycemic drugs, but the risk of adverse effects, particularly hypoglycemia (because insulin stimulation is glucose dependent) and weight gain, is much less (3). It has been assumed that the increase in active GLP-1 accounts for glucose lowering by DPP-4 inhibition in type 2 diabetes—an assumption now shown to be only partly correct. The elegant study by Nauck et al. (9) reported in this issue of Diabetes provides important insights regarding the mechanisms underlying glucose lowering by DPP-4 inhibition that are consistent with, and complimentary to, outcomes reported by Aulinger et al. (10) (Fig.1). In the study by Nauck et al. (9), a total of 32 patients with type 2 diabetes managed by diet or metformin and 29 healthy control subjects received vildagliptin 100 mg or placebo for 10 days in a crossover design. Meal tests, with concurrent assessment of gastric emptying, were performed on days 9 and 10, with and without intravenous exendin [9-39], a GLP-1 receptor antagonist. The primary end point was a modified insulinogenic index, i.e., the ratio of the insulin secretory response relative to plasma glucose, for 4 h postprandially. In patients with type 2 diabetes, insulinogenic indices (based on insulin, C-peptide, or insulin secretory rates) were increased by vildagliptin and reduced by exendin [9-39], but, importantly, the difference between these conditions was only ;50%. Hence, as concluded by Aulinger et al. (10) who evaluated the effects of sitagliptin, only about half of the insulinotropic effect of DPP-4 inhibition could be attributed to GLP-1. Exendin [9-39] also accelerated gastric emptying, whereas vildagliptin apparently had no effect. The original prespecified primary end point (glucagon) could, unfortunately, not be used because of cross-reactivity of exendin [9-39] in the immunoassay for glucagon. Although both Nauck et al. (9) and Aulinger et al. (10) evaluated the contribution of GLP-1 to the insulinotropic effect of DPP-4 inhibition, there were substantial differences in study design: 1) different DPP-4 inhibitors were used, 2) the duration of DPP-4 treatment differed from 2 to 9–10 days, 3) the dose of exendin [9-39] used by Nauck et al. was lower, 4) Nauck et al. studied a mixed meal rather than a 75-g glucose drink, and 5) Nauck et al. included a control group without diabetes. The studies performed by Aulinger et al. (10) and Nauck et al. (9) provide compelling evidence that GLP-1–dependent