Introduction

Glucagon-like peptide 1 (GLP-1) and its derivatives have generated considerable interest [1]. GLP-1 has the potential to normalise glucose concentrations in patients with type 2 diabetes; however, this peptide hormone is degraded and eliminated from the circulation too rapidly to be of therapeutic value. The initial step of proteolytic cleavage and inactivation is mediated by dipeptidyl peptidase-IV (DPP-IV), following which, only 10–15% of the peptide remains in its intact, biologically active form during the continuous infusion of GLP-1 [2]. DPP-IV inhibitors reverse this effect, increasing circulating levels by a factor of 4–6 during the infusion of exogenous GLP-1 [3], thus providing the rationale for their use in diabetes therapy. Phase II studies have shown that DPP-IV inhibitors can lower HbA1c levels by approximately 1%, thus confirming their therapeutic potential [4].

DPP-IV inhibitors have been developed to augment circulating concentrations of intact, biologically active, endogenously secreted GLP-1, and it is widely believed that the effects of DPP-IV inhibition are largely mediated by GLP-1. Gastric inhibitory peptide (GIP) is also a substrate of DPP-IV, but its insulinotropic effect is reduced in patients with, or at risk of, type 2 diabetes [5]. Other incretin candidates have not been convincingly characterised, and DPP-IV inhibitors are ineffective in mice in which both the GLP-1 and GIP receptors have been knocked out [6]. These observations would seem to point to GLP-1 as the sole (or major) mediator of the therapeutic effect of DPP-IV inhibition; however, there are other arguments that oppose this view.

DDP-IV inhibition causes little increase in endogenous GLP-1

Although DPP-IV inhibition results in a fourfold to sixfold increase in circulating levels of active GLP-1 following infusion of the peptide, no comparable effect upon endogenous GLP-1 secretion has been observed (Table 1, Fig. 1a). In contrast, the majority of studies have failed to demonstrate an at least 100% increase in the concentration of intact biologically active GLP-1 [7–12], and the infusion of an equivalent amount of GLP-1 (e.g. 0.2–0.4 pmol·kg−1·min−1) has little insulinotropic activity in healthy subjects and even less activity in patients with type 2 diabetes [13]. In the latter study, physiological replacement doses of GLP-1 did not produce significant increases in insulin secretion under hyperglycaemic clamp conditions in patients with type 2 diabetes [13]. Other studies have indeed found effects of relatively low doses of exogenous GLP-1 in type 2 diabetes [14], but on closer analysis of their dose–response curves for insulinotropic effects it is not clear that minor elevations in plasma GLP-1 (as demonstrated with DPP-4 inhibition; Table 1, Fig. 1) would produce a clinically meaningful insulinotropic response.

Table 1 Effects of DPP-IV inhibitors on plasma concentrations of intact, biologically active GLP-1 and total (i.e. including metabolites) GLP-1
Full size table
Fig. 1
figure 1

GLP-1 concentrations after the oral intake of nutrients expressed as the peak values after DPP-IV inhibition relative to placebo. a Intact, biologically active GLP-1, as determined with an N-terminal-specific antiserum. b Total GLP-1 including metabolites generated by DPP-IV action, as determined with a non-specific immunoassay, and indicating the total amount of DPP-IV secreted. Details of the references are provided in Table 1. All studies reported significantly higher concentrations of active GLP-1 with DPP-IV inhibition. Some, but not all studies, demonstrated a significant reduction in total GLP-1 with DPP-IV inhibition. aStudies with normal diets. bStudies with high-fat diets. *Peak GLP-1 concentrations (or integrated incremental values) significantly different vs placebo

Full size image

DPP-IV inhibitors have little effect on gastric emptying

GLP-1 slows gastric emptying [15], and dose–response curves indicate that even the lowest concentrations of (active) GLP-1 associated with insulinotropic [15] and glucose-lowering [16] effects lead to the significant deceleration of gastric emptying. This is also true for other incretin mimetics, such as exenatide [17] and liraglutide [18]. It should be noted that a decrease in the rate of gastric emptying has not been observed in clinical studies on DPP-IV inhibitors [19]. Furthermore, the studies published to date do not provide indirect evidence (e.g. delayed increments in glucose and parameters of insulin secretion following nutrient intake) of delayed gastric emptying [4]. The single exception to this is a study in monkeys, but in this case the differences in gastric emptying occurred later than the increase in intact, biologically active GLP-1 [11]. It may therefore be considered unlikely that DPP-IV inhibition leads to plasma concentrations that are high enough to delay gastric emptying and, hence, stimulate insulin secretion.

GLP-1 and incretin mimetics cause nausea/vomiting while DPP-IV inhibitors do not

In clinical studies, incretin mimetics such as exenatide [20] and liraglutide [21] invariably cause nausea and vomiting in a proportion of patients. This adverse effect is less marked with GLP-1 [1], but the absence of such side effects in studies with DPP-IV inhibitors can be interpreted as evidence against the presence of high (‘therapeutic’) levels of endogenously secreted intact GLP-1.

Meal-stimulated levels of GLP-1 fall in response to DPP-IV inhibition

Meal or oral glucose-stimulated GLP-1 secretion is reduced, even following a single oral dose of a DPP-IV inhibitor (Fig. 1b), in both dogs [22] and healthy humans [23]. The reported reduction in humans was 75%, as measured in terms of integrated incremental concentrations of total GLP-1 after an oral glucose load, using a non-specific radioimmunoassay that is not sensitive to the presence or absence of an intact N-terminus. This implies that the maximum possible circulating concentration of intact, biologically active GLP-1 is around 25% of that seen in placebo-treated subjects, leaving little potential for plasma concentrations of endogenously secreted GLP-1 to rise into the therapeutically relevant range. Furthermore, the longer-term effects of the repeated administration of DPP-IV inhibitors upon post-meal increments in intact GLP-1 are unknown. However, it appears certain that the reduction in GLP-1 secretion produced by DPP-IV inhibition is not compensated for by the hypersecretion of other incretin hormones, since GIP secretion is significantly reduced after treatment with LAF237 [23]. This finding is supported by a study showing smaller increments in intact GLP-1 and intact GIP in rats lacking DPP-IV [24].

DPP-IV inhibition has delayed effects on glucose homeostasis

GLP-1 has glucose-lowering effects, both during and after its administration [1], as do incretin mimetics such as exenatide and liraglutide [17, 18]. Based on the assumption that the effects of DPP-IV inhibition are mediated by GLP-1, one might expect that the maximum clinical effectiveness of inhibition would be reached within a few hours. This is not the case, and clinical studies have shown that reductions in fasting glucose and other measures of glycaemic control improve much more slowly, reaching maximum values after weeks or even months [4].

Conclusion

Taken together, these arguments cast doubt on the assumption that GLP-1 is the only, or at least the major, mediator of the clinical effects of DPP-IV inhibition. Therefore, more effort should be put into elucidating the role of other potential incretin hormones or neuropeptides, such as pituitary adenylate cyclase-activating polypeptide [25]. The importance of neuropeptides as mediators may have been underestimated due to the fact that their effect cannot be related to measurable plasma concentrations. Regardless of this, it may be assumed that better knowledge of the mechanisms involved in the therapeutic effects of DPP-4 inhibition will allow these agents to be used more effectively to meet the needs of patients with type 2 diabetes.