Mechanism of Action of DPP-4 Inhibitors—New Insights

Glucagon-like peptide-1 (GLP-1) is an incretin hormone that results in glucose-dependent insulin secretion, suppression of glucagon secretion, a delay in gastric emptying, and a decrease in caloric intake likely secondary to centrally mediated signaling (1). It arises from posttranslational processing of proglucagon primarily in intestinal L cells and is secreted in two major forms: GLP-1(7,36) and GLP-1(7,37) (2). The majority of known biological actions of GLP-1 depend on the presence of the two N-terminal amino acids; these are removed by the enzyme, dipeptidyl peptidase-4 (DPP-4), whose substrates are polypeptides with an alanine or a proline at the second position from the N-terminal side (3). Hence, the intact (7,36) and (7,37) peptides are often referred to as “active” GLP-1, whereas the truncated (9,36) and (9,37) peptides are known as “inactive” GLP-1. The activity, affinity, and wide distribution of DPP-4 results in GLP-1 having a half-life of approximately 1 min in the circulation (4).

Consequently, if GLP-1 itself is to be harnessed effectively as a therapy for diabetes, continuous infusion would be necessary. Indeed, continuous infusion of GLP-1 in pharmacological concentrations has been shown to be an effective short-term treatment for type 2 diabetes (5). More pragmatic alternatives to this relatively cumbersome approach have included the development of GLP-1 receptor agonists with significant sequence homology to native GLP-1—and therefore capable of binding and stimulating the GLP-1 receptor—but resistant to the actions of DPP-4. In general, this class of compounds, which are administered by intermittent sc injections, exhibits similar characteristics to those observed when GLP-1 is infused in a manner that results in sustained increases of active GLP-1 concentrations. Gastrointestinal side effects and weight loss are frequent accompaniments of therapy with GLP-1 receptor agonists.

The other class of pharmacotherapeutic agents that use the incretin system are DPP-4 inhibitors, which inhibit the principal enzyme responsible for the degradation of endogenous GLP-1. By decreasing clearance of GLP-1, concentrations of active GLP-1 are increased, resulting in a lowering of fasting and postprandial glucose concentrations. There are differences in the glucose-lowering efficacy, the effect on body weight, and the side effect profiles between DPP-4 inhibitors and GLP-1 receptor agonists (6). These differences have led to speculation that alternative mechanisms of action may explain the effects of DPP-4 inhibition. Scientific inquiries in this area may ultimately help to inform a better understanding of the incretin system (7, 8).

In practice, DPP-4 inhibitors increase concentrations of both active incretin hormones, GLP-1 and glucose-dependent insulinotropic polypeptide (secreted by the enteroendocrine L and K cells, respectively, which are substrates for DPP-4). This results in improved β-cell responsiveness to prevailing glucose concentrations and suppression of glucagon secretion (9). That the incretin hormones mediate these effects has been conclusively demonstrated in the double-incretin receptor knockout (DIRKO) mouse (10). Despite increases in active GLP-1 concentrations, there are no effects, however, of these compounds on directly measured gastric emptying and gastric accommodation (11). This is a significant difference from GLP-1 receptor agonists that is not as yet completely understood. Although it is true that concentrations of active GLP-1 may not be high enough or sustained enough during DPP-4 inhibition to alter gastrointestinal motility, this is not always the case (12). Indeed DPP-4 is necessary for the activation of some pro-satiety factors (such as peptide YY) also produced by the L cell. In such cases, it is conceivable (but not proven) that the net result of DPP-4 inhibition on satiation is neutral because the increase in active GLP-1 is countered by a decrease in activated peptide YY.

It has been suggested that decreased incretin secretion may play a part in the pathogenesis of type 2 diabetes. However, the direct measurement of GLP-1 concentrations in response to an oral challenge across the spectrum of prediabetes (13) and in established diabetes (14) suggests that this is not the case. Given the decline in β-cell function associated with increasing glucose intolerance and frank diabetes, despite normal incretin hormone concentrations, it stands to reason that β-cell responsiveness to these secretagogues is progressively impaired. However, it remains unclear whether defective responsiveness to incretins is a specific defect or merely part of a global defect in responsiveness to multiple secretagogues due to a fundamental altered ability of the β-cell to secrete insulin. Indeed, a blunted insulin secretory response to multiple other secretagogues such as arginine has been well documented in the prediabetic state (15).

Another often-overlooked subtlety of DPP-4 inhibitors is the fact that they likely decrease incretin hormone secretion through negative feedback inhibition by active hormone on enteroendocrine cells (16). The mechanism by which this occurs is unclear. Moreover, the effect this has, if any, on the magnitude of glucose lowering during chronic treatment with these compounds is unknown.

In the current issue of the JCEM, Muscelli et al. (17) reexamine the mechanisms by which sitagliptin, a DPP-4 inhibitor, achieves its glucose-lowering effect. To do so, 50 subjects with type 2 diabetes were randomized to drug or placebo in a double-blind fashion after a 4-wk washout of other antihyperglycemic medication. Two studies were performed at baseline and then after a 6-wk treatment period. One of the investigations consisted of a dual-label mixed meal to enable measurement of glucose fluxes before and after treatment, whereas the other involved the infusion of dextrose in a fashion that reproduced the glucose profiles observed during the mixed meal test. Sitagliptin resulted in greater suppression of endogenous glucose production and decreased meal appearance. This was attributed to improved β-cell function and decreased glucagon concentrations. Intriguingly, their data suggested a beneficial effect of DPP-4 inhibition on insulin action. Measurement of the incretin effect (as quantified by the oral-to-iv ratio of insulin secretory responses) suggested that this was not decreased compared with nondiabetic subjects (as previously demonstrated) and was actually unaffected by sitagliptin therapy.

These effects are not wholly unexpected, given the known effects of DPP-4 inhibitors on β- and α-cell secretion, although they have not been reliably demonstrated in prior experiments using similar designs (9, 11). A likely explanation is that given the observed variability of meal appearance and endogenous glucose production, prior experiments were not powered to detect these relatively small effects. Moreover, effects of these magnitudes are more difficult to observe in subjects with better glycemic control (16). Increased portal insulin concentrations coupled with decreased portal glucagon concentrations would be expected to suppress endogenous glucose production and to a lesser extent increase hepatic meal extraction. Although the rate of systemic meal appearance is a function of gastric emptying as well as hepatic extraction, the fact that the timing of peak meal appearance as well as the overall pattern of meal appearance was unchanged strongly suggests that gastric emptying was unchanged by sitagliptin. This is in keeping with other studies where gastric emptying was measured directly (11). Muscelli et al. (17) also suggest— appropriately—that effects on insulin action are likely to be a function of an improvement in glucose toxicity rather than a direct effect of incretin hormones or the drug itself on this parameter. Indeed, GLP-1 does not seem to have very significant effects on insulin action and glucose effectiveness (see review in Ref. 18); the fact that an effect of DPP-4 inhibition on insulin action has only been observed after prolonged therapy is in accordance with the contention that this is attributable to amelioration of glucose toxicity (19).

The current study reinforces our current understanding of the mechanism of action of DPP-4 inhibitors in type 2 diabetes. However, the degree to which some of the observed effects are attributable to lowering of glucose toxicity remains unclear and would need to be addressed in future studies that control for the effect of glucose lowering per se on insulin secretion and insulin action as well as endogenous glucose production and meal appearance.

A final area of uncertainty is related to the numerous observations (3) that GLP-1-mediated insulin secretion and suppression of glucagon are glucose-dependent. There is currently some evidence to suggest that the effects of DPP-4 inhibition are less marked when the glycemic state is close to normal fasting and postprandial concentrations (16). Understanding the thresholds at which benefit occurs may help inform the rational use of such agents as monotherapy or combination therapy in diabetes.