What if Gut Hormones Aren't Really Hormones: DPP-4 Inhibition and Local Action of GLP-1 in the Gastrointestinal Tract

It is increasingly clear that the gastrointestinal (GI) tract plays a major role in glucose regulation after meal ingestion. A prominent mechanism of GI influence over blood glucose is through the secretion of glucose-dependent insulinotropic polypeptide and glucagon-like peptide 1 (GLP-1), together termed incretins because of their actions to stimulate internal secretions (i.e., insulin). Indeed, common estimates are that 70% of β-cell release after meals is attributable to the incretins, and both of these peptides are necessary for normal glucose tolerance (1). Moreover, iv administration of GLP-1, and to a much lesser degree glucose-dependent insulinotropic polypeptide, enhances insulin secretion and improves glucose disposal in persons with type 2 diabetes. This has led to the development of two new classes of medication for diabetes, GLP-1 receptor agonists, which mimic the effects of endogenous GLP-1, and dipeptidyl peptidase IV (DPP-4) inhibitors, which prevent metabolism of the native peptide and extend its activity (2).

The pharmacological actions of the GLP-1 receptor agonists are largely predictable from what is known about the physiology of GLP-1. Because these agents are not susceptible to inactivation by DPP-4 and reach supraphysiological levels in the plasma for hours after administration, it has been presumed that they interact directly with GLP-1 receptors on β-cells to mediate their effects. The mechanism of action of DPP-4 inhibitors is less clear. These orally available small molecules provide substantial inhibition of DPP-4, at least as reflected by reduced activity of the enzyme in plasma and an increased percentage of circulating incretins retained in the intact, bioactive form. However, the absolute concentration of intact GLP-1 during DPP-4 inhibition is limited by the amount of peptide secreted from the gut, and levels increase only 1.5- to 3-fold in postprandial plasma (3). In addition, despite effective blockade of DPP-4 activity, and extension of the half-life of intact GLP-1 from approximately 1 min to 5 min (4), the residence time of incretins in the circulation is still relatively short in treated patients. Thus, there is a great disparity in the circulating amount of GLP-1 receptor agonist between diabetic individuals treated with GLP-1 receptor agonists and those given DPP-4 inhibitors. This presents a conundrum in that the clinical benefit, as measured by chronic improvement in glycemic control, does not differ dramatically between these classes of drug (5).

The relatively small dynamic range of plasma GLP-1, and the short residence time of intact GLP-1 in the systemic circulation, has raised doubts about a purely endocrine mechanism for its control of blood glucose. The central question is whether the small and short-lived fluctuations in plasma GLP-1 are sufficient to affect direct actions on target tissues such as islet β-cells. And given the only modest enhancement of GLP-1 pharmacokinetics with blockade of DPP-4, this question applies not only to normal physiology but also to patients treated with DPP-4 inhibitors. One model that has been advanced to resolve glucoregulatory actions of GLP-1 that appear outsized relative to its concentration in peripheral plasma is that of portal sensing. The hepatic portal vein, beyond acting as the conduit of all postintestinal blood, functions as a sensor for substrates and hormones. The wall of the portal vein is densely innervated, and visceral afferents respond to portal infusion of glucose and hormones. Moreover, the GLP-1 receptor is present on axons in the portal vein (6), and infusion of GLP-1 or a GLP-1 antagonist limited to the hepatic-portal bed has effects on glucose metabolism (6–9). Since the portal vein has the highest GLP-1 concentrations of any major vessel in the circulation, the portal sensing model has face validity.

Interestingly, the same visceral afferents that serve the portal vein also innervate the intestine, including the mucosa where they run near the L-cells that produce GLP-1 (10). Estimates are that nearly 50% of GLP-1 is inactivated by DPP-4 in the capillaries of the gut before it reaches the portal vein (11), so the potential for intestinal nerves to mediate GLP-1 action has been advanced as a more plausible explanation for GLP-1 action. Intestinal lymph has 10- to 20-fold higher concentrations of GLP-1 than portal plasma (12), suggesting that levels are highest in the interstitium of the gut. The enteral theory of GLP-1 action, first proposed by Holst and Deacon (12), proposes that GLP-1 activates visceral afferent nerves in the substance of the gut, initiating a neural reflex that includes responses to lower blood glucose.

In this issue of Endocrinology, Waget and co-workers (13) provide the most in-depth test of the enteral hypothesis of GLP-1 action to date. In this paper, they examine the impact of modulating local intestinal activity of DPP-4 on GLP-1 actions. Interestingly, orally administered doses of the DPP-4 inhibitor sitagliptin too low to affect systemic plasma levels of DPP-4 activity cause significant inhibition of DPP-4 in extracts of the gut. Moreover, these doses of sitagliptin improve glucose tolerance in wild-type but not GLP-1 receptor-null mice. The authors conclude from these data that DPP-4 inhibitors mediate glucose lowering in part through protection of GLP-1 in the gut, facilitating a gut-brain axis of GI hormone signaling. The authors also report that the His-Ala dipeptide that is cleaved from GLP-1 by DPP-4 causes glucose intolerance when given to mice in supraphysiological amounts and inhibits insulin secretion from isolated islets. This latter finding is completely novel, that and if verified could add a further dimension to the action of DPP-4 inhibitors as minimizing the liberation of a compound disruptive to glucose homeostasis.

The Waget study is an important advance toward understanding how DPP-4 inhibitors work and provides the most supportive data yet for a model that has generated some enthusiasm for several years. The study leaves some gaps unfilled; the evidence for neural activity due to enhanced GLP-1 activity in the gut is not strong, and the sites of relevant DPP-4 inhibition and GLP-1 receptor activation are not provided. Nonetheless, the investigators from the Burcelin lab have provided a valuable first step in defining a novel mechanism through which GLP-1 may act.

DPP-4 inhibitors have become increasingly popular agents to treat diabetes. They are reliable, safe, and well tolerated by patients. Although specifically designed for a known and rational drug target, the physiological mechanism through which DPP-4 inhibitors work has been a subject of debate. Clarifying this issue has the potential to enhance the utility of this drug class and the application of the GLP-1 system in therapeutics. Focusing on nonendocrine mechanisms of drug activity as suggested by Waget et al. (13) is a new concept, but one whose time may have come for the DPP-4 inhibitors.