Insulin-Regulated Glucagon-Like Peptide-1 Release from L Cells: Actin’ Out

Incidence of type 2 diabetes has now reached epidemic proportions, afflicting 8% of adults in the United States. More worrisome are predictions based upon current trends indicating incidence of 30%: one of every three children born in the year 2000 to develop diabetes in their lifetime (1). Type 2 diabetes is characterized by chronic hyperglycemia resultant from the combined dysfunction of insulin secretion from pancreatic β-cells in the face of peripheral insulin resistance. Thus, type 2 diabetic patients require multifactorial treatment strategies to regain and maintain euglycemia, which can strain both patient compliance and patient’s pocketbooks. The answer to this dilemma is to devise a magic pill that could simultaneously coordinate increases in insulin secretion with peripheral glucose clearance. Moreover, because obesity trends pair with those of type 2 diabetes, wouldn’t it be nice if such a magic pill could induce weight loss, too?

Glucagon-Like Peptide-1 (GLP-1) Is the Answer

Within the last decade, such a magic pill substance was discovered: the incretin hormone GLP-1. A seminal study from 2002 determined that continuous infusion of GLP-1 decreased fasting hyperglycemia, improved insulin sensitivity and β-cell function, and reduced body weight (2). These observations correlated beautifully with data demonstrating a paucity of GLP-1 release from the L cells of type 2 diabetic patients (3,4,5,6). Importantly, unlike insulin injection or oral sulfonylurea therapies, GLP-1 injection enhances insulin secretion but without risk of hypoglycemic episodes. Because GLP-1 specifically amplifies insulin release triggered by nutrient stimulation (7), as opposed to triggering secretion itself absent of nutrient stimulation, GLP-1 does not inappropriately elevate insulin secretion under fasting conditions. Rather, GLP-1 action works in sync with the regulated insulin secretion machinery, increasing the quantity of insulin released only when needed. To exert effects upon weight loss, GLP-1 action is coupled through decreases in gastric emptying, intestinal motility, and food intake. Despite its wonder-drug profile, development of GLP-1 therapy has met with challenges due to its inordinately short half-life (2 min) resulting from degradation upon secretion from the L cell of the intestine by the serine protease dipeptidyl peptidase-IV, and a propensity for patient intolerance due to nausea. Interestingly, efforts to increase endogenous GLP-1 levels by elongating half-life via inhibition of dipeptidyl peptidase-IV have not yielded the desired weight loss effect that is seen with GLP-1 administration (8,9), and as such, agonists to boost GLP-1 release are being sought (10,11). Given the hyperinsulinemia associated with prediabetes and insulin resistance, the current study by Lim et al. (12) raises the question as to whether hyperinsulinemia might be a direct cause of impaired GLP-1 release in type 2 diabetic patients.

The GLP-1 Release Pathway

Under normal conditions, the L cell of the intestine senses nutrients such as glucose and fats through microvilli on the apical surface and responds by triggering the fusion of GLP-1-containing secretory granules with the basolateral plasma membrane to release the processed bioactive GLP-1 peptides. Glucose-stimulated GLP-1 release is biphasic, initially rising within 10–15 min of food intake, peaking by 40 min (13). In a previous study, Lim and colleagues (14) discovered that, surprisingly, acute insulin stimulation could also trigger GLP-1 secretion from the L cell, concurrent with the activation of MAPK kinase (MEK)-1/2 and on to ERK1. Lacking, however, was a pathway describing how the insulin signal led to mobilization of the GLP-1 secretory granules within L cells. Based upon analogies drawn between islet β-cell and intestinal L cell release patterns, Lim et al. (12) sought to determine whether filamentous actin (F-actin) remodeling was involved in insulin-stimulated GLP-1 release from L cells.

F-Actin Remodeling in the L Cell

F-actin remodeling to promote vesicle/granule exocytosis events is a conserved mechanism where the ratio of polymerized F-actin to depolymerized globular actin (G-actin) transiently changes to effect mobilization or repositioning of secretory granules with respect to surface membranes. F-actin remodeling is considered a requisite event in processes prominent in the regulation of euglycemia, including glucose-stimulated insulin release (islet β-cells) (15,16) and insulin-stimulated glucose uptake/GLUT4 vesicle translocation (adipose and skeletal muscle tissues) (17,18). Common to F-actin remodeling in these processes is the use of small Rho family GTPases, in particular Cdc42, and its downstream effector protein p21-activated kinase (Pak1). Dissimilar is the impact of F-actin depolymerization upon these processes. For example, conversion of polymerized F-actin to monomeric G-actin using agents such as latrunculin or cytochalasin will potentiate insulin release from β-cells but inhibit glucose uptake into adipocytes. Lim et al. (12) show that L cells respond to latrunculin akin to that of the islet β-cells by potentiating GLP-1 release. Furthermore, a recent study suggests that the mechanisms of glucose-stimulated GLP-1 release and glucose-stimulated insulin release share a biphasic pattern of release (19) wherein intracellularly localized granules require mobilization to the cell surface to support the second phase of release. Considerable support exists for the concept that F-actin remodeling may function as a permissive barrier, corralling granules to simultaneously mobilize and position granules to support release, but only in response to appropriate stimuli (20). In another capacity, F-actin remodeling has also been suggested to facilitate the distal docking/fusion steps of exocytosis in 3T3-L1 adipocytes and β-cells, because their t-SNARE (target membrane soluble N-ethylmaleimide-sensitive factor attachment protein receptor) isofor syntaxin 4 can interact with F-actin directly, or indirectly via actin-binding proteins (21,22,23).

Actin remodeling events in β-cells involve upstream signaling via Rho family GTPases Cdc42 and Rac1. In this issue, Lim et al. (12) elegantly demonstrate a novel pathway within intestinal L cells, whereby acute insulin stimulation triggers Cdc42 activation and Pak1 activation (Fig. 1), although Rac1 appears to be dispensable in this process. Signaling through this Cdc42-Pak1 axis was a requirement for actin remodeling in L cells and the remodeling essential for insulin-stimulated GLP-1 release. Interestingly, the Cdc42-Pak1 axis was also required for activation of MEK1/2 and ERK1/2, placing Cdc42 as a critical proximal regulator of insulin action in L cells. The cell-type specificity for Cdc42 activation by insulin is notable, given that not all cells that express insulin receptors exhibit insulin-induced Cdc42 activation (Wang, Z., and D. C. Thurmond, unpublished data), suggesting involvement of signaling intermediates that are cell-type specific upstream of Cdc42 in the L cell pathway. Actin remodeling has been captured by microscopic methods in other cell types, although capture of net changes by biochemical methods has been more challenging (24). Remarkably, significant changes in F-actin to G-actin ratios were detect able in this L cell system, and in a time-dependent and dynamic manner. From a cell biology perspective, the L cell may present novel opportunities to study actin remodeling in a quantifiable manner. It will be important in future studies to determine whether insulin-stimulated GLP-1 release is also biphasic.

Figure 1.

Schematic model of insulin-induced GLP-1 release from intestinal L cells. Insulin binds to insulin receptors present on L cells to trigger the activation of Cdc42 within 10–15 min. Cdc42 expression is required for subsequent activation of Pak1 and MEK1/2 signaling cascades. Pak1 activation proceeds to induce F-actin remodeling, whereas MEK1/2 signals downstream to ERk1/2, with both pathways resulting in release of GLP-1. Insulin signaling may occur at alternate surfaces based upon reports of insulin receptor localization (25,26). Nutrient sensing through sodium-dependent glucose transporter-1 (SGLT1) and G protein-coupled receptors (GPCR) occurs as nutrients pass cell in the intestinal luminal surface (top).

Perspectives for the Future

How will characterization of this new pathway help treat type 2 diabetes? Clearly, this marks the first stage of pathway characterization for the role of insulin upon the L cells. Surprisingly, insulin stimulation under these acute experimental conditions actually increased GLP-1 release, consistent with the concept that the chronic hyperinsulinemia characteristic of early diabetes and insulin resistance may exert damaging effects upon L cell function via a feed-forward cycle. Mechanistically, one could speculate that under the hyperglycemic conditions of prediabetes, insulin release from β-cells is amplified, and then insulin triggers GLP-1 release from L cells. In turn, GLP-1 travels to β-cells to potentiate more insulin release, and the cycle repeats until euglycemia is restored or until both cell types exhaust into dysfunction and type 2 diabetes ensues. Given these scenarios, would uncoupling this new element of insulin action in L cells be a good or bad target for therapeutic intervention to combat type 2 diabetes?