Chronic hyperglycemia downregulates GLP-1 receptor signaling in pancreatic β-cells via protein kinase A

Abstract

Objective: Glucagon-like peptide 1 (GLP-1) enhances insulin secretion and protects β-cell mass. Diabetes therapies targeting the GLP-1 receptor (GLP-1R), expressed in numerous tissues, have diminished dose-response in patients with type 2 diabetes compared with healthy human controls. The aim of this study was to determine the mechanistic causes underlying the reduced efficacy of GLP-1R ligands.

Methods: Using primary mouse islets and the β-cell line MIN6, outcomes downstream of the GLP-1R were analyzed: Insulin secretion; phosphorylation of the cAMP-response element binding protein (CREB); cAMP responses. Signaling systems were studied by immunoblotting and qRT-PCR, and PKA activity was assayed. Cell surface localization of the GLP-1R was studied by confocal microscopy using a fluorescein-tagged exendin-4 and GFP-tagged GLP-1R.

Results: Rodent β-cells chronically exposed to high glucose had diminished responses to GLP-1R agonists including: diminished insulin secretory response; reduced phosphorylation of (CREB); impaired cAMP response, attributable to chronically increased cAMP levels. GLP-1R signaling systems were affected by hyperglycemia with increased expression of mRNAs encoding the inducible cAMP early repressor (ICER) and adenylyl cyclase 8, reduced PKA activity due to increased expression of the PKA-RIα subunit, reduced GLP-1R mRNA expression and loss of GLP-1R from the cell surface. To specifically examine the loss of GLP-1R from the plasma membrane a GLP-1R-GFP fusion protein was employed to visualize subcellular localization. Under low glucose conditions or when PKA activity was inhibited, GLP-1R-GFP was found at the plasma membrane. Conversely high glucose, expression of a constitutively active PKA subunit, or exposure to exendin-4 or forskolin led to GLP-1R-GFP internalization. Mutation of serine residue 301 of the GLP-1R abolished the glucose-dependent loss of the receptor from the plasma membrane. This was associated with a loss of an interaction between the receptor and the small ubiquitin-related modifier (SUMO), an interaction that was found to be necessary for internalization of the receptor.

Conclusions: These data show that glucose acting, at least in part, via PKA leads to the loss of the GLP-1R from the cell surface and an impairment of GLP-1R signaling, which may underlie the reduced clinical efficacy of GLP-1R based therapies in individuals with poorly controlled hyperglycemia.

Keywords: GLP-1 receptor; Hyperglycemia; Protein kinase A; Small ubiquitin-related modifier.

Figure 1

Chronic high glucose impairs insulin secretory responses. Primary mouse islets (A) and MIN6 cells (B) were cultured chronically at low glucose (white bars) or high glucose (black bars). Islets and MIN6 cells pre-cultured at low glucose were either left at low glucose (LG) or switched to high glucose in the absence (HG) or presence of exendin-4 (HG + E). Islets and MIN6 cells pre-cultured at high glucose were either left at high glucose (HG) or had exendin-4 added to the medium (HG + E). Insulin values were expressed as a percentage of the cellular insulin content (*, P < 0.05; ns, not significant by 1-way ANOVA).

Figure 2

Chronic high glucose impairs CREB phosphorylation activation. The phosphorylation of the cAMP-response element binding protein (pCREB) in lysates of MIN6 cells in response to high glucose + exendin-4 (HG + Ex) for 15 min was determined by immunoblotting. MIN6 cells were pre-cultured chronically (20 h) either at low or high glucose. Immunoblotting was controlled by blotting for total CREB (tCREB) and tubulin. (A). MIN6 cells were chronically maintained at low or high glucose then treated for 15 min with low glucose (LG), high glucose (HG) or high glucose in the presence of the GLP-1R agonist exendin-4 (HG + Ex). (B). MIN6 cells were switched from low glucose to high glucose for the time indicated (0–20 h) to determine the time-course of the loss of the CREB phosphorylation response. (C). MIN6 cells were switched from high to low glucose for the time indicated (0–20 h) to determine the time-course of the restoration of the CREB phosphorylation response.

Figure 3

Chronic hyperglycemia reduces the cAMP response to exendin-4. The generation of cAMP was studied using a FRET-based cAMP reporter in isolated islets (A) and MIN6 cells (B) or by cAMP ELISA in MIN6 cells (C). Changes in cAMP were measured using the Epac-cAMP FRET biosensor during a 5 min low glucose period after which high glucose with exendin-4 was added to the culture chamber (A, B). Change in FRET ratio values in the low glucose versus high glucose plus exendin-4 periods were expressed relative to time = 0 values. Changes in Epac-cAMP FRET following high glucose plus exendin-4 addition were analyzed by t-tests (data expressed as mean ± SD; P < 0.0001, n = 7–9). The levels of cAMP were measured using a cAMP ELISA in MIN6 cells pre-cultured for 20 h at low or high glucose. The levels of cAMP were measured 10 min after addition of high glucose (HG) or HG with exendin-4 (HG + E). Data was analyzed by 1-way ANOVA; with *** indicating P < 0.001. n = 5–14).

Figure 4

Hyperglycemia down-regulates GLP-1R signaling. MIN6 cells cultured chronically at low or high glucose were analyzed for effects of hyperglycemia upon GLP-1R signaling systems. RNA was prepared for qRT-PCR analysis (A, B, D, G) or protein lysates were prepared for PKA activity assays (C) and immunoblotting (E). RNA expression of adenylyl cyclase 8 (AC8; A), the inducible cAMP early repressor (ICER; B), the PKA subunits (PKA-Cα, RIα, RIIα, and RIIβ; D) and the GLP-1R (G) was determined by qRT-PCR. Lysates prepared from MIN6 cells cultured at low glucose (LG; white bars) or high glucose (HG; black bars) were analyzed for: (C) PKA activity in the presence of 100 nm added cAMP and (D) for expression of the PKA subunits PKA-Cα, RIα, RIIα and RIIβ by immunoblotting, with β-tubulin presented as a loading control. (F) MIN6 cells chronically cultured at low or high glucose were fixed in 4% PFA, and the binding of fluorescein-tagged exendin-4 quantified using an anti-fluorescein-HRP antibody. Binding was quantified as the integrated density of HRP activity (Student's t-test, P < 0.0001, n = 4).

Figure 5

PKA activity correlates with loss of GLP-1R from the cell surface. MIN6 cells were transfected with a GFP-tagged GLP-1R (GLP-1R-GFP) expressed under the control of the constitutive CMV promoter. (A) GLP-1R-GFP transfected MIN6 cells were cultured at low glucose (3 mM; LG) or high glucose (25 mM; HG) for 4 h in the presence or the absence of the PKA inhibitor, H89 (HG + H89). GFP was visualized in green, nuclei stained with dapi (blue) and β-catenin immune-stained to mark the plasma membrane (red). (B) GLP-1R-GFP transfected MIN6 cells were infected with a recombinant adenovirus expressing a constitutively active PKA catalytic subunit (caPKA). Cells were cultured at low glucose (LG), and the localization of the GLP-1R-GFP was determined by fluorescence microscopy. Cells infected with the caPKA adenovirus were identified using an antibody against the FLAG-tag of the caPKA (red). (C) GLP-1R-GFP transfected MIN6 cells were cultured for 4 h at high glucose (HG) with: forskolin (Fsk), to raise cAMP levels through the activation of adenylyl cyclases; H89, to inhibit PKA; or expression of a dominantly negative PKA regulatory subunit (dnPKA). GLP-1-GFP was visualized in green, nuclei by staining with dapi (blue), and the FLAG-tag epitope on the dnPKA in red.

Figure 6

GLP-1R serine 301 mediates the glucose-dependent membrane downregulation. (A) MIN6 cells expressing wild type GLP-1R-GFP or the mutant gfp-tagged receptors S193A, S301A, or S430A, were chronically cultured at low glucose (LG) or high glucose (HG) before cell surface proteins were biotinylated and purified by streptavidin affinity. Biotinylated (membrane) and non-biotinylated (cytosol) proteins were immunoblotted for GFP and quantified by densitometry (A. t-test, P = 0.002, n = 3). (B) MIN6 cells infected with either the wild type GLP-1R-GFP or the S301A mutated GLP-1R-GFP were chronically maintained at high glucose and GLP-1R-GFP localization determined by fluorescence microscopy. (C) Determination of GLP-1R/SUMO-1 interaction using FRET of CFP-tagged GLP-1R (GLP-1R-GFP) and YFP-tagged SUMO-1 (SUMO-YFP); Blue, MIN6 cells expressing wild type GLP-1R-GFP and wild type SUMO-1-mCherry; Red, MIN6 cells expressing wild type GLP-1R-GFP and SUMO-GG mutant SUMO-1-mCherry; Green, MIN6 cells expressing S301A mutant GLP-1R-GFP and wild type SUMO-1-mCherry. Data collected from 8 to 10 cells on multiple plates. (D) Fluorescence microscopy image of MIN6 cells cultured at low glucose showing localization of GLP-1R-GFP (wild type or S301A mutant, green) with the expression of SUMO-1-mCherry (wild type or SUMO-GG mutant, red).

Figure 7

Chronic exposure of β-cells to the GLP-1R agonist exendin-4. (A) MIN6 cells were chronically cultured (20 h) at low glucose in the absence (LG) or the presence (LG + Ex4) of the GLP-1R agonist exendin-4, fixed and incubated with fluorescein-tagged exendin-4 to bind cell surface (but not internal) GLP-1R. (B) MIN6 cells transfected with GLP-1R-GFP were cultured for 4 h at high glucose with exendin-4 in the absence (HG + Ex) or the presence of H89 (HG + Ex + H89), and GLP-1R-GFP localization was determined by fluorescence microscopy (green). Cells were stained for β-catenin to mark membranes (red) and dapi to identify nuclei (blue). (C) To determine the in vivo effects of chronic exendin-4 exposure, male mice were administered exendin-4 at 4–6 h intervals for 24 h or were administered saline at each time-point (controls). Mice were then given a 5 μg/kg body exendin-4 dose and an hour later a 3 g/kg intraperitoneal glucose bolus. Plasma insulin levels were measured for 15 min following the glucose challenge (C) and blood glucose levels for 120 min (D). Data were analyzed by 2-way ANOVA with Bonferroni post hoc tests. *, P < 0.05; ***, P < 0.001. n = 6–8 mice for both C and D).