Opposing effects on regulated insulin secretion of acute vs chronic stimulation of AMP-activated protein kinase

Aims/hypothesis Although targeted in extrapancreatic tissues by several drugs used to treat type 2 diabetes, the role of AMP-activated protein kinase (AMPK) in the control of insulin secretion is still debatable. Previous studies have used pharmacological activators of limited selectivity and specificity, and none has examined in primary pancreatic beta cells the actions of the latest generation of highly potent and specific activators that act via the allosteric drug and metabolite (ADaM) site. Methods AMPK was activated acutely in islets isolated from C57BL6/J mice, and in an EndoC-βH3 cell line, using three structurally distinct ADaM site activators (991, PF-06409577 and RA089), with varying selectivity for β1- vs β2-containing complexes. Mouse lines expressing a gain-of-function mutation in the γ1 AMPK subunit (D316a) were generated to examine the effects of chronic AMPK stimulation in the whole body, or selectively in the beta cell. Results Acute (1.5 h) treatment of wild-type mouse islets with 991, PF-06409577 or RA089 robustly stimulated insulin secretion at high glucose concentrations (p<0.01, p<0.05 and p<0.001, respectively), despite a lowering of glucose-induced intracellular free Ca2+ dynamics in response to 991 (AUC, p<0.05) and to RA089 at the highest dose (25 μmol/l) at 5.59 min (p<0.05). Although abolished in the absence of AMPK, the effects of 991 were observed in the absence of the upstream kinase, liver kinase B1, further implicating ‘amplifying’ pathways. In marked contrast, chronic activation of AMPK, either globally or selectively in the beta cell, achieved using a gain-of-function mutant, impaired insulin release in vivo (p<0.05 at 15 min following i.p. injection of 3 mmol/l glucose) and in vitro (p<0.01 following incubation of islets with 17 mmol/l glucose), and lowered glucose tolerance (p<0.001). Conclusions/interpretation AMPK activation exerts complex, time-dependent effects on insulin secretion. These observations should inform the design and future clinical use of AMPK modulators. Graphical abstract Supplementary Information The online version contains peer-reviewed but unedited supplementary material available at 10.1007/s00125-022-05673-x.

Islets were isolated by digestion with collagenase as described [7]. In brief, pancreata were inflated with a solution of collagenase from clostridium histolyticum (1 mg/mL; Nordmark, Germany) and placed in a water bath at 37 ⁰C for 12 min. Islets were washed and purified on a Histopaque gradient (Sigma-Aldrich, UK). Isolated islets were cultured for 24 h in RPMI 1640 containing 11.1 mmol/l glucose, 10% foetal bovine serum and L-glutamine (Sigma-Aldrich, UK) and allowed to recover overnight.

Insulin secretion
Insulin secretion assays on isolated mouse islets were performed as previously described [7].
Secreted and total insulin were quantified using a HTRF insulin kit (Cisbio, France) in a PHERAstar reader (BMG Labtech, UK) following the manufacturer's guidelines.

Intracellular free calcium and cytosolic ATP/ADP imaging
Whole isolated islets were incubated with Cal520 (Aatbio, USA) for intracellular calcium measurement [8] or with an adenoviral Perceval sensor construct [9] for measurements changes in ATP/ADP in response to 17 mmol/L glucose, following 1 hr preincubation at 3 mmol/L glucose. Fluorescence imaging was performed using a Nipkow spinning disk head (Yokogawa CSU-10; Runcorn, UK), allowing rapid scanning of islet areas for prolonged periods of time with minimal phototoxicity. Volocity software (PerkinElmer Life Sciences, UK) provided interface while islets were kept at 37⁰C and constantly perifused with KHB containing 3 mmol/L or 17 mmol/L glucose or 30 mmol/L KCl (for intracellular calcium measurements only).

Protein isolation and Western immunoblotting
Isolated whole islets proteins were extracted in lysis buffer (150 mmol/l NaCl, 50 mmol/l Tris-

Total internal reflection of fluorescence (TIRF) and spinning disc confocal imaging
Mouse islets were dissociated after being washed with PBS using accutase at 37⁰C for 5 minutes. Cells were then separated by gently pipetting up and down. Cells were then pelleted, resuspended in normal culture media and left to attach on a glass slide treated with polylysine. Imaging was performed as described previously [10] using a Nikon Eclipse Ti microscope equipped with a ×100/1.49 numerical aperture (NA) TIRF objective, a TIRF iLas2 module to control laser angle (Roper Scientific), and a Quad Band TIRF filter cube (TRF89902, Chroma).
Acquisitions were performed using a 488-nm laser line, and images were captured with an ORCA-Flash 4.0 camera (Hamamatsu, Japan), both in TIRF mode and widefield mode.
Metamorph software (Molecular Devices, USA) was used for data capture and the laser angle was selected for an imaged section thickness of 150-180 nm.
For insulin secretion assay, cells were seeded onto ECM/Fibronectin-coated 96-well plates at 7 x 10 4 cells per well. Two days after seeding, cells were incubated overnight in a glucose starving medium (glucose-free DMEM supplemented with 2 % Albumin from bovine serum fraction V, 50 ml b-mercaptoethanol, 10 mM nicotinamide, 5.5 mg/ml transferrin, 6.7 ng/ml sodium selenite, 100 units/ml, penicillin, 100 mg/ml streptomycin and 2.8 mM glucose). The was then removed to a fresh tube and centrifuged at 3,000 rpm for 5 min at 4°C. Insulin content was measured using an insulin ultra-sensitive assay kit (Cisbio).

RNA isolation and quantitative PCR
RNA was isolated from epididymal and subcutaneous adipose tissue, liver and pancreatic islets with TRIzol following manufacturer's instructions (Invitrogen, UK). RNA purity and concentration were measured by spectrophotometry (Nanodrop, Thermo Scientific, UK). Only RNA with absorption ratios between 1.8-2.0 for 260/280 and 260/230nm were used. RNA integrity was checked on an agarose gel. RNA was reversed transcribed using High-Capacity cDNA Reverse Transcription kit (Applied Biosystems, UK). qPCR was performed with Fast SYBR green master mix (Applied Biosystems, UK). The comparative Ct method (2 -CT ) was used to calculate relative gene expression levels using βactin as an internal control. The primers sequences are listed in ESM Table 1.

Statistical analysis
GraphPad Prism 9.0 (www.graphpad.com) was used for statistical analyses. Significance was evaluated by unpaired Student's t-tests or by one-or two-way ANOVA, with multiple comparisons tests, as appropriate. P values of <0.05 were considered statistically significant.
Data are shown as mean ± SEM. Table 1: Primer sequences used for RT-qPCR analysis in this study.