Glucagon-Like Peptide 1 Protects Pancreatic β-Cells From Death by Increasing Autophagic Flux and Restoring Lysosomal Function

Macroautophagy (hereafter referred to as autophagy) is the principle lysosomal-mediated mechanism for the degradation of misfolded proteins and damaged organelles to sustain intracellular homeostasis and core metabolic functions (1). Cytosolic components are engulfed by autophagosomes that fuse with lysosomes to allow for degradation of contents by lysosomal hydrolases. This process generates amino acids, lipids, and nucleotides for biosynthesis or for use as energy sources during starvation and can also serve to remove superfluous nutrients in metabolically active tissues. Under most conditions, autophagy acts to prevent or delay cell death, but can result in cell death through deregulation or exacerbation. Lysosomal function and fusion with autophagosomes are critical mediators of autophagic flux and its prosurvival versus prodeath role (1–3).

Recent studies have implicated a role for β-cell autophagy in the development of type 2 diabetes (4). The overt β-cell failure evident in type 2 diabetes (5) is a consequence of both deterioration of function and loss of cell mass (6) and is in part attributed to the cytotoxic effects of chronically elevated glucose and fatty acids (glucolipotoxicity [GLT]) (7). The impact of gluco-/lipotoxicity on β-cell autophagy has been proposed to be both beneficial and detrimental to cell survival. Autophagy-deficient β-cells show compromised function and survival upon high-fat feeding, supporting a protective role for autophagy (8), a finding supported by in vitro studies of gluco-/lipotoxicity (9–13). In contrast, other studies report that chronic lipotoxicity causes defective autophagic turnover, leading to cell death (14–16), consistent with increased autophagosome accumulation in human and rodent type 2 diabetes (17,18). Such studies have suggested that the impairment in autophagic flux may be secondary to defects in lysosomal function (14,16), but the underlying mechanisms have not been elucidated, and the impact of this defect on lysosomal-mediated cell death has not been explored.

Glucagon-like peptide 1 (GLP-1) receptor agonists lower blood glucose through distinct actions on various organs including β-cells (19). GLP-1 agonists enhance insulin secretory capacity by increasing glucose-induced insulin secretion and preserving β-cell mass by stimulation of β-cell proliferation and inhibition of cell death (20). Studies in other tissues, including liver and spinal cord, have shown that GLP-1 agonists activate autophagy and that this mechanism is essential for the cell-survival effects of the drug (21,22). However, the precise role of GLP-1 in regulating β-cell autophagy and its impact on cell fate has not been defined. To this aim, we have explored the role for autophagy in the prosurvival effects of GLP-1 in β-cells exposed to GLT. The results provide a mechanistic insight into the deregulation of autophagy and lysosomal dysfunction in pancreatic β-cells exposed to excess nutrients and identify restoration of these pathways as critical components for the prosurvival effect of the GLP-1 receptor agonist exendin-4.

Eight- to 10-week-old C57BL/6 mice fed ad libitum were euthanized by cervical dislocation. All procedures conformed to Home Office Regulations and approved by the Newcastle University Ethical Committee. Pancreases were perfused with collagenase-P via the common bile duct and islets purified using a Histopaque gradient (23) with a final filtration step using a 70-µm filter. Islets were cultured overnight in RPMI 1640 supplemented with 10% FBS, 50 U/mL penicillin, and 50 μg/mL streptomycin prior to treatment.

Human islets were isolated from seven donors without diabetes at the Clinical Islet Laboratory, University of Alberta, Canada, or the Islet Isolation Unit, Kings College London, U.K., with appropriate ethical approval. Islets were transported to Newcastle and maintained in CMRL media supplemented with 0.5% human albumin serum, 50 U/mL penicillin, and 50 μg/mL streptomycin for 24 h prior to treatment.

Pancreatic tissue/islets from eight donors (three patients with type 2 diabetes: aged 59.0 ± 6.6 years; BMI 26.8 ± 4.3 kg/m²; metformin: two patients; five patients without diabetes: aged 42.9 ± 14.2 years; BMI 27.1 ± 4.1 kg/m²) were studied. Ethical approval was acquired from Newcastle and North Tyneside ethics committee and research consent obtained. After fixation in formalin and embedding in paraffin, indirect immunofluorescence staining was carried out as in White et al. (24). Tissue was imaged using an Axio Imager II microscope (Carl Zeiss), and a minimum of five islets from one section was imaged per patient. The p62 signal intensity of the islets was quantified using ImageJ (National Institutes of Health) and nonislet staining subtracted. Data are expressed as p62 intensity per islet.

INS-1E cells were cultured in RPMI 1640 media containing 50 μmol/L β-mercaptoethanol, 1 mmol/L sodium pyruvate, 50 U/mL penicillin, 50 μg/mL streptomycin, and 5% FBS. Palmitate/BSA was prepared as in Choi et al. (11). For Atg5 silencing, INS-1E cells were transfected with scrambled small interfering RNA (siRNA) or siRNA targeted against Atg5 for 48 h using Lipofectamine RNAiMAX (Thermo Fisher Scientific, Loughborough, U.K.) according to the manufacturer’s instructions.

INS-1E cells were transfected with pBABE-puro mCherry EGFP-LC3B (25) (a gift from Jayanta Debnath; Addgene plasmid 22418) and cells selected using 1 µg/mL puromycin. INS-1E(mCherry-GFP-LC3) cells were plated on chambered coverslips and nuclei stained with 4.5 µg/mL Hoechst for 5 min. Live cell imaging was performed in a heated environmental chamber (37°C, 5% CO₂) using either a TE2000 (×100; Nikon) or an A1R (×100; Nikon). Images were processed using NIS Element Imaging software and puncta quantified using Volocity.

INS-1E cells were transfected with EGFP-LC3 (26) (gift from Karla Kirkegaard; Addgene plasmid 11546) and cells selected using 1 µg/mL puromycin. INS-1E(GFP-LC3) were plated onto coverslips and costained with 75 nmol/L LysoTracker Red DND-99 (Thermo Fisher Scientific) for 30 min before fixation with 4% paraformaldehyde. Nuclei were stained with 4.5 µg/mL Hoechst 33342 and coverslips mounted onto slides. Cells were imaged using an Eclipse E400 (×60; Nikon). Colocalization between GFP-LC3 and lysosomes was visually quantified as the number of GFP-LC3 puncta costaining for LysoTracker. This analysis was performed by two independent observers, one unaware of sample identity.

For propidium iodide (PI)/Hoechst staining, INS-1E/human islets were incubated with 10 µg/mL PI and 10 µg/mL Hoechst-33342 for 20 min. Images were taken with a TE2000 (×20; Nikon). For INS-1E, >1,000 cells were counted per condition and analysis performed using ImageJ software. For human islets, a minimum of 10 islets was analyzed and percent viability estimated visually by two independent observers, one unaware of sample identity. Apoptosis was determined using the Caspase-Glo 3/7 Assay (Promega, Southampton, U.K.) or by Western blotting for cleaved caspase-3.

Proteins were fractionated using 4–12% SDS-PAGE gels (Bio-Rad, Hertfordshire, U.K.) and transferred to polyvinylidene difluoride. Membranes were probed with primary antibody (Supplementary Table 1) at 4°C overnight. After incubation with secondary antibody conjugated to horseradish peroxidase, bands were detected using enhanced chemiluminescence. Immunoblots were quantified using ImageJ.

Cells plated onto coverslips were fixed with 4% paraformaldehyde and immunostaining performed as in Stubbs et al. (27). Cells were imaged using a TE2000 (×100; Nikon). Lysosomal puncta were quantified using Blobfinder software (Uppsala University, Uppsala, Sweden). Transcription factor EB (TFEB) translocation was quantified by measurement of the mean pixel intensity for the nuclear and cytoplasmic compartments using ImageJ and the nuclear/cytoplasmic (N/C) ratio calculated.

For cathepsin activity, INS-1E plated in 24-well plates was extracted into 200 µL cell lysis buffer and cathepsin B/D detected according to the manufacturer’s instructions (CBA001, Merck Millipore, Billerica, MA; and 68AT-CathD-S100, RayBiotech, Norcross, GA), respectively).

RNA was extracted using TRIzol (Thermo Fisher Scientific), and cDNA was synthesized from 1 μg RNA with Maloney murine leukemia virus (Promega). Real-time RT-PCR was performed as in Arden et al. (28) using custom-designed primers (Supplementary Table 2). Relative mRNA levels were calculated by delta cycle threshold, corrected for cyclophilin A, and expressed relative to control.

For analysis of TFEB translocation, nuclear and cytosolic extracts were prepared as in Suzuki et al. (29). For analysis of lysosomal membrane permeabilization (LMP), cytosolic extracts were prepared as in Marques et al. (30).

Results are expressed as means ± SEM for the number of cell preparations and values compared using either the Student t test (paired or unpaired) or one- or two-way ANOVA followed by Bonferroni test.

We first confirmed the protective effects of the GLP-1 agonist exendin-4 over β-cell death. Treatment of INS-1E with GLT (25 mmol/L glucose and 0.5 mmol/L palmitate) increased cell death (Fig. 1A) and apoptosis (Fig. 1B), which were partially prevented by cotreatment with 100 nmol/L exendin-4 (Fig. 1A and B). GLT also increased cell death and apoptosis in human islets, and exendin-4 partially prevented this (Fig. 1C–E). GLT increased basal autophagic activity, as evidenced by increased LC3 II in INS-1E and in mouse and human islets (Fig. 1F–H). Exendin-4 had no effect in the absence of GLT but exacerbated the GLT-induced increase in LC3 II (Fig. 1F–H).

To investigate autophagic flux, we evaluated LC3 II in the presence of the lysosomal inhibitor chloroquine (31). As expected, chloroquine increased LC3 II (Fig. 2A). GLT continued to increase LC3 II in the presence of chloroquine, suggesting autophagic stimulation. However, the effect of GLT was greater in the absence of chloroquine than in its presence (GLT without chloroquine, 4.8-fold ± 0.23 increase vs. GLT plus chloroquine, 2.2-fold ± 0.46 increase in LC3 II/GAPDH; P < 0.005), suggesting partial blockage of autophagic flux.

To further explore this, we used two more robust methods to assess autophagic flux: mCherry-GFP-LC3 analysis and p62 accumulation (31). We first generated INS-1E stably expressing mCherry-GFP-LC3 (32). Validation of this system confirmed that blockage of autophagic flux with chloroquine caused accumulation of yellow puncta (Fig. 2B). Exposure of INS-1E(mCherry-GFP-LC3) to GLT also increased yellow puncta (Fig. 2C), indicating impairment in flux, whereas treatment with exendin-4 decreased the number of yellow puncta (Fig. 2C), indicating restoration of flux. These observations were further supported by observations of GLT-induced p62 accumulation (Fig. 2D), consistent with the increase in p62 in tissues from patients with type 2 diabetes (Fig. 2E), indicating impaired autophagic flux (30). Cotreatment of INS-1E with exendin-4 decreased p62 accumulation (Fig. 2D). Collectively, these data demonstrate that the GLT-induced increase in LC3 II represents both autophagic stimulation and also an impairment in flux. Cotreatment with exendin-4 prevents the impairment in autophagic flux and further stimulates autophagy.

To determine whether autophagy contributes to the prosurvival effects of exendin-4, we first explored the impact of impaired autophagic flux on cell survival. Live cell imaging of INS-1E(mCherry-GFP-LC3) treated with GLT showed that cells with apparent autophagic dysfunction, as evident by large yellow puncta (at 4–6 h), showed evidence of cell death (cell rounding, detachment, and loss of fluorescence) at later time points (Fig. 3A, arrowheads, and Supplementary Video 1). Further assessment using siRNA knockdown of the key autophagic gene Atg5 showed that silencing of Atg5 (Fig. 3B) partially prevented the exendin-4–induced increase in LC3 II (Fig. 3B and C). Silencing of autophagy did not prevent the GLT-induced increase in cell death but did reverse the protective effect of exendin-4, as evidenced by the increase in PI staining (Fig. 3D). The impact of exendin-4 on autophagy is therefore essential for its protective role in β-cell survival.

The mechanisms for the impairment in autophagic flux in response to GLT are not well understood, although a role for lysosomal dysfunction has been proposed (14,16). To understand how exendin-4 prevents the deregulation of autophagic flux, lysosomal function was assessed. Treatment of INS-1E with GLT decreased lysosomal staining as determined using LysoTracker dye and immunostaining for Lamp2, which was partially prevented by cotreatment with exendin-4 (Fig. 4A and B). GLT also decreased cathepsin B activity, and exendin-4 reversed this (Fig. 4C). These effects were not due to changes in protein expression (Supplementary Fig. 1A and B). The decrease in lysosomal staining was accompanied by an increase in lysosomal vesicle size, an effect partially reversed by cotreatment with exendin-4 (Fig. 4D). To assess whether the enlarged vesicles represent lysosomes or autolysosomes, lysosomes were stained with LysoTracker and autophagosomes with GFP-LC3 and the accumulation of these vesicles monitored upon treatment with GLT for 4–8 h. After 4-h treatment, GFP-LC3–containing autophagosomes increased (Fig. 4E). Colocalization of lysosomal puncta (red) with autophagosomes (green) increased after 4-h treatment with GLT but decreased at later time points (Fig. 4E and F), suggesting loss of autophagosome–lysosomal fusion. Cotreatment with exendin-4 maintained colocalization between lysosomes and autophagosomes, consistent with restored autophagic flux (Fig. 4E and F).

To determine whether the changes in lysosomal function could be explained by alterations in lysosomal biogenesis, we investigated the regulation of TFEB. TFEB upregulates expression of genes involved in lysosomal biogenesis and function, autophagosome biogenesis, and autophagosome–lysosome fusion (33,34). Inactive in the cytoplasm under basal conditions, TFEB translocates to the nucleus and activates gene transcription upon stimulation with starvation or lysosomal stress (34). Immunostaining for endogenous TFEB showed that under basal conditions, TFEB localized to the cytoplasm (Fig. 5A). Treatment with GLT caused translocation of TFEB to the nucleus (Fig. 5A), quantified as a twofold increase in the N/C ratio (Fig. 5B). Further analysis using subcellular fractionation confirmed the nuclear translocation of TFEB (Fig. 5C). Real-time PCR analysis showed increased mRNA expression of a number of TFEB downstream targets, including TFEB, MCOLN1, BECN1, and UVRAG (Fig. 5D), confirming that GLT promotes lysosomal/autophagosomal biogenesis. Costimulation with exendin-4 further exacerbated the GLT-induced translocation of TFEB as determined by the N/C ratio (Fig. 5A and B), subcellular fractionation (Fig. 5C), and also increased downstream gene transcription (Fig. 5D), suggesting that upregulation of lysosomal/autophagosomal biogenesis may contribute to the beneficial effects of exendin-4.

The accumulation of defective lysosomes with their high concentration of hydrolytic enzymes is potentially hazardous for the cell. Release of these components into the cytoplasm via LMP can initiate cell death (35). To determine whether the accumulation of defective lysosomes directly contributes to β-cell death, LMP was investigated. Staining of INS-1E for cathepsin B and D showed loss of punctate staining after treatment with GLT, which was prevented by cotreatment with exendin-4 (Fig. 6A and B). Such a staining pattern in the absence of changes in protein expression (Supplementary Fig. 1B and C) can be indicative of LMP. Further assessment by isolation of cytoplasmic fractions showed that exposure to GLT caused release of cathepsin D into the cytoplasm, which was prevented by cotreatment with exendin-4 (Fig. 6C). The impact of LMP on total cell death was assessed using cathepsin inhibitors. Cotreatment with antipain, a nonspecific inhibitor, partially prevented the GLT-induced increase in β-cell death (Fig. 6D). Further analysis using specific inhibitors of cathepsin B (CBi; CA-074) and cathepsin D (CDi; Pepstatin A) (Supplementary Fig. 1D and E) showed that whereas CBi had no effect (Fig. 6E), CDi prevented the GLT-induced increase in cell death (Fig. 6F). This was confirmed in murine islets in which CDi partially reversed the GLT-induced increase in apoptosis (Fig. 6G), supporting a role for cathepsin D in LMP-induced cell death.

To further explore this mechanism, tissue samples from control subjects and patients with type 2 diabetes were stained for cathepsin D. When compared with control subjects, diabetic tissue showed decreased staining of cathepsin D in insulin-positive cells (Fig. 6H and Supplementary Fig. 2). Although a decrease in cathepsin D expression cannot be excluded, higher-magnification images (Fig. 6I) indicate loss of punctate staining, similar to that displayed in GLT-treated INS-1E.

To elucidate the mechanisms underlying GLT-induced lysosomal dysfunction and its reversal by exendin-4, we investigated whether the lysosomal dysfunction was dependent on autophagy. Autophagy was chemically modulated using 3-methyladenine (3MA). Despite inhibition of LC3 II formation in response to GLT (Fig. 7A), there was no effect of 3MA on the GLT-dependent decrease in lysosomal staining or its reversal by exendin-4 (Fig. 7B), suggesting that these changes are independent of autophagy.

Endoplasmic reticulum (ER) stress is a major driver in GLT-induced cell death (36). To determine whether ER stress also plays a role in GLT-induced lysosomal dysfunction, ER stress was assessed. GLT increased ER stress, as evident by an increase in phosphorylated eukaryotic translation initiation factor 2a (p-eIF2a) and CHOP (Fig. 7C and D). Blockage of ER stress with the inhibitor tauroursodeoxycholic acid (Fig. 7F) partially prevented the increase in LC3 II induced by GLT (Fig. 7G) and also prevented the decrease in lysosomal staining (Fig. 7H) and the GLT-induced increase in cell death (Fig. 7I). These data support a model by which ER stress stimulates autophagy but also impairs lysosomal function, preventing fusion of lysosomes with autophagosomes, leading to blockage of autophagic flux. Consistent with previous studies (37,38), the GLT-induced increase in ER stress was attenuated by coincubation with exendin-4 (Fig. 7E). This attenuation of ER stress by exendin-4 may underlie the improvement in lysosomal function.

We next sought to elucidate the signaling pathways linking ER stress with autophagy and lysosomal dysfunction. Previous studies have linked ER stress to autophagy via JNK signaling (39), but whether JNK regulates lysosomal function is unknown.

GLT increased JNK phosphorylation (Fig. 7J). Inhibition of JNK-P using SP600125 prevented the GLT-induced increase in LC3 II (Fig. 7J and K), but not GLT-induced decrease in lysosomal staining (Fig. 7L). These data show that although JNK signaling mediates the downstream ER stress mechanisms linking to autophagy, it does not mediate the lysosomal dysfunction induced by GLT. There was no effect of exendin-4 on JNK phosphorylation, and JNK inhibition did not prevent exendin-4–induced LC3 II nor attenuation of lysosomal dysfunction (Fig. 7J–L), suggesting that exendin-4 does not mediate its protective effects via JNK.

GLP-1 agonists improve β-cell function and increase β-cell mass through upregulation of proliferation and inhibition of cell death (19,20), although the underlying mechanisms remain unknown. In the current study, we have shown that regulation of autophagic flux is a critical component in the prosurvival effects of exendin-4. Our data supports a model in which GLT induces lysosomal dysfunction, leading to deregulation of autophagy and resulting lysosomal cell death. Treatment with exendin-4 reverses the lysosomal dysfunction, relieving the impairment in autophagic flux, and improves cell survival by stimulating autophagy.

Whether the regulation of β-cell autophagy by gluco-/lipotoxicity represents a prosurvival or prodeath mechanism is contentious (8–18). Some studies support a protective role for autophagy in regulating β-cell mass in response to excess nutrients (8–13), whereas others suggest that these stimuli impair autophagic flux, leading to cell death (14–18). Our data show that GLT stimulates β-cell autophagy but that it also impairs flux through the induction of lysosomal dysfunction, as indicated by decreased lysosomal staining and decreased cathepsin activity, consistent with previous models of β-cell lipotoxicity (14,16). This does not appear to be caused by an impairment in lysosomal biogenesis, because GLT causes nuclear translocation of TFEB and upregulation of downstream gene transcription. The enlarged lysosomal vesicles are distinct from the accumulated autophagosomes, as evident from the loss of overlap between lysosomal stains and autophagosomal GFP-LC3, indicating loss of lysosomal–autophagosomal fusion at later time points. Chemical inhibition of autophagy and ER stress show that the defect in lysosomal dysfunction is independent from autophagy but downstream of the GLT-induced increase in ER stress. ER stress also appears to be a major driver in the increase in autophagic signaling, consistent with a previous study (39). However, further analysis indicate differing downstream pathways regulating autophagy and lysosomal dysfunction, JNK dependent and independent, respectively. These findings support a model by which GLT induces ER stress, which both stimulates autophagy and induces lysosomal dysfunction. We predict that the increased activation of autophagy in the presence of impaired lysosomal fusion leads to deregulation of autophagic flux and accumulation of autophagosomal and lysosomal vesicles (Fig. 8).

The accumulation of autophagic vesicles in tissue from patients with type 2 diabetes by both p62 accumulation (current study and 17) and ultrastructure analysis (18) suggests that the impairment in autophagic flux may contribute to the loss of β-cell mass in type 2 diabetes. This hypothesis is supported by our mCherry-GFP-LC3 data showing that after exposure to GLT, cells with impaired autophagic flux appear to undergo cell death at later time points, unlike cells with normal flux, which are more likely to survive. Furthermore, the improvement in β-cell survival induced by exendin-4 was dependent on restoration of autophagic signaling. This is similar to that reported for the autophagy activators rapamycin and carbamazepine, which improved β-cell survival under lipotoxic conditions (11,13,15). Studies in other cell types have shown that an impairment in autophagic flux can lead to cell death via apoptosis, nonapoptotic mechanisms, and also lysosomal-mediated cell death (35). Alterations in lysosomal enzymes/protein expression have been reported in both human and rodent type 2 diabetes (18,40), but their contribution to cell death has not been explored. The current study provides strong evidence that the defect in lysosomal function contributes directly to cell death via LMP (Fig. 8). Loss of punctate cathepsin staining without alterations in protein expression suggest release of cathepsin from the lysosomes, a finding supported by detection of cathepsin D in cytoplasmic fractions. This loss in staining pattern was also evident in tissue from patients with type 2 diabetes compared with control subjects, although a decrease in cathepsin D protein expression cannot be excluded. Treatment with various cathepsin inhibitors indicate that release of cathepsin D from these cells appears to drive GLT-induced cell death. These findings are consistent with recent studies showing lysosomal dysfunction and LMP in the renal tubule and neuronal cortex when exposed to elevated lipid concentrations (41,42). The finding that human type 2 diabetic tissue shows defects in both autophagic flux (p62 accumulation) and lysosomal function (cathepsin D staining) strongly supports a role for lysosomal/autophagic-induced cell death as a major driver of β-cell death and dysfunction in type 2 diabetes.

GLP-1 agonists improve the insulin secretory response by increasing β-cell mass in animal models of type 2 diabetes, which is due, in part, to a decrease in β-cell death (19,20). Previous studies investigating the impact of GLP-1 agonists on β-cell autophagy have shown that GLP-1 increases LC3 II levels and autophagosomal formation in conditions of high-fat diet/lipotoxicity (43,44) but that it inhibits autophagic stimulation in a model of tacrolimus-induced diabetes (45). Through the rigorous assessment of autophagic stimulation and flux, we show that GLP-1 stimulated flux through autophagy. This was evident by an increase in LC3 II formation accompanied by a decrease in autophagosomal accumulation as assessed by mCherry-GFP-LC3 and p62. This is consistent with previous models of nutrient excess (43,44) but not tacrolimus-induced diabetes (45), suggesting that the underlying mechanism of autophagic dysfunction may determine the impact of GLP-1 agonists. There was no effect of exendin-4 on autophagy in the absence of GLT, in agreement with a previous study (17). siRNA knockdown of Atg5 prevented the exendin-4–induced increase in LC3 II and prevented the prosurvival effect of exendin-4, suggesting that activation of autophagy is a critical component of exendin-4–mediated cell survival, although additional pathways impacting on cell survival cannot be excluded. Further assessment of autophagic flux showed that exendin-4 prevents the GLT-induced defects in lysosomal function. This is consistent with the protective effects of exendin-4 in tacrolimus-induced diabetes (45), highlighting the importance in restoring lysosomal function in the prosurvival effects of the drug. Through its ability to restore lysosomal function, exendin-4 also prevents LMP and further stimulates lysosome biogenesis. Restoration of lysosomal function accompanied by activation of autophagy will promote removal of the accumulated autophagosomes and lysosomes, thus improving cell survival. Given that the GLT-induced increase in ER stress mediates lysosomal dysfunction in our β-cell model, it is likely that the resolution of ER stress by exendin-4 leads to an improvement in lysosomal function and accordingly restoration of autophagic flux.

In conclusion, the results of our study show that lysosomal dysfunction mediates the loss of β-cell mass in response to GLT. We show that treatment with the GLP-1 agonist exendin-4 can restore autophagic flux and rescue lysosomal function. These findings support exploration of a potential therapeutic role for autophagy and lysosomal modulation for the treatment and/or prevention of type 2 diabetes and also for the further elucidation of the GLP-1 signaling mechanisms impacting on this pathway.

Acknowledgments. The authors thank Dr. Helen Marshall (Institute of Cellular Medicine, Newcastle University) for technical assistance with the rodent islet isolations and Philip Home (Institute of Cellular Medicine, Newcastle University) for additional funding support.

Funding. The authors thank Dr. Claus Wollheim (University Medical Center, Geneva, Switzerland) for the gift of INS-1E cells and the Clinical Islet Laboratory, University of Alberta, Canada, and the Islet Isolation Unit, Kings College, London, U.K., for provision of the human islets. This work was supported by project grants from Diabetes UK (12/0004544) and the Diabetes Research & Wellness Foundation (SCA/OF/11/12).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. F.P.Z. acquired the data, performed data analysis, and contributed to the writing of the manuscript. K.S.C. acquired the data and performed data analysis. M.H.-S. assisted with rodent islet isolation and human islet preparation. J.A.M.S. contributed to the concept and study design. P.E.L. contributed to the concept and study design and provided constructs. C.A. developed the concept and study design, performed data interpretation, and wrote the manuscript. C.A. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in poster form at the 51st Annual Meeting of the European Association for the Study of Diabetes, Stockholm, Sweden, 14–18 September 2015, and the Diabetes UK Professional Conference, Glasgow, Scotland, U.K., 2–4 March 2016.