Inhibition of Autophagic Turnover in β-Cells by Fatty Acids and Glucose Leads to Apoptotic Cell Death*

Background: Autophagy is essential for β-cell function and survival. Results: Autophagic turnover is impaired in β-cells when treated with metabolic stressors. Conclusion: Diminished autophagy leads to apoptotic β-cell death. Significance: Therapeutic interventions using pharmacological agents, which can improve ER folding capacity, as well as target the autophagy machinery, could provide promising strategies for treating human diseases such as T2D. Autophagy, a cellular recycling process responsible for turnover of cytoplasmic contents, is critical for maintenance of health. Defects in this process have been linked to diabetes. Diabetes-associated glucotoxicity/lipotoxicity contribute to impaired β-cell function and have been implicated as contributing factors to this disease. We tested the hypothesis that these two conditions affect β-cell function by modulating autophagy. We report that exposure of β-cell lines and human pancreatic islets to high levels of glucose and lipids blocks autophagic flux and leads to apoptotic cell death. EM analysis showed accumulation of autophagy intermediates (autophagosomes), with abundant engulfed cargo in palmitic acid (PA)- or glucose-treated cells, indicating suppressed autophagic turnover. EM studies also showed accumulation of damaged mitochondria, endoplasmic reticulum distention, and vacuolar changes in PA-treated cells. Pulse-chase experiments indicated decreased protein turnover in β-cells treated with PA/glucose. Expression of mTORC1, an inhibitor of autophagy, was elevated in β-cells treated with PA/glucose. mTORC1 inhibition, by treatment with rapamycin, reversed changes in autophagic flux, and cell death induced by glucose/PA. Our results indicate that nutrient toxicity-induced cell death occurs via impaired autophagy and is mediated by activation of mTORC1 in β-cells, contributing to β-cell failure in the presence of metabolic stress.

exists in two distinct complexes, mTORC1, which is rapamycin-sensitive, and mTORC2, which is insensitive to rapamycin. mTORC1 plays an important role in ␤-cell mass expansion and improved glucose tolerance (15), whereas prolonged inhibition by rapamycin causes loss of ␤-cell function and mass (16). However, recent studies have linked mTORC1 hyperactivation to insulin resistance and endoplasmic reticulum (ER) stress development resulting in decline in both ␤-cell mass and function (17,18). Type 2 diabetes (T2D) is a complex metabolic disorder characterized by a progressive decrease in ␤-cell function and overt ␤-cell mass (19). Pancreatic ␤-cells overproduce insulin to compensate for insulin resistance in the early stages of T2D but eventually become dysfunctional, leading to hyperglycemia and clinical onset of diabetes. Nutrient overload has been postulated as the main cause of deterioration of ␤-cells in T2D. Increased free fatty acids (FFAs) alone or in combination with glucose have been proposed to impair insulin secretion and trigger the loss of ␤-cells by apoptosis (20,21). Saturated fatty acids were found to be particularly cytotoxic to ␤-cells, whereas unsaturated fatty acids appear to have a protective role (22). Interestingly, fatty acids induce AP formation and suppress autophagic turnover in a rat insulinoma cell line (INS-1) (2). Increased early stage AP formation has been reported in pancreatic ␤-cells in diabetic db/db and in nondiabetic high fat-fed C57BL/6 mice, suggesting an impairment of AP maturation (23). FFAs have been hypothesized to be the underlying cause of obesity and diabetes (24). Here, human islets treated with PA resulted in impaired autophagy and decreased expression of genes related to lysosomal function that may affect lysosomeautophagosome fusion (25). However, there have been conflicting reports concerning whether FFAs induce or inhibit autophagy (23, 26 -28). Glucolipotoxicity has been shown to cause ␤-cell failure and contribute to diabetes development, but the underlying mechanism has not been definitively elucidated. Therefore, in the present study, we tested the hypothesis that PA and glucose alone or in combination inhibit autophagic turnover in pancreatic ␤-cells. We also investigated the role of metabolic challenge on ␤-cell apoptosis and function. Our results demonstrate that autophagic turnover is impaired in metabolically stressed ␤-cells, ultimately leading to apoptotic cell death. Stimulation of autophagy by treatment with an mTORC1 inhibitor alleviated the suppression of autophagic turnover, and this was accompanied with a protection from apoptotic cell death.
Human Islet Procurement and Culture-Human islets from donors (aged 20 -50 years) without a history of diabetes were obtained from Prodo Laboratories (Table 1) and, upon arrival, washed and cultured in CMRL (5.5 mM glucose; Invitrogen) ϩ 10% FBS ϩ penicillin/streptomycin and glutamine in 6-well, ultralow adherence plates (Corning) at a concentration of 1 islet equivalent per 1 l of medium. The islets were cultured (37°C in 5% CO 2 ) in 6-well, ultralow adherence plates (Corning) at a concentration of 1 islet equivalent per 1 l of medium. After 48 h, the medium was changed to culture medium containing 5.5, 16.5, and 33.3 mM glucose supplemented with or without 0.4 mM palmitic acid (PA) for 48 h. Fatty acid free BSA was used as a control.
Viability Assay-CellTiter-Glo luminescent cell viability assay was used to determine the number of viable cells in culture based on quantitation of the ATP present, which signals the presence of metabolically active cells. Briefly 96-well plate was equilibrated for 30 min at room temperature. An equal volume of CellTiter-Glo was added to each well of the plate. The plates were put on an orbital shaker for 2 min to induce cell lysis. The plates were then allowed to stand at room temperature for 10 min, and luminescence was then recorded.
Lysosomal and Immunostaining-For detection of fusion between autophagosomes and lysosomes, GFP-LC3B-MIN6 cells were stained with 50 nM LysoTracker Red (Molecular Probes) for 20 min in DMEM. The cells were then washed with PBS and immediately observed under fluorescent microscope at 540-nm excitation and 560-nm emission.
For immunohistochemistry, islets were harvested, washed, and fixed for 24 h in 4% paraformaldehyde in 1.5-ml microcentrifuge tubes. After fixation, the islets were precipitated (200 g for 3 min), aspirated, and embedded in 1.5% agarose. The agarose plugs were subsequently embedded in paraffin and sectioned at 4 m. For immunohistochemistry, sections were dewaxed and rehydrated, and these steps were followed by antigen retrieval in 25 mM Tris, 1 mM EDTA (pH 8.5). Endogenous peroxidase activity was then quenched with 3% H 2 O 2 in methanol. Following washes in PBS, sections were blocked with 3% goat serum in PBS. Primary and secondary (biotin-labeled) antibody incubations were followed by VectaStain ABC reagent (Vector Laboratories, Burlingame, CA) and then development with diaminobenzidine.
For immunofluorescence, cells were fixed in 4% PFA for 20 min, washed with PBST, and blocked with 10% goat serum in PBS for 1 h. The fixed cells were then incubated with primary antibody (in 10% goat serum) for 2 h at 4°C. The cells were then washed with PBST, incubated with secondary antibody (Invitrogen 1:1000) in 10% goat serum with PBS for 1 h, and washed with PBST. Confocal microscopy analyses were carried out using Zeiss LSM510.
TUNEL Staining-The free 3-OH strand breaks resulting from DNA degradation were detected by TUNEL, using an in situ death detection kit, fluorescein (Roche) as per the kit instructions. Coverslip mounting with DAPI was performed and images obtained as above. Overlay and counting was performed with ImageJ software (National Institutes of Health).
Electron Microscopy-MIN6 cells were seeded into 6-well plates and grown in DMEM supplemented with 15% FBS (MIN6) and 10% FBS (INS-1), 10 mM HEPES buffer,1 mM pyruvate, 50 M 2-␤-mercaptoethanol, 50 units/ml penicillin, and 50 g/ml streptomycin. The cells were then incubated in either control medium or medium containing 0.4 mM PA for 16 h. The cells were then rinsed in PBS and fixed with 3.0% glutaraldehyde in 0.1 cacodylate buffer (pH 7.4) at 4°C for 2 h. The wells were rinsed four times in 0.1 M cacodylate with 5% sucrose for 5 min each. Cells were osmicated in 1% OsO 4 for 45 min at 4°C, dehydrated in a graded series of ethanol (50 -100%), infiltrated with eponate 12 resin, and placed in a 60°C oven for 48 h. Sectioning was performed with using a Reichert Ultracut E microtome and mounted on copper 300 mesh grids before staining with uranyl acetate and lead citrate. The grids were examined in a Philips Tecnai 12 BioTWIN transmission electron microscope, and images were captured with a Gatan 4K ϫ 4K digital camera.
Data Analysis-All values are expressed as means Ϯ S.D. Student's t test and one-way analysis of variance were employed for two-group and multiple comparisons, respectively. All measurements were considered significant if the p value was Յ0.05 (*) or Յ0.01 (**).

RESULTS
Exposure of ␤-Cells to Glucose and PA Leads to Apoptotic Cell Death-Because hyperglycemia and FFAs result in impaired ␤-cell function and are contributing factors to diabetes, we investigated the effects of saturated fatty acid PA (0.4 mM) alone or in combination with increasing concentrations of glucose (from 5.5 to 25 mM) on ␤-cells at graded incubation times (from 6 to 24 h). As shown in Fig. 1, A-C, incubation of INS-1 cells with varying concentrations of glucose and PA alone or in combination reduced cell viability. A similar situation was also observed for MIN6 cells (Fig. 1, D and E). Next, we investigated the mechanism of PA-induced cell death by evaluating the apoptotic pathway. Immunoblotting analysis showed an increase in the cleaved form of caspase 3 ( Fig. 1, F-H) and PARP, suggesting caspase activation and apoptosis. To confirm our immunoblotting results, treated ␤-cells were stained with annexin V and propidium iodide and analyzed by flow cytometry to discriminate between apoptotic and necrotic cells (Fig. 2, A-C). We found that the proportion of annexin V-positive cells was ele-vated with increasing concentrations of glucose or PA alone, suggesting apoptosis (Fig. 2, A-C). To extend our findings to human ␤-cells, insulin and TUNEL staining was performed on histological sections from isolated pancreatic islets treated with either PA (0.4 mM) or glucose (16.5 mM) to look for changes in the expression of insulin and an increase in the number of apoptotic ␤-cells (Fig. 2, D-H). These experiments revealed that exposure to glucose (16.5 mM) and PA (0.4 mM) decreased the expression of insulin and induced 6-and 7-fold increases in the number of TUNEL-positive ␤-cells, respectively, relative to islets at 5.5 mM (normal) glucose (  cose induce apoptotic cell death in ␤-cells both alone or in combination. PA and Glucose Treatment Results in the Accumulation of Autophagosomes in ␤-Cells-Impaired autophagy has been linked to the pathophysiology of T2D (31). Recent studies on pancreatic ␤-cells from T2D patients have shown accumulation of vacuoles and APs, concomitant with an increase in apoptotic ␤-cell death (31). We therefore hypothesized that metabolic stress leads to accumulation of APs, thus resulting in apoptotic cell death. To test this hypothesis, we first examined the appearance of autophagic vacuoles in ␤-cells treated with PA and glucose using EM. Our EM analysis showed an increase in APs with abundant engulfed cargo in ␤-cells treated with PA, whereas relatively few APs were present in control samples (Fig.  3A). Because FFAs have been shown to increase AP formation (2), we treated the ␤-cell lines and human pancreatic islets with PA and measured levels of the autophagic marker, LC3B-II (which is a well characterized marker used to visualize APs and also represents the membrane bound form of LC3B on autophagic vacuoles). As shown in Fig. 3B, the levels of LC3B-II were significantly higher in PA (0.4 mM)-treated ␤-cells, thus indicating an increase in the number of APs. Blockade of autophagy by treatment with bafilomycin (200 nM), which inhibits the fusion of APs with lysosomes, further increased the levels of LC3B-II and Sequestosome-1 (SQSTM1/p62) when concomitantly treated with PA (0.4 mM) (Fig. 3, B and C), thus reflecting an increase in the rate of AP formation. Our results reveal that whereas PA inhibited autophagy, it induced a shift in LC3B localization to a punctate pattern (Fig. 3D). In addition, PA (0.4 mM) treatment of MIN6 cells stably expressing GFP-LC3B significantly increased the number of GFP-LC3B puncta per cell as compared with control (p ϭ 0.003), (Fig. 3E). The APs were larger, and the GFP fluorescence intensity (FI) was brighter in ␤-cells treated with PA (0.4 mM) than under starvation (positive control), suggesting impairment in AP degradation. To confirm increased AP accumulation, WIPI-1 (WD repeat domain phosphoinositide-interacting protein 1), a marker for phagophores, was measured in MIN6 cells stably expressing WIPI-GPF and treated with PA (0.4 mM) alone or in combination with glucose (20 mM). We found that the number of WIPI-1 puncta per cell increased from 5 Ϯ 1.4 in the BSA-treated control cells to 18 Ϯ 5 in the PA-treated cells, which was significant (p ϭ 0.002) (Fig. 3, F and G). No further increases in puncta were observed following concomitant PA (0.4 mM) and glucose (20 mM) exposure (22 Ϯ 4.2, p ϭ 0.002) (Fig. 3, F and G). Insulin and p62 double staining in human islets treated with glucose (33.3 mM) or PA (0.4 mM) alone or in combination showed co-localization of insulin-depleted ␤-cells with p62 (Fig. 3H), suggesting that the expression of p62 reflects damaged ␤-cells. Our results demonstrate that lipotoxicity leads to accumulation of APs and p62, potentially reflecting a defect in the lysosomal degradation machinery of ␤-cells.
PA and Glucose Suppress Autophagic Turnover-Long-lived proteins are degraded during the process of autophagy. To test whether protein degradation was affected by FFAs, we performed a pulse-chase autophagic flux assay. We found that the release of [ 14 C]valine was decreased in INS-1 cells treated with PA alone or in combination with glucose at 18 h (p ϭ 0.03 and p ϭ 0.04, respectively; Fig. 4, A and B). Bafilomycin blocked 40% of [ 14 C]valine release, indicating that proteins are degraded by autophagy. The decrease in [ 14 C]valine release by PA was almost unchanged whether bafilomycin was present or absent in the media, indicating that the suppression of protein degra- dation by PA was mediated by autophagy. We then determined levels of SQSTM1/p62, a key substrate of autophagy whose primary function is thought to involve turnover of ubiquitinated proteins and aged or damaged cell organelles (32). We measured SQSTM1/p62 expression levels and found that they were increased in ␤-cell lines treated with PA (0.4 mM) and glucose (5.5-35 mM) (Fig. 4, C, D, F, and G). Ubiquitinated proteins, which are normally removed by SQSTM1/p62, have been shown to accumulate in ␤-cells of T2D patients (33). We tested whether the increased levels of SQSTM1 led to clearance of ubiquitinated proteins. Instead we found the reverse: accumulation of ubiquitinated proteins in ␤-cells treated with PA (0.4 mM) alone or in combination with glucose (15 mM) (Fig. 4, E and  H). Starvation (which potently induces autophagy) of MIN6-GFP-LC3-expressing cells did not prevent the accumulation of SQSTM1/p62 when treated with PA (0.4 mM) for 14 h (Fig. 4I), indicating that PA provides an overwhelming impairment in autophagic turnover. PA also increased unprocessed free GFP (Fig. 4I), as did 3-methyl adenine, an autophagy inhibitor (Fig. 4J), thus suggesting impaired autophagy. To determine whether PA and glucose affect autophagic turnover in human pancreatic islets, we treated human cadaveric islets with PA (0.4 mM) and glucose (16.5 mM) and found accumulation of several early normally degraded autophagy markers such as LC3B-II, p62, and ATG5 (Fig. 5, A and B). These results suggest impaired autophagic protein turnover.
Autophagic turnover is dependent on the fusion of APs with lysosomes resulting in acidic autolysosomes. We asked whether PA affects lysosomal machinery, and to this end measured GFP fluorescence as it decreases within the acidic lysosomal environment. We utilized flow cytometry to directly measure GFP fluorescence intensity within treated ␤-cells (34), thus estimating the level of fusion of APs with lysosomes. In this assay, decreased autophagic turnover leads to an increase in total FI in cells expressing GFP-LC3, whereas stimulation of autophagy by starvation or treatment with rapamycin leads to decreased GFP FI. Treatment of the MIN6-GFP-LC3 cell line with PA (0.4 mM) significantly increased the GFP FI by 26 Ϯ 6% as compared with control. However, this result was significantly smaller than that obtained following treatment with the autophagy inhibitor bafilomycin (200 nM) (58 Ϯ 10%), suggesting that PA induces partial blockage of autophagic flux (Fig. 5C). Treatment of the MIN6-GFP-LC3 cell line with oleate (0.4 mM) increased the GFP FI at 12 h (Fig. 5C), but the GFP FI intensity decreased significantly at 18 and 24 h compared with PA-treated cells (Fig.  5D). Western blot of the oleate treated MIN6 cells showed decreased expression of p-mTOR and CHOP, thus suggesting stimulation of autophagy (Fig. 5E). Furthermore, co-localization between GFP-LC3B (Green) and LysoTracker Red (a marker for lysosomes) was reduced in cells treated with PA compared with untreated controls, suggesting defects in lysosomal machinery (Fig. 5F). Lysosomal acidification is dependent on ATP hydrolysis by the H ϩ pump, and we found decreased cellular ATP levels in MIN6 cells treated with PA (0.4 mM) as compared with untreated control (47 Ϯ 7 nmol/mg) and (80 Ϯ 9) nM/mg protein, respectively (Fig. 5G). Our results demonstrate that autophagic turnover in ␤-cells is impaired when treated with PA and glucose, thus leading to accumulation of ubiquitinated and long-lived proteins.
Ultrastructural Changes in ␤-Cells When Treated with PA-Because p62 targets ubiquitinated proteins and damaged organelles to lysosomes for degradation, we hypothesized that decreased autophagic turnover may lead to accumulation of damaged organelles in ␤-cells. We performed EM to analyze ultrastructural changes and found cisternal distention of ER, massive vacuolization and aberrant and degenerated mitochondrial morphology in MIN6 cells treated with (0.4 mM) PA alone (Fig. 6A). Mitochondria in PA-treated ␤-cells showed damaged structures and lost cristae (Fig. 6B). EM studies also showed an increase in the total number of mitochondria in these cells (Fig. 6D, right panel). In addition, mitochondrial staining with MitoTracker Deep Red was increased in MIN6 cells treated with PA, suggesting a decrease in mitochondrial turnover (Fig. 6C, right panel). Our results demonstrate accumulation of mitochondria, which is attributed to impaired or inefficient autophagy.
PA Activates mTOR Signaling-SQSTM1/p62 activates mTORC1 by associating with Raptor, a critical subunit of the mTORC1 (35). mTORC1 is also activated by nutrient overload (14). Because treatment of ␤-cells with PA and glucose increases p62 levels, we asked whether PA and glucose exposure induce the activation of mTORC1 in ␤-cells. Immunoblot analysis showed an increase in mTORC1 activation in a doseand time-dependent manner in MIN6 and INS-1 cells when treated with PA (0.2-0.5 mM) alone or in combination with glucose (Fig. 7, A-C). Increased ␤-cell apoptosis has been observed in mice expressing a constitutively active form of S6K1, which is a downstream target of mTORC1 (37). We therefore determined whether proteins downstream of mTORC1 were also activated. Western blot analysis of MIN6 cells treated with increasing concentrations of PA (0.2-0.5 mM) showed increased phosphorylation of key proteins downstream of mTORC1, including 4EBP1 (eukaryotic translation initiation factor 4E-binding protein 1) and rpS6 (ribosomal protein S6) (Fig. 7D). Recent studies have shown that hyperactivation of mTORC1 in TSC1 or TSC2 (tuberous sclerosis 1 and 2)-deficient fibroblasts leads to an increase in the unfolded protein response caused by ER stress (14). Because accumulation of polyubiquitinated proteins in ␤-cells leads to increased ER stress (38), we next asked whether activation of mTORC1 in ␤-cell lines and human islets induces ER stress. We measured the expression of CCAAT enhancer-binding protein homologous protein (CHOP), a transcription factor and key mediator of cell death in response to ER stress, and found its expression was higher in ␤-cell lines and human islets when treated with increasing concentrations of PA (0.2-0.5 mM) and glucose (16.5 mM) (Fig. 7, E and F). Furthermore, IHC staining of human cadaver islets treated with PA (0.4 mM) and glucose (16.5 mM) showed intense staining for CHOP, compared with control islets (5.5 mM glucose) (Fig. 7G). We next asked whether stimulation of autophagy in ␤-cells decreased ER stress. We measured the expression of ER stress-related genes DDIT3, ATF4, and DNAJC3 by RT-PCR and found lower expression of these genes when treated with rapamycin (Fig. 7, H-J). We conclude that activation of mTORC1 by PA alone or in combination with glucose inhibits autophagic turnover, which in turn induces ER stress in ␤-cells, leading to their apoptotic cell death.
Blockage of Autophagy in MIN6 Cells Leads to Cell Death-Because our results suggest that PA and glucose inhibit autophagic turnover, we next asked whether modulating autophagic flux would affect PA-induced ␤-cell death. Using RNA interference, we knocked down Bif-1 (Bax-interacting factor 1, also known as Endophilin B). Bif-1 interacts with Beclin1 through UVRAG (ultraviolet irradiation resistance-associated gene) and regulates autophagy (39). As shown in Fig. 8, A and B, shRNA targeting Bif-1 reduced mRNA and subsequent protein levels of Bif-1. Lentiviral shRNA targeting Bif-1 reduced AP formation in MIN6 cells treated with PA (0.4 mM) for 16 h (Fig. 8C). Levels of p62, increased in MIN6-Bif-1 knockdown cells (Fig. 8C), suggesting impairment in autophagy. Serum starvation (a positive control of autophagy) of MIN6-Bif-1 knockdown cells inhibited cell death (Fig. 8D) as shown in earlier studies in autophagydeficient Atg5 and Bif-1 Ϫ/Ϫ MEFs (39), thus suggesting that autophagy is the primary cell death mechanism under starva- tion conditions. To determine whether inhibition of autophagy reduced the viability of MIN6 cells, we treated cells with PA (0.4 mM) and found that viability of the MIN6-Bif-1 knockdown cells was reduced by 40% compared with control (Fig. 8E), whereas PA-induced apoptosis was augmented (Fig. 8, F and  G). Treatment of MIN6 cells with autophagy inhibitors monensin (a protein transport inhibitor) and bafilomycin A1 (a V-ATPase inhibitor) (200 nM) significantly increased the proportion of annexin V-positive cells as compared with BSA and PA-treated controls (Fig. 9, A and B), thus suggesting an increase in apoptosis. Co-treatment of MIN6 cells with PA (0.4 mM) and bafilomycin (200 nM) (an inhibitor of fusion between APs and lysosomes), significantly decreased the viability of cells by 30% compared with PA-treated cells (Fig. 9C). Conversely, stimulation of autophagy with rapamycin (200 nM) (an inhibitor of mTORC1) and PA (0.4 mM) significantly increased the viability of MIN6 cells by roughly 30% (Fig. 9C) compared with PA treatment alone. In addition, treatment of MIN6-GFP cells with rapamycin also significantly decreased GFP FI when co-treated with PA as compared with PA alone (Fig. 9D), suggesting an increase in autophagic turnover. Lastly, we determined whether apoptosis was augmented in MIN6 cells treated with PA (0.4 mM) and rapamycin (200 nM). Western blot analysis revealed a decrease in the cleaved form of caspase 3 (Fig. 9E) as compared with cells treated with PA (0.4 mM) alone, suggesting a decrease in apoptotic cell death when cells were treated with PA and rapamycin. To confirm our immunoblotting results, we assayed the treated MIN6 cells with annexin V and propidium iodide (Fig. 9F). FACS analysis showed a decrease in annexin V-positive cells when MIN6 cells were treated with PA (0.4 mM) and rapamycin (200 nM) as compared with PA-treated cells, demonstrating that rapamycin was able to reduce PA-induced apoptosis (Fig. 9, F and G). We therefore conclude that inhibition of autophagy in ␤-cells decreased their viability via increases in apoptotic cell death, whereas stimulating autophagy by rapamycin increases the autophagic flux and inhibits cell death.

DISCUSSION
Glucolipotoxicity is a main contributor to the progression of diseases, such as obesity and diabetes, caused by oxidative ER stress and accumulation of ubiquitinated proteins within cells leading to apoptotic cell death (23,40). Autophagy plays an essential role for the maintenance of ␤-cell mass and function (41)(42)(43), whereas its dysregulation has been associated with obesity, insulin resistance, and glucose intolerance (44). Autophagy regulates lipid metabolism (45), but little is known about the effect of lipid injury on autophagy levels or activity. We sought to evaluate the consequences of nutrient overload in ␤-cells based on this being the main feature of obese and T2D patients. To test our hypothesis, we examined the effects of PA and glucose alone or in combination on autophagic turnover in ␤-cells. We found that treatment of ␤-cells with different concentrations of glucose and PA alone or in combination decreased cell viability and induced apoptotic cell death. ␤-Cell apoptosis was associated with suppressed autophagic turnover explaining the toxic effects of nutrient overload, which correlated with the studies showing that ␤-cell specific knock-out of Atg7 in mice leads to decreased ␤-cell mass, along with increased ␤-cell death (23). Recent studies using mice fed with high fat diet or genetic models of obesity (ob/ob) and diabetes (db/db) showed reduced autophagy in liver as indicated by down-regulation of different autophagy markers (26). Conversely, it has also been reported that PA induces autophagy in both in vivo and in vitro model systems (23,46). These inconsistencies are partly due to technical issues associated with performing in vivo autophagic assays and also attributed to factors such as cell types used, ages of the in vivo experimental animal models, the concentration and duration of FFA treatment, and the ratio of conjugated BSA to FFA used (47).
Our data point to diverse consequences elicited by impaired autophagic turnover. We found that defective autophagy caused by nutrient overload leads to accumulation of LC3B, an early autophagy biomarker, which was previously attributed to an increase in autophagic flux. However, recent advances have provided ample evidence questioning this interpretation (7,8). Increased LC3B was also associated with an increase in accumulation of APs, which is evidenced by an increase in autofluorescent dots in GFP-LC3-transfected cells and an increase in WIPI-1 puncta. This was also accompanied with an accumulation of large ubiquitin-containing proteins with overexpression of the LC3-binding protein SQSTM1/p62, which otherwise are continuously cleared by autophagy, thus resulting in apoptotic cell death. Our observation of decreased autophagic turnover leading to accumulation of key autophagy substrate proteins, like p62, is supported by other studies showing that diabetesinduced oxidative stress and ␤-cell specific knock-out of Atg7 led to accumulation of p62 and ubiquitinated proteins (13). Studies in high fat diet-fed mice and db/db mice also showed higher levels of LC3B, accompanied with an increased number of autophagosomes (48). Also in vivo hepatic suppression of Atg7 in mice caused insulin resistance and ER stress, whereas suppression in in vitro cell models caused defective insulin signaling (26).
Autophagy plays an important role in clearance of large longlived ubiquitinated proteins. Our studies showed a decrease in long-lived protein degradation in ␤-cells treated with PA alone or in combination with glucose. We also provided evidence indicating a blockage of autophagy by FFAs by showing an increase in GFP-LC3 FI and accumulation of different autophagy substrates. Our EM studies identified accumulation of APs with abundant engulfed cargo in PA-treated ␤-cells, which is comparable with observations in the ␤-cells of T2D patients (31). Proper lysosomal acidification is required for autophagic turnover. Any slight changes in ATP levels impair lysosomal H ϩ pump activity, and we observed a decrease in cellular ATP levels when ␤-cells were treated with FFAs. Our results were also supported by recent studies in INS-1 cells, which showed that treatment with PA led to decreased expression of ATP6AP2 and its adaptor protein AP2M1, both of which contribute to lysosomal acidification (25). Similar changes have also been reported in several neurodegenerative diseases (9,12). Recent reports in T2D patients have described altered expression of LAMP-2 and cathepsin B and D in ␤-cells of T2D patients, suggesting impairment in the latter stages of autophagy (31).
Metabolically stressed human ␤-cells display markers of ER stress and activation of the apoptosis pathway (49). Similar forces in obese and T2D patients, i.e. enhanced demand for insulin production accompanied by elevated levels of free fatty acids, place a strain on ␤-cells leading to chronic ER stress (44). Autophagy plays an important protective role in ␤-cells during ER stress (50), but its impairment may precipitate the onset of ER stress (51), thus eliciting the expression of transcription factor Ddit3/CHOP, which is regarded as a key mediator of cell death in response to ER stress. Our results demonstrate that CHOP is highly up-regulated in ␤-cells when treated with PA and glucose, thus indicating that autophagy-deficient ␤-cells are more vulnerable to FFA-induced ER stress. On the other hand, stimulation of autophagy in ␤-cells by rapamycin decreased ER stress.
Autophagy plays an important role in cell survival under starvation conditions, but it also participates in cell death. Therefore, we investigated the role of autophagy in PA-induced ␤-cell death by stimulation or inhibition of autophagy. Inhibition of autophagy in ␤-cell lines by down-regulating Bif-1 expression or treatment with bafilomycin A1 increased FFA-induced apoptotic cell death, suggesting a protective role for autophagy in ␤-cells. mTORC1, which inhibits autophagy, is activated by nutrient overload (14). We speculate that activated mTORC1 in ␤-cells treated with PA and glucose inhibits autophagy leading to accumulation of SQSTM1/p62 and ubiquitinated proteins. This is consistent with recent observations (14), where hyperactivation of mTORC1 in Ulk Ϫ/Ϫ islets and Tsc2 Ϫ/Ϫ fibroblasts leads to accumulation of SQSTM1/p62. Defective mitophagy by mTORC1 activation leads to an increase in mito-chondrial mass (52), and we have observed accumulation of damaged or dysfunctional mitochondria in PA-treated ␤-cells. Also, recent studies show FFAs induce the arrest of mitochondrial fusion and fragmentation (53), concomitant with mitochondrial damage in ␤-cells (54), thus correlating with our studies. These findings imply that mitochondrial dysfunction and ER stress leads to ␤-cell death. Inhibiting mTORC1 using rapamycin alleviated the suppression of autophagic turnover by PA and glucose, and this was also accompanied with a protection from glucolipotoxicity-induced cell death. Recent studies in diabetic mouse models have shown improved glycemia, increased pancreatic insulin content, and decreased ␤-cells apoptosis when treated with rapamycin (55). Also drugs already in clinical use such as metformin, which potentiates AMPK activity and stimulates autophagy, have been shown to decrease ␤-cell death in isolated islets from T2D, as well as nondiabetic islets cultured with PA (56). On the other hand, chronic inhibition of mTORC1 by rapamycin has been shown to cause glucose intolerance in mice (15), as well as insulin resistance in humans (57). This is because chronic inhibition of mTORC1 with rapamycin disrupts the second mTOR complex, mTORC2, which is required for the insulin-mediated suppression of hepatic gluconeogenesis (36), thus leading to glucose intolerance and reduced ␤-cell function.
In conclusion, nutrient overload results in impairment of autophagy leading to accumulation of autophagosomes, autophagy substrates like p62, ER stress, accumulation of aberrant mitochondria, lysosomal dysfunction, and activation of mTORC1, thus contributing to ␤-cell death (Fig. 10). Stimulation of autophagy by treatment with mTORC1 inhibitor rapamycin alleviated the suppression of autophagic turnover by PA and glucose decreased ER stress, and this was also accompanied with a protection from glucolipotoxicity induced cell death. Thus, therapeutic interventions using pharmacological agents, which can improve ER folding capacity and stabilize misfolded proteins, as well as target the autophagy machinery, could provide promising strategies for treating human diseases such as T2D. FIGURE 10. Inhibition of autophagic turnover by palmitatic acid leads to apoptotic cell death. By activating mTORC1, palmitic acid impairs autophagic turnover, resulting in increased ER stress and decreased mitophagy, thus triggering ␤-cell death.