Chronic exposure to elevated saturated free fatty acid (FFA) levels has been shown to induce endoplasmic reticulum (ER) stress that may contribute to promoting pancreatic β-cell apoptosis. Here, we compared the effects of FFAs on apoptosis and ER stress in human islets and two pancreatic β-cell lines, rat INS-1 and mouse MIN6 cells. Isolated human islets cultured in vitro underwent apoptosis, and markers of ER stress pathways were elevated by chronic palmitate exposure. Palmitate also induced apoptosis in MIN6 and INS-1 cells, although the former were more resistant to both apoptosis and ER stress. MIN6 cells were found to express significantly higher levels of ER chaperone proteins than INS-1 cells, which likely accounts for the ER stress resistance. We attempted to determine the relative contribution that ER stress plays in palmitate-induced β-cell apoptosis. Although overexpressing GRP78 in INS-1 cells partially reduced susceptibility to thapsigargin, this failed to reduce palmitate-induced ER stress or apoptosis. In INS-1 cells, palmitate induced apoptosis at concentrations that did not result in significant ER stress. Finally, MIN6 cells depleted of GRP78 were more susceptible to tunicamycin-induced apoptosis but not to palmitate-induced apoptosis compared with control cells. These results suggest that ER stress is likely not the main mechanism involved in palmitate-induced apoptosis in β-cell lines. Human islets and MIN6 cells were found to express high levels of stearoyl-CoA desaturase-1 compared with INS-1 cells, which may account for the decreased susceptibility of these cells to the cytotoxic effects of palmitate.
- endoplasmic reticulum
lipotoxicity contributes to β-cell dysfunction during the development of type 2 diabetes (19, 42, 48). Chronically elevated levels of free fatty acids (FFAs), particularly saturated FFAs such as palmitate, have been shown to be cytotoxic to pancreatic β-cells (8, 10, 24, 30–33, 41, 46, 50). The monounsaturated FFA oleate on the other hand has been shown to not cause apoptosis of β-cells in some studies (16, 17, 31, 32, 35), while other studies (11, 24, 27, 33, 50) found that it does, particularly those that used the INS-1 β-cell line.
Palmitate-induced cytotoxicity can result from multiple mechanisms, such as increases in reactive oxygen species (8, 18, 33), ceramide, and nitric oxide (NO) levels (30, 45) and mitochondrial perturbations (8, 18, 33). In addition, recent studies (23–25) have shown that endoplasmic reticulum (ER) stress pathways are activated by chronic palmitate exposure.
ER stress is caused when the protein folding capacity of the ER is not sufficient to deal with protein folding demands or when an excess of misfolded or aggregated proteins accumulate. Such conditions activate the unfolded protein response (UPR) that attempts to reduce the amount of new protein synthesis, increase folding capacity, and degrade terminally misfolded proteins (20, 53). Prolonged ER stress can lead to apoptosis induction (37, 53) and thus may contribute to palmitate-induced β-cell death. The relative contribution of the various mechanisms by which excess FFAs cause apoptosis induction in different pancreatic β-cell models, however, is not well established.
In this study, we examined palmitate-induced apoptosis and ER stress in human islets and two commonly used pancreatic β-cell lines. Unexpectedly, we found that human islets and MIN6 cells are more resistant to palmitate-induced ER stress and apoptosis than INS-1 cells. Palmitate is able to induce apoptosis in INS-1 cells at concentrations that do not cause ER stress, while knocking down GRP78 expression in MIN6 cells did not render them significantly more susceptible to palmitate-induced apoptosis. Our results suggest that mechanisms other than ER stress are likely playing a more prominent role in inducing apoptosis in vitro. MIN6 cells and human islets express high levels of the enzyme stearoyl-CoA desaturase-1 (SCD1), which may protect these cells from palmitate-induced cytotoxic effects.
MATERIALS AND METHODS
Rat INS-1 pancreatic β-cells were obtained from the laboratory of C. Wollheim (2) and maintained in RPMI 1640 (11.1 mM glucose, 1 mM sodium pyruvate, and 10 mM HEPES) supplemented with 10% FBS, 2 mM l-glutamine, and 55 μM β-mercaptoethanol containing antibiotics (100 U/ml penicillin and 100 μg/ml streptomycin) at 37°C and 5% CO2. Mouse MIN6 pancreatic β-cells were cultured in DMEM (25 mM glucose, 2 mM l-glutamine, and 1 mM sodium pyruvate) supplemented with 10% FBS and 55 μM β-mercaptoethanol at 37°C and 5% CO2.
Human islet isolation and culture.
Human islets were obtained from the Clinical Transplantation Laboratory at the University of Alberta from nondiabetic human cadavers. Human islets were isolated following the Edmonton Protocol (36) and cultured in RPMI 1640 medium supplemented with 10% FBS, 25 mM HEPES, 2 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin, at 37°C in a humidified chamber. Human islets were incubated in RPMI 1640 and treated as indicated in the figure legends.
FFA preparation, cell treatment, and lyses.
FFA/BSA solutions were prepared as described previously (13, 23). For most experiments, 0.5 mM FFA complexed with 0.5% BSA or 1% BSA was used (FFA:BSA: molar ratio of 6.6:1 and 3.3:1, respectively). INS-1 or MIN6 cells were cultured on 10-cm dishes or 12-well plates depending on the experiment. FFA/BSA was added in serum-free media as indicated in the figure legends. One micromolar thapsigargin (1 and 6 h) or 2 μg/ml tunicamycin (16 h) were used as positive controls for induction of ER stress in some experiments. At the indicated time points, the cells were washed in PBS and lysed in ice-cold lysis buffer (1% Triton X-100, 20 mM HEPES, pH 7.4, 100 mM KCl, 2 mM EDTA, 1 mM PMSF, 10 μg/ml leupeptin, and 10 μg/ml aprotinin). For experiments in which phosphorylated proteins were examined, the lysis buffer also contained phosphatase inhibitors (10 mM NaF, 2 mM Na3VO4, and 10 nM okadaic acid). The cells were lysed on ice for 30–60 min and centrifuged at 13,000 rpm for 10 min at 4°C. The supernatant was then transferred to a new tube, and the protein concentration was determined with BCA reagent (Pierce).
For data presented in Fig. 1, human islets were dispersed using Dispase II (Roche Applied Science) according to the manufacturer's instructions. Subsequently, dispersed cells were plated on poly-l-lysine treated coverslips in 24-well plates. The human islets were then treated as described in Fig. 1 and stained with FITC-annexin V and propidium iodide, as per the manufacturer's instructions (Vybrant Apoptosis Assay Kit 3, Molecular Probes) and visualized using an Olympus BX51W1 fluorescence microscope fitted with ×20/0.95 water immersion objective and cooled CCD camera (U-TVCA, Olympus). All cells in a given field of view were counted (∼400 cells/condition were counted). Cells that bound FITC-annexin V but excluded propidium iodide were termed apoptotic (annexin V-positive) and cells permeant to propidium iodide were deemed necrotic.
The cell death detection kit ELISAPLUS (Roche Diagnostics) was used to monitor FFA-induced apoptosis (see Figs. 2, 6, and 8). One day before the experiments INS-1 or MIN6 cells were seeded in 24-well plates (200,000 cells/well; see Fig. 2), 12-well plates (500,000 cells/well; see Fig. 6) or 24-well plates (50,000 cells/well; see Fig. 8). The cells were treated as indicated in the figure legends. After the treatment conditions, the cells were lysed and oligonucleosomes in the cytosol (indicative of apoptosis-associated DNA degradation) were quantified according to the manufacturer's instructions. Cells grown in the presence of serum had near background levels of ELISA signal compared with lysis buffer alone indicative of a low level of apoptotic cells in the control population. Consequently, all treatment conditions were normalized to cells grown in the presence of serum at each time point.
GRP78/BiP and green fluorescent protein recombinant adenovirus production.
Full-length human GRP78/BiP DNA was amplified by PCR from pCMV-GRP78 WT, a plasmid obtained from Dr. R. Prywes (Columbia University), and SpeI and EcoRV restriction sites were introduced to flank the DNA fragment. The SpeI/EcoRV PCR fragment was cloned into the pCRII-TOPO vector (Invitrogen). The fragment was then cloned into the SpeI/EcoRV site of the pShuttle-IRES-hr-GFP-2 vector (Stratagene). Correct ligation of the insert into the vector was confirmed by DNA sequencing. To produce an adenovirus expressing GRP78, the AdEasy XL Adenoviral Vector System (Stratagene) was used according to the instructions provided. Briefly, the pShuttle-IRES-hrGFP-2 vector containing the GRP78 insert and a supercoiled viral DNA plasmid (pAdEasy-1) were used to produce a recombinant adenovirus plasmid by homologous recombination. This plasmid was purified and transfected into the AD-293 cell line using the MBS mammalian transfection kit (Stratagene). After a primary viral stock was prepared by freeze-thawing infected AD-293 cells, the viral particles were subjected to one round of amplification. The amplified adenovirus was then used to infect four 10-cm plates of AD-293 cells and was purified using the Vivapure AdenoPACK 100 kit (Vivascience). The purified virus was resuspended in buffer (20 mM Tris·HCl, 25 mM NaCl, and 2.5% glycerol, pH 8.0), and aliquots were stored at −80°C. The viral titer (plaque-forming units/ml) of the purified adenovirus was determined by optical absorbance at 260 nm. Control recombinant enhanced green fluorescent protein (GFP)-expressing adenovirus was generated and purified as described above.
INS-1 and MIN6 cell infection.
INS-1 or MIN6 cells (500,000 cells/12-well plate) were infected with 9 × 108 plaque-forming units/ml of Ad-GRP78 or Ad-GFP and incubated at 37°C and 5% CO2 for 2 h with gentle shaking every 30 min. The cells were then washed once with PBS and replaced with fresh RPMI 1640 medium. After being cultured for 24 h, the cells were treated with either thapsigargin or palmitate as indicated in the figure legends. As a control, cells were cultured in the same conditions without the addition of Ad-GRP78 or Ad-GFP. After the treatments, the cells were washed once with PBS and lysed for either Western blot analysis or apoptosis measurment as described above.
Small interfering RNA transfection.
Knockdown of GRP78 expression in MIN6 cells was performed using small interfering RNA (siRNA). MIN6 cells were transfected with GRP78 siRNA (Invitrogen) or control GFP siRNA (obtained from Dr. O. Ouerfelli, Sloan-Kettering Institute) using Lipofectamine RNAiMAX reagent (Invitrogen) according to the manufacturer's instructions. Briefly, per transfection 6 pmol of siRNA was diluted in 100 μl Opti-MEM I medium without serum (Invitrogen) directly in one well of a 24-well plate and mixed gently. After incubation at room temperature for 5 min, 1 μl of Lipofectamine RNAiMAX reagent (Invitrogen) was added to each well containing the diluted siRNAs, mixed gently, and incubated at room temperature for 20 min. During the incubation period, MIN6 cells were diluted in DMEM medium without antibiotics to obtain a concentration of 100,000 cells/ml. To each well containing siRNA/Lipofectamine RNAiMAX complexes, 0.5 ml of the diluted cells was added (i.e., 50,000 cells/well) to obtain a final siRNA concentration of 10 nM. The cells were mixed gently and incubated at 37°C for 72 h. The cells were treated as indicated in the figure legends and processed for either Western blot analysis or apoptosis measurement using the cell death ELISA assay (Roche Diagnostics).
Measurement of XBP-1 mRNA splicing.
Total RNA was isolated from INS-1 cells as described above. Rat XBP-1 cDNA was amplified by RT-PCR (QIAGEN OneStep RT-PCR kit) using primers that flank the intron excised by IRE1 exonuclease activity as described previously (52). Primer sequences used to amplify rat XBP-1 were 5′-AAA CAG AGT AGC AGC ACA GAC TGC-3′ and 5′-TCC TTC TGG GTA GAC CTC TGG GAG-3′. The protocol used for the RT-PCR reaction was: 50°C (30 min); 95°C (15 min); 30 cycles of 94°C (1 min), 62°C (1 min), and 72°C (1 min); and 72°C (10 min). RT-PCR products were resolved on a 3% agarose gel and visualized using ethidium bromide. The 480-bp product (unspliced rat XBP-1) and the 454 bp product (spliced XBP-1) were scanned, and the bands were quantified using Scion Image software. The percentage of spliced XBP-1 vs. full-length XBP-1 mRNA was determined.
RNA isolation and real-time quantitative PCR.
For real-time PCR data presented in Fig. 1, total RNA was isolated from human islets using the Trizol Reagent (Invitrogen). After extraction, total RNA was treated with rDNAse I (Ambion) then run on a 1% agarose gel to confirm its integrity. First-strand synthesis of complementary DNA (cDNA) was carried out using Superscript II RNase H reverse transcriptase and oligo(dT) primers following the manufacturer's instructions (Invitrogen). The resulting cDNA (10 ng per reaction) was used for amplification in quantitative real-time PCR (qPCR). Serial dilutions of human genomic DNA (gDNA) was used for the generation of a standard curve. Briefly, gDNA or cDNA (4 μl/well) was added to a qPCR mixture (6 μl/well) containing the following components: 3.475 μl water, 1 μl of 10× PCR buffer, 0.6 μl of 50 mM MgCl2, 0.2 μl of 50 μM primer mix (or 0.1 μl forward and 0.1 μl reverse), 0.2 μl of 10 mM dNTP mixture, 0.2 μl ROX reference dye, 0.3 μl SYBR green I (stock diluted 1:1,000 in water), and 0.025 μl (or 125 U) platinum Taq polymerase (Invitrogen). For PCR amplification the following general protocol was used: 95°C (3 min); 40 cycles of PCR: 95°C (10 s), 65°C (15 s), 72°C (20 s); 95°C (15 s), 60°C (15 s), and 95°C (15 s). qPCR was performed in the ABI Prism 7900 HT sequence detection system (Applied Biosystems). Gene-specific oligonucleotide primers were designed using the Primer Quest SciTool from Integrated DNA Technologies. Primer sequences are available in the online version of this article that contains supplemental data. The expression level of the various transcripts was calculated using the standard curve method. Values were normalized to expression of human β-actin mRNA and represent the average of four independent experiments.
For real-time data (see Fig. 9), total RNA was isolated from INS-1 and MIN6 cells and human islets using the Trizol Reagent (Invitrogen). Total RNA was reverse-transcribed to single-stranded cDNA using the high-capacity cDNA reverse transcription kit (Applied Biosystems). The resulting cDNA was used for real-time PCR analysis using the TaqMan Gene Expression system (Applied Biosystems). SCD1 (rat Scd1, Rn00594894_g1; mouse Scd1, Mm00772290_m1; and human Scd1, Hs00748952_s1) or β-actin-specific primers (rat β-actin, 4352931E; mouse β-actin, 4352933E; and human β-actin, 4333762F) and TaqMan MGB probes were obtained from Applied Biosystems. Serial dilutions of INS-1 cDNA were used to generate a standard curve. A final real-time PCR reaction of 25 μl containing 10 μl cDNA, 1.25 μl ddH2O, 1.25 μl TaqMan gene expression assay (20×), and 12.5 μl TaqMan universal PCR master mix (2×) were loaded onto each well of a ABI PRISM 96-well optical reaction plate. Reactions were run on an ABI Prism 7900HT sequence detection system using the following protocol: 10 min at 95°C, 40 cycles of 15 s at 95°C, and 1 min at 60°C. The standard curve and corresponding values for each sample were determined by the SDS 2.1 software of the ABI Prism 7900HT instrument. Values were normalized to expression of β-actin mRNA for each species and presented as means ± SE of three independent experiments.
Western blot analysis.
Proteins were resolved by SDS-PAGE and transferred to nitrocellulose membranes. After incubation with secondary antibody conjugated to horseradish peroxidase, the bands were detected with the enhanced chemiluminescence system (Amersham Bioscience). Immunoblots were scanned and quantified using Scion Image software. For detection of insulin by immunoblotting the samples were resolved using 4–12% NuPage gels (Invitrogen). The following primary antibodies were used: phospho-eukaryotic initiation factor-2α (eIF2α) (Cell Signaling, #9721, 1:500), GADD153/CHOP (Santa Cruz Biotechnology, sc-575, 1:500), anti-KDEL (StressGen, SPA-827, 1:1,000), PDI (StressGen, SPA-890, 1:4,000), γ-tubulin (Sigma, T6557, 1:1,000), guinea-pig anti-insulin (Dako, 1:250), anti-actin (StressGen, CSA-400, 1:1,000), and GM130 (Transduction Laboratories, G65120, 1:500).
Results are means ± SE. Statistical significance between two experimental conditions was analyzed using Student's two-sample t-test assuming equal variance. Data from several or more groups were analyzed by ANOVA followed by Tukey's post hoc test. P < 0.05 was considered statistically significant.
Differential susceptibility to apoptosis and ER stress of human islets, rat INS-1, and mouse MIN6 β-cells induced by chronic FFA exposure.
Chronic treatment with elevated levels of FFA and in particular saturated FFA induces apoptosis in human pancreatic β-cells in vitro (6, 16, 17, 30, 31). Recent evidence has shown that islets from type 2 diabetic patients have increased levels of some markers of ER stress pathways such as ER chaperones and the ER stress-associated apoptosis promoting gene CHOP (25). Here we examined the effect of FFA on apoptosis and ER stress response in human islets. Isolated human islets were plated in culture dishes and treated for 16 h with 1 mM palmitate complexed to 1% BSA or 1 mM oleate complexed to 1% BSA (6.6:1 FFA:BSA) in media without serum. A small, but significant, increase in apoptosis was observed in islets treated with palmitate, but not oleate, compared with BSA-treated islets (Fig. 1A). Real-time PCR analysis revealed that palmitate treatment caused the upregulation of some genes induced by ER stress (Fig. 1B). The mRNA levels of XBP-1, ATF-4, and GRP78 were slightly elevated, but the effects did not reach statistical significance. The mRNA levels of the CHOP transcription factor, however, were significantly increased (Fig. 1B). In palmitate-treated islets, the amount of phospho-eIF2α was increased compared with control BSA-treated islets, indicating that the PERK ER stress sensor is activated (Fig. 1C). Interestingly, treatment of human islets with thapsigargin or tunicamycin, pharmacological inducers of ER stress, induced PERK pathway activation (eIF2α phosphorylation), but the effect was less pronounced than in INS-1 cells (Fig. 1D).
In addition to activating ER stress pathway signaling, palmitate also causes ER morphological changes that in INS-1 cells are evident after only a few hours of exposure (23). Similar changes were observed in human islets but only after prolonged (16 h) incubation with palmitate (Supplemental Figs. S1 and S2). Thus, human islets appear to be more resistant to ER stress compared with INS-1 cells. This is consistent with the observation that activation of the PERK pathway (eIF2α phosphorylation) by thapsigargin or tunicamycin is less effective in human islets compared with INS-1 cells (Fig. 1D).
We next compared the effect of FFA on apoptosis in two pancreatic β-cell lines, rat INS-1 and mouse MIN6 β-cells. Consistent with our previous results (23), treatment of INS-1 cells with 0.5 mM palmitate complexed with 0.5% BSA (6.6:1 palmitate:BSA) for 16 and 24 h resulted in a marked increase in apoptotic cells over BSA controls, while the unsaturated FFA oleate (6.6:1 oleate:BSA) induced less apoptosis at the 16-h time point (Fig. 2). Under the same experimental conditions, however, the extent of apoptosis induced by palmitate in MIN6 cells was markedly lower and oleate had no effect (Fig. 2). Thus MIN6 cells are considerably more resistant to FFA-induced apoptosis than INS-1 cells.
To examine the susceptibility of these cell lines to palmitate-induced ER stress, the cells were treated with palmitate and markers of ER stress signaling were monitored. The transcription factor CHOP in particular is activated by ER stress conditions and is involved in mediating in part ER stress-induced apoptosis (37, 41). As expected, CHOP is induced in INS-1 and MIN6 cells that are exposed to thapsigargin (Fig. 3, A and B). We have previously documented that maximal expression of CHOP is observed after only 6 h of treatment with 1 mM palmitate complexed to 1% BSA (6.6:1 palmitate:BSA) in INS-1 cells (23). A representative experiment is shown in Fig. 3A. In MIN6 cells, CHOP induction is not maximal until 24 h of treatment with the same concentration of palmitate (Fig. 3B). The presence of an unknown nonspecific band in MIN6 cells made accurate quantification of the CHOP signal intensity difficult, but the result is presented in Fig. 3C. Clearly, CHOP is not induced until at least 16 h of palmitate exposure. These results suggest that MIN6 cells are more resistant to ER stress than INS-1 cells.
ER stress is not the primary mechanism by which palmitate induces apoptosis in vitro.
GRP78 is known to bind the luminal domains of the three ER stress transducer proteins PERK, IRE1, and ATF6 to keep them inactive (3, 9, 29, 44). A recent study (25) has shown that overexpression of GRP78 in MIN6 cells protects these cells from palmitate-induced apoptosis and activation of ER stress pathways (25). We therefore examined if overexpression of GRP78 in INS-1 cells could protect these cells from ER stress and apoptosis induced by chronic palmitate treatment. We generated recombinant adenoviruses to overexpress GRP78 in INS-1 cells given that pancreatic β-cell lines are difficult to transfect with conventional methods (1).
We first determined whether GRP78 overexpression can protect INS-1 cells from ER stress induced by thapsigargin, which depletes ER calcium stores and is known to cause the accumulation of misfolded and unfolded proteins in the ER. An approximately fivefold overexpression of GRP78 was achieved at a viral dose that had no effect on cell morphology or growth (Fig. 4, A and B). A recombinant GFP-expressing virus at the same viral titer was used as a control. We examined the ability of thapsigargin to cause ER stress by monitoring the activation of the PERK and IRE1 UPR pathways. The PERK pathway initiates an immediate response to reduce protein load in the ER by attenuating general protein translation via phosphorylation of eIF2α (43). A 1-h treatment with thapsigarin induces an approximately fivefold increase in the levels of phospho-eIF2α in control INS-1 cells (Figs. 1D and 4C). As shown in Fig. 4A, treatment of INS-1 cells with thapsigargin for 1 h led to the phosphorylation of eIF2α and reduction of steady-state proinsulin expression levels in GFP- and GRP78-overexpressing cells. Thapsigargin treatment is expected to reduce steady-state proinsulin levels since insulin is the major secretory protein being synthesized in β-cells. The increase in phosphorylated eIF2α tended to be lower in GRP78-overexpressing compared with control GFP-expressing cells in response to thapsigargin (Fig. 4, A and C), but thapsigargin gave more variable results in infected cells and this did not reach significance. GRP78-overexpressing cells also tended to have higher steady-state proinsulin levels compared with GFP-expressing cells, although by band densitometry analysis the increase was not statistically significant (Fig. 4, A and D).
ER stress also triggers the IRE1 signaling pathway. Activated IRE1 cleaves an intron of the XBP-1 mRNA, leading to efficient translation and the production of the active XBP-1 transcription factor, which subsequently translocates to the nucleus and induces UPR target gene transcription (7, 26, 51). We tested the effect of GRP78 overexpression on the activation of IRE1 pathway by measuring XBP-1 mRNA splicing by RT-PCR. Thapsigargin treatment of INS-1 cells led to efficient splicing of XBP-1 mRNA, as shown by the appearance of the smaller spliced form (Fig. 4E). XBP-1 mRNA splicing in response to thapsigargin treatment was significantly reduced in cells overexpressing GRP78, compared with GFP-overexpressing or uninfected cells (Fig. 4, E and F). Overall, these results show that overexpression of GRP78 reduces both PERK and IRE1 pathway signaling, indicating that this approach affords INS-1 cells some protection against ER stress.
We next examined the effect of overexpression of GRP78 on palmitate-induced ER stress and apoptosis. As shown in Fig. 5A, exposure of INS-1 cells to 0.5 mM palmitate complexed to 1% BSA (3.3:1 palmitate:BSA) for 16 h in media without serum leads to phosphorylation of eIF2α. The effect was variable, however, and band intensity analysis did not reveal a statistically significant difference in phosphorylated eIF2α levels in GRP78 compared with GFP-overexpressing cells (Fig. 5C). However, palmitate (3.3:1 palmitate:BSA) exposure did cause an approximately threefold increase in phosphorylated eIF2α levels in control uninfected cells that was statistically significant (Fig. 5C). This level of palmitate, however, did not cause detectable XBP-1 splicing (Fig. 5D). These results suggests that little if any ER stress is induced under these conditions. Exposure of INS-1 cells to 6.6:1 palmitate:BSA induced XBP-1 splicing, but this was not affected by overexpressing GRP78 (Fig. 5E). Therefore, despite expressing approximately threefold higher levels of GRP78 (Fig. 5B), INS-1 cells were not protected against palmitate-induced activation of ER stress pathways or palmitate-induced apoptosis (Fig. 6A). However, at concentrations of palmitate (3.3:1 palmitate:BSA) that induced minimal or no ER stress (Fig. 5, C and D), significant apoptosis was still observed (Fig. 6B), suggesting that ER stress may not be the main mechanism contributing to palmitate-induced apoptosis in INS-1 cells.
We also examined whether overexpression of GRP78 in MIN6 cells could protect these cells from palmitate-induced apoptosis, but no significant effect was observed (results not shown). MIN6 cells, however, already express high levels of ER chaperone proteins (Fig. 7). Thus, overexpressing GRP78 by approximately twofold in MIN6 cells may not afford sufficiently higher protection against ER stress to make a large enough impact to reduce palmitate-induced apoptosis. Indeed, MIN6 cells are more differentiated than INS-1 cells, express higher ER chaperone levels (Fig. 7), and are more resistant to ER stress (Fig. 3), which might be why they are significantly more resistant to palmitate-induced apoptosis (Fig. 2). To test this hypothesis, we knocked-down GRP78 expression in MIN6 cells using siRNA to examine the effect on ER stress and apoptosis. As shown in Fig. 8A, GRP78 expression was reduced by ∼60% compared with control transfected cells. MIN6 cells depleted of GRP78 had higher levels of apoptosis in the control conditions (either grown in serum or the absence of serum; Fig. 8B). Palmitate treatment induced apoptosis to a similar extent in both control (GFP siRNA-transfected) and GRP78-depleted cells. In contrast, GRP78-depleted cells exposed to tunicamycin had markedly higher levels of apoptosis compared with control siRNA-transfected cells (Fig. 8B). These results suggest that ER stress has a minimal contribution to palmitate-induced apoptosis in MIN6 cells in vitro.
Human islets and MIN6 cells express high levels of stearoyl-coenzyme A desaturase-1.
In addition to reduced apoptosis and ER stress in MIN6 cells in response to chronic palmitate exposure, we also found that, similar to human islets, MIN6 cells were resistant to the effects of palmitate on cell morphology (Supplemental Fig. S3). The appearance of angular clefts in the cytoplasm as a result of altered ER morphology were only observed after prolonged (>12 h) treatment with palmitate in MIN6 cells, whereas this was evident after only a couple of hours in INS-1 cells. It is possible that MIN6 cells do not transport palmitate as efficiently as INS-1 cells. However, we found that the uptake of [3H]palmitate or C16-BODIPY was similar in the two cell lines (Supplemental Fig. S4).
A recent study (5) has identified palmitate-resistant MIN6 clones that express higher levels of stearoyl-coenzyme A desaturase-1 (SCD1) compared with the palmitate-sensitive parental cells. SCD is an ER-localized enzyme that converts saturated fatty acids to monounsaturated fatty acids by introducing a single double bond in fatty acyl-CoA substrates, including palmitoyl- and stearoyl-CoA (15, 22). This changes the saturated FFA species to less toxic unsaturated forms. Thus, we hypothesized that human islets and MIN6 cells being palmitate-resistant may express high levels of SCD1. We compared the mRNA expression levels of SCD1 in human islets and INS-1 and MIN6 cells by real-time PCR. For the real-time analysis, primers were used that were specific for human, rat, or mouse SCD1 and the expression level in each cell type was normalized to its own levels of β-actin. As shown in Fig. 9, human islets and MIN6 cells express significantly higher steady-state levels of SCD1 mRNA (relative to β-actin) than INS-1 cells. This may contribute to the increased resistance of human islets and MIN6 cells to chronic palmitate exposure.
Obesity is a major risk factor for the development of type 2 diabetes, and chronically elevated FFAs may contribute to the progression of this disease by promoting pancreatic β-cell apoptosis (19, 42, 48). Exploration into the mechanisms of FFA-induced β-cell apoptosis have used pancreatic islets and β-cell lines treated with FFA in vitro. Such studies have identified that saturated FFAs such as palmitate induce considerable β-cell apoptosis (5, 8, 10, 24, 30, 32, 33, 46). In addition, chronic treatment with palmitate has also been shown to cause apoptosis in many different cell types (14, 21, 28, 40, 49).
Different cell types have differential susceptibility to FFA. When we compared the effects of chronic FFA treatment on apoptosis in human islets and two commonly used β-cell lines, we found that human islets and MIN6 cells were resistant to FFA-induced apoptosis compared with INS-1 cells. The differential susceptibility of INS-1 and MIN6 cells to palmitate-induced apoptosis is striking (Fig. 2) and is in line with published studies (5, 25). In addition to greater resistance to apoptosis caused by palmitate, we found that MIN6 cells and human islets are also more resistant to ER stress. Induction of ER stress markers by palmitate or ER stress-inducing toxins is less marked and takes longer in MIN6 and human islets compared with INS-1 cells. MIN6 express higher steady-state levels of ER chaperones (GRP78, GRP94, and PDI) than INS-1 cells, which may contribute to their greater resistance to ER stress.
Chronic conditions associated with diabetes and obesity such as low level inflammation, hyperglycemia, and hyperlipidemia have recently been shown to induce ER stress in different cell types (38). The contribution of ER stress induction to the detrimental effects of these treatments is beginning to be examined. Reducing ER stress in the ob/ob obesity-associated diabetic mouse model has recently been shown to improve insulin resistance in liver and adipose cells (39). In this study, we examined if the ER stress that is caused by chronic palmitate exposure in pancreatic β-cells in vitro contributes to causing the apoptotic event.
We attempted to increase the ability of INS-1 cells to deal with ER stress conditions by overexpressing the major ER stress sensor GRP78. Using an adenoviral approach, we successfully increased GRP78 levels by approximately three- to fivefold in INS1 cells, which afforded some protection against ER stress caused by thapsigargin (Fig. 4). This approach significantly reduced IRE1 pathway activation but only modestly reduced PERK pathway activation. The reason for the differential effect may relate to the fact that the PERK pathway is more sensitive to ER stress caused by ER calcium depletion than the IRE1 pathway (17). Consequently, the reduction in the amount of misfolded proteins in GRP78-overexpressing cells is enough to significantly reduce IRE1 activation but is less effective in reducing PERK activation. Interestingly, INS-1 cells overexpressing GRP78 tended to have higher steady-state proinsulin levels, indicating that GRP78 may be important for insulin biosynthesis. We examined the effect of chronic palmitate exposure on ER stress and apoptosis in GRP78-overexpressing β-cells. However, no significant reduction of either ER stress pathway activation or apoptosis was observed (Figs. 5 and 6A). Since palmitate-induced ER stress was not reduced in GRP78-overexpressing cells we are unable to determine the effect of ER stress in contributing to apoptosis induction with this approach. However, in INS-1 cells 0.5 mM palmitate (3.3:1 palm:BSA) was able to induce significant β-cell apoptosis but minimal ER stress pathway activation. Thus apoptosis can be initiated by palmitate independent of significant ER stress pathway activation in INS-1 cells.
We also failed to observe a protective effect by overexpressing GRP78 in MIN6 cells. MIN6 cells, however, already express high levels of GRP78 (Fig. 7). GRP78 was therefore knocked down in MIN6 cells using siRNA. Depleting GRP78 expression in MIN6 cells significantly increased the susceptibility of these cells to tunicamycin (ER stress-induced apoptosis) but not to palmitate-induced apoptosis (Fig. 8). Thus, ER stress is unlikely to be a major mechanism by which palmitate induces apoptosis in pancreatic β-cells under in vitro experimental conditions.
Our results are somewhat different from those of a recent study (25) showing that overexpressing GRP78 can protect MIN6 cells from palmitate-induced apoptosis. Palmitate promoted apoptosis in approximately one-half of the population of GRP78-overexpressing cells compared with control transfected cells. We did not observe a protective effect in either INS-1 or MIN6 cells following a protocol of administering FFA in the absence of serum, conditions that likely make the cells more susceptible to oxidative stress (33). We attempted the protocol used by Laybutt et al. (25), adding BSA-complexed palmitate in the presence of serum for 72 h but did not observe any significant apoptosis induction in our MIN6 cell line. It is possible that we failed to observe a protective effect due to our adenoviral approach, as opposed to the electroporation transfection approach used by Laybutt et al. (25). The GRP78 overexpression levels obtained may be insufficient to afford enough ER stress protection. However, this is difficult to compare given that there is no quantified data for the level of ER stress reduction in that study. Our results using GRP78 siRNA in MIN6 cells (Fig. 8) indicate that cells with reduced GRP78 are not more sensitive to palmitate, and thus ER stress has a minimal role in causing palmitate-induced apoptosis. Furthermore, as we reported in INS-1 cells (23), palmitate induces limited ER stress as indicated by the lack of significant GRP78 protein upregulation compared with ER stress toxins such as tunicamycin (Fig. 8A). Chronic treatment of β-cells with saturated FFA has been shown to cause numerous other cellular alterations that can initiate apoptosis including generation of reactive oxygen species (8, 18, 33), mitochondrial perturbations (8, 18, 33), and generation of ceramide and NO (30, 45). These mechanisms are likely to be more important in initiating and causing apoptosis in response to chronic FFA exposure. Furthermore, it is possible that some of these changes may be responsible for inducing subsequent ER stress. Oxidative stress for example can cause ER stress and activate the UPR (34).
Why might human islets and MIN6 cells be dramatically more resistant to palmitate-induced apoptosis than INS-1 cells? As shown in Fig. 9, MIN6 cells and human islets express significantly higher levels of SCD1, the rate-limiting enzyme in the biosynthesis of monounsaturated fats (12, 15). SCD1 is the major gene target of leptin and appears to be a metabolic control point partitioning fat towards storage or oxidation in the liver (12). The role of this enzyme in the β-cell is not established. Recently, a subclone of MIN6 cells was isolated that is resistant to palmitate-induced apoptosis and has elevated SCD1 expression (5). This clone of MIN6 cells is better able to handle excess saturated FFA by desaturating them to less toxic species. Saturated FFAs, but not unsaturated FFAs, are substrates for ceramide production (47) and ceramide has been implicated in palmitate-induced β-cell apoptosis (30, 46).
Elevated SCD1 levels in human islets and MIN6 cells may also indirectly protect these cells from ER stress conditions. In INS-1 cells treatment with palmitate leads to gross morphological changes that are observed by light and electron microscopy (23, 35). These changes likely result from both the storage of excess palmitate in the form of tripalmitin triglyceride in the ER (35) and incorporation of palmitate into ER phospholipids making them rigid (4). These effects may perturb the ER protein folding environment leading to suboptimal protein folding conditions resulting in ER stress-induced apoptosis. In MIN6 cells, morphological changes are induced by chronic palmitate exposure (Supplemental Fig. S3), but these changes are only observed after prolonged treatment with palmitate, while they are readily apparent after only a few hours in INS-1 cells (23). Thus, it is possible that increased SCD1 expression protects MIN6 cells from these effects by converting the palmitate to unsaturated species that do not accumulate in the ER but stay in the cytosol as relatively harmless triglyceride (35). It remains to be determined whether enhancing SCD1 expression can protect cell types such as β-cells that are not specialized for fat storage from the harmful effects of excess saturated FFAs.
In summary, we show that the saturated FFA palmitate causes apoptosis in human islets and cultured β-cell lines. Human islets and mouse MIN6 cells are resistant to ER stress compared with INS-1 cells, although ER stress does not appear to be a major mechanism by which saturated FFAs induce β-cell apoptosis in vitro. The resistance of human islets and MIN6 cells to palmitate may relate to high levels of SCD1 expression found in these cells, but this requires future study. MIN6 cells behave more like human islets and are a better model system for studying the effects of FFAs than INS-1 cells.
This work was supported, in part, by operating grants to A. Volchuk from the Canadian Diabetes Association (GA-2-06-2138-AV) and to M. B. Wheeler from the Canadian Institutes of Health Research (MOP-12898). Infrastructure support for the A. Volchuk laboratory was provided by funding from the Canadian Foundation for Innovation and the Ontario Innovation Trust. A. Volchuk is the recipient of a Tier II Canada Research Chair award. E. Lai is supported in part by a Banting and Best Diabetes Centre graduate studentship. G. Bikopoulos is supported by a Natural Sciences and Engineering Research Council of Canada graduate studentship.
We thank E. Karaskov for starting this project and contributing to some of the experiments in this study.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 by American Physiological Society