There are strong links between obesity, elevated free fatty acids, and type 2 diabetes. Specifically, the saturated fatty acid palmitate has pleiotropic effects on β-cell function and survival. In the present study, we sought to determine the mechanism by which palmitate affects intracellular Ca2+, and in particular the role of the endoplasmic reticulum (ER). In human β-cells and MIN6 cells, palmitate rapidly increased cytosolic Ca2+ through a combination of Ca2+ store release and extracellular Ca2+ influx. Palmitate caused a reversible lowering of ER Ca2+, measured directly with the fluorescent protein-based ER Ca2+ sensor D1ER. Using another genetically encoded indicator, we observed long-lasting oscillations of cytosolic Ca2+ in palmitate-treated cells. In keeping with this observed ER Ca2+ depletion, palmitate induced rapid phosphorylation of the ER Ca2+ sensor protein kinase R-like ER kinase (PERK) and subsequently ER stress and β-cell death. We detected little palmitate-induced insulin secretion, suggesting that these Ca2+ signals are poorly coupled to exocytosis. In summary, we have characterized Ca2+-dependent mechanisms involved in altered β-cell function and survival induced by the free fatty acid palmitate. We present the first direct evidence that free fatty acids reduce ER Ca2+ and shed light on pathways involved in lipotoxicity and the pathogenesis of type 2 diabetes.
- free fatty acids
- fluorescence resonance energy transfer
- calcium homeostasis
- endoplasmic reticulum
type 2 diabetes is characterized by peripheral insulin resistance, defects in insulin secretion, and β-cell apoptosis (4, 13, 52, 65). Obesity is a well-known risk factor for diabetes and is characterized by elevated levels of circulating free fatty acids. Chronic elevation of the long-chain saturated fatty acid palmitate has been shown to cause β-cell death (27, 36, 37, 48, 56). In contrast, unsaturated fatty acids such as oleate are considered to be less cytotoxic and in some cases protective against palmitate toxicity (20, 27, 37, 38). Acute administration of palmitate is thought to cause a modest increase in insulin secretion (9, 24), while prolonged exposure to this fatty acid increases basal insulin secretion while decreasing the relative response to glucose (66). Nevertheless, the mechanisms by which palmitate affects β-cells are not fully understood.
One model of lipotoxicity proposes that palmitate initiates β-cell endoplasmic reticulum (ER) stress (16, 24, 32) and insulin resistance in multiple tissues (43). The ER is a major cellular Ca2+ store, and ER Ca2+ homeostasis is critical for ER function (55) as well as proinsulin folding and processing (21). Accordingly, depletion of ER Ca2+ by cytokines or inhibitors of the sarco(endo)plasmic reticulum Ca2+-ATPase (SERCA) uptake pump triggers an ER stress response in β-cells (5, 34). However, whether palmitate depletes ER Ca2+ stores remains unclear. Recent studies employing indirect measurements of the β-cell ER filling state have provided conflicting information (27, 10). The ability of palmitate to alter cytosolic Ca2+ is well established in HIT-T15, INS-1, and primary mouse β-cells (49, 54). In HIT-T15 cells, the maintenance of these signals was dependent on extracellular Ca2+ influx through voltage-gated Ca2+ channels (49). However, the observation that a Ca2+ transient remained in the absence of extracellular Ca2+ suggested that a component of these Ca2+ signals might be initiated by intracellular stores (49). Activation of G protein-coupled fatty acid receptors would be expected to mobilize inositol 1,4,5-trisphosphate (IP3)-sensitive Ca2+ stores (3, 23, 30). This notwithstanding, the effects of palmitate on ER Ca2+ levels have not been measured directly with targeted ER Ca2+ sensors (46). Although these previous studies suggest that Ca2+ dynamics play an important role in the effects of palmitate, the precise mechanisms by which palmitate affects β-cell Ca2+ homeostasis remain unresolved, especially in human β-cells.
In the present study, we used a combination of cytosolic Ca2+ imaging of human β-cells and fluorescence resonance energy transfer (FRET)-based Ca2+ imaging in MIN6 cells to test the hypothesis that both extracellular Ca2+ and ER Ca2+ contribute to the effects of palmitate on β-cell Ca2+ homeostasis. Our data provide new insight into the mechanisms involved in fatty acid-induced signaling and ER stress.
All reagents were from Sigma (St. Louis, MO), unless otherwise indicated. The acetoxymethyl ester of the cytoplasmic Ca2+ dye fura-2 (fura-2 AM; Molecular Probes, Eugene, OR) was stored in dimethyl sulfoxide (DMSO) as a 1 mM stock solution. Thapsigargin and xestospongin C (Cayman Chemical, Ann Arbor, MI) were stored in DMSO as a 1 mM stock solution. Carbachol (Calbiochem, San Diego, CA) was dissolved in water and stored as a 100 mM stock solution. The cameleon FRET probes D1ER and D3cpv were gifts from Dr. Amy Palmer (Dept. of Chemistry and Biochemistry, University of Colorado, Boulder, CO). Drug solutions were prepared immediately before each experiment.
Palmitic acid (Nu-Check Prep, Elysian, MN) was dissolved in 65 mM NaOH and complexed with 20% essentially fatty acid-free bovine serum albumin (BSA). The complex was added to Dulbecco's modified Eagle's medium (DMEM; Invitrogen, Burlington, ON, Canada) containing 5 mM or 25 mM glucose, 10% fetal bovine serum (Invitrogen), and 1% penicillin-streptomycin (Invitrogen) for a final palmitate-to-BSA molar ratio of 6:1 (7), unless otherwise indicated. For imaging experiments, sodium palmitate was prepared as described previously (49). Briefly, a stock solution of 100 mM was prepared by dissolving palmitate in 60 mM NaOH and added to prewarmed Ringer solution (3 mM glucose) containing 0.1% essentially fatty acid-free BSA in a molar ratio of 6:1. 2-Bromopalmitate solutions were prepared in the same way. Oleate solutions were prepared as described previously (54) by dissolving oleic acid in hot 100 mM NaOH to a 100 mM stock solution and diluted by addition to prewarmed Ringer solution containing 0.1% essentially fatty acid-free BSA in a molar ratio of 6:1. For imaging experiments, the final concentration of palmitate, oleate, and 2-bromopalmitate was 100 μM unless otherwise shown. Vehicle controls contained BSA and NaOH and did not induce cell death or Ca2+ fluxes (data not shown).
The mouse insulinoma-derived MIN6 cell line (40) was cultured as described previously (14). Human islets were provided by Dr. Garth Warnock (Ike Barber Transplantation Unit, Dept. of Surgery, Vancouver General Hospital, Vancouver, BC, Canada) with support from the Michael Smith Foundation for Health Research-funded Centre for Human Islet Transplant and β-Cell Regeneration. Donors were both men and women from 20 to 70 yr of age without a history of diabetes. Human islets were cultured in RPMI-1640 medium (Invitrogen) with 10% FBS and 2% penicillin-streptomycin in humidified 5% CO2-95% air (vol/vol) at 37°C. Primary human islets were dispersed onto sterile coverslips 2 days before experiments with a previously described technique (33).
MIN6 cells were grown to a confluence of 30–60% for cameleon or fura-2 imaging. MIN6 cells were transfected with D1ER or D3cpv cameleon DNA 1 day after plating on sterile 22-mm coverslips with Lipofectamine 2000 reagent (Invitrogen) and 1 μg/μl midi-prepped construct DNA in Opti-MEM (Invitrogen). Cells were left to adhere for 3 h, and then fresh DMEM was added and cells were imaged within 2–3 days.
Live cell imaging.
Single-cell imaging was performed in HEPES-buffered Ringer solution containing (in mM) 144 NaCl, 5.5 KCl, 1 MgCl2, 2 CaCl2, 20 HEPES, and 3 glucose, unless otherwise indicated (25). Cytosolic Ca2+ was imaged in fura-2 AM-loaded cells as described previously (25, 33) or in D3cpv-expressing cells (46). ER luminal Ca2+ was imaged with the FRET-based D1ER cameleon (34, 45, 46). Preheated solutions were applied by stable perifusion at 1 ml/min, and complete solution changes were achieved in <30 s. Images were taken with a ×20 objective, unless otherwise indicated, with a Zeiss 200m inverted microscope, a Sutter Lambda DG4 excitation source, a Roper CoolSnap HQ camera, and Slidebook Software (Intelligent Imaging Innovations, Denver, CO). The cyan fluorescent protein (CFP) component of the cameleons was excited at 430 nm with a S430/×25 filter (Chroma). CFP emission and yellow fluorescent protein (YFP) FRET emission were alternately collected at 470 and 535 nm, respectively, with S470/30m and S535/30m filters mounted in a Sutter Lambda 10-2 emission filter wheel. The FRET signal was normalized to the CFP emission intensity, and changes in ER or cytosolic Ca2+ were expressed as the FRET-to-CFP ratio. There was no correlation between the apparent Ca2+ levels and the intensity of the FRET probes (D3cpv and D1ER; data not shown). The analysis of D1ER-measured ER Ca2+ responses was performed on a cell-by-cell basis (34). A response was considered to have occurred if 1) there was an inflection of the calcium signal within the treatment period that deviated from baseline by >15% and 2) this inflection was reversed after the treatment was removed.
Protein extraction and Western blotting.
MIN6 cells were grown on six-well plates (Becton Dickinson Labware, Franklin Lakes, NJ) to a confluence of 80–90% and then lysed with Cell Signaling Lysis Buffer (Beverly, MA) for CHOP and Xbp-1 or Cell Signaling Lysis Buffer supplemented with phosphatase inhibitors (in mM: 50 sodium fluoride, 10 sodium pyrophosphate decahydrate, and 10 β-glycerophosphate) for phospho-protein kinase R-like ER kinase (PERK) (34). The cells were then sonicated for 30 s and cleared by centrifugation at 13,000 rpm for 5–15 min. Protein concentrations were determined via bicinchoninic acid (BCA) assay (Pierce, Rockford, IL), and samples were loaded onto SDS-PAGE gels. Separated proteins were transferred onto polyvinylidene difluoride (PVDF) membranes via a semidry transfer apparatus and blocked with I-Block (Tropix, Bedford, MA) containing 0.1% Tween 20. Rabbit polyclonal Chop primary antibody and rabbit polyclonal Xbp-1 primary antibody (Santa Cruz Biotechnology, Santa Cruz, CA) incubations were done overnight at 4°C, while rabbit monoclonal phospho-PERK (Cell Signaling, Boston, MA) was done over two nights at 4°C. Secondary antibodies were incubated at room temperature for 1 h in I-Block (Chop) or 5% fat-free milk in 1× TBS (phospho-PERK and Xbp-1) containing 0.1% Tween 20.
On-line cell death assay.
Cell death was analyzed as described previously (2, 34). Briefly, cells were plated onto clear-bottom 96-well microplates (Perkin Elmer, Waltham, MA) and treated as indicated. Cell death was monitored with a Cellomics (Pittsburgh, PA) KineticScan Reader to image the incorporation of propidium iodide (500 ng/ml), a dye that is normally nonfluorescent but fluoresces brightly once it passes through the compromised plasma membrane of a dying cell and binds to DNA.
Supernatants from MIN6 cell cultures were assayed for insulin as described previously (2). Mouse islets were cultured overnight after isolation, after which groups of 100 size-matched islets were suspended with Cytodex microcarrier beads (Sigma) in the 300-μl plastic chambers of an Acusyst-S perifusion apparatus (Endotronics, Minneapolis, MN). Islets were perifused at 37°C and 5% CO2 at 0.5 ml/min with a Krebs-Ringer buffer containing (in mM) 129 NaCl, 5 NaHCO3, 4.8 KCl, 2.5 CaCl2, 1.2 MgSO4, 1.2 KH2PO4, 10 HEPES, and 3 glucose. Radioimmunoassay-grade BSA (5 g/l; Sigma) was also included. Before sample collection, islets were equilibrated under basal (3 mM glucose) conditions for 1 h. Insulin secretion was measured by radioimmunoassay (Rat Insulin RIA Kit, Linco Research, St. Charles, MO).
All studies were replicated at least three times. At least three separate human donors were used for the primary cell experiments. Statistical analysis was performed as indicated in the text. ANOVA (Student-Newman-Keuls) or unpaired two-tailed t-test was used where appropriate. A P value of <0.05 was considered significant.
Acute effects of palmitate on cytosolic Ca2+.
Palmitate is known to stimulate Ca2+ signals in some β-cell models (49, 54), but there are no reports of palmitate-induced Ca2+ signals in human β-cells. Thus, to extend previous investigations, we employed human islet β-cells, identified by their response to 20 mM glucose. In these cells, we observed robust and complex cytosolic Ca2+ signals in response to 100 μM palmitate (Fig. 1A). Qualitatively similar Ca2+ signals were observed when MIN6 cells were treated with palmitate (Fig. 1B). The amplitude and response rate of cytosolic Ca2+ signals was analyzed over a range of palmitate concentrations, from 25 μM to 500 μM palmitate (all palmitate-to-BSA ratio of 6:1). These experiments demonstrate that palmitate induces a significant number and amplitude of responses at the lowest concentration tested (Fig. 1E). Next, we examined the effect of 2-bromopalmitate, a nonmetabolizable analog of palmitate that inhibits long-chain acyl-CoA transfer into the mitochondria (50). Compared with palmitate, 2-bromopalmitate evoked a distinct, blunted response that never showed oscillations in either MIN6 or human β-cells (Fig. 1, C and D). These observations suggest that while the large and oscillatory responses require mitochondrial fatty acid oxidation, modest monophasic Ca2+ signals can be induced in the absence of palmitate metabolism. Palmitate action through G protein-coupled receptors would not be expected to require lipid metabolism, and these data point to the possibility that a component of these palmitate-induced Ca2+ signals may be due to the mobilization of intracellular Ca2+ stores, possibly due to the activation of cell surface receptors for fatty acids.
Mechanisms of palmitate-induced Ca2+ signals.
Next, we sought to identify the sources of the palmitate-induced Ca2+ signals. First we tested whether extracellular Ca2+ is required for the effects of palmitate by perifusing β-cells with Ca2+-free buffer. In human β-cells, palmitate caused a transient Ca2+ signal but not a sustained Ca2+ elevation in the absence of extracellular Ca2+ (Fig. 2A). Similar results were obtained with MIN6 cells (Fig. 2B). We also found that removing extracellular Ca2+ inhibited palmitate-induced Ca2+ signals that had already been initiated (Fig. 2C). These results demonstrate that palmitate can initiate Ca2+ signals in the absence of extracellular Ca2+, but also that Ca2+ influx from the extracellular space is required for continued propagation of palmitate-induced Ca2+ signals. Our findings implicate intracellular Ca2+ stores as a component of fatty acid-induced Ca2+ signaling, in agreement with studies using other β-cell models (49, 54). To determine the extent to which IP3 receptors are involved in this intracellular store release, we treated cells with palmitate in the presence of xestospongin C, a drug that blocks IP3-stimulated Ca2+ signals in β-cells (26). We pretreated MIN6 cells with 1 μM xestospongin C for 10 or 20 min and quantified the amplitude of the initial (1st phase) and sustained (2nd phase) palmitate-induced Ca2+ signals (Fig. 2, D and E). The first minute of response was expected to involve intracellular stores (Fig. 2, A and B), and we noted that xestospongin C tended to reduce the first phase of the palmitate-induced Ca2+ signals. However, this difference did not reach statistical significance despite the large number of cells studied. This result suggests that the IP3 receptor (IP3R) may not play a dominant role in palmitate-induced intracellular store Ca2+ release.
Acute palmitate treatment decreases ER Ca2+.
The ER is thought to be the main dynamic intracellular Ca2+ storage compartment in β-cells (16). We used a recently developed targeted “cameleon” probe (D1ER) to dynamically measure Ca2+ levels in the ER lumen of MIN6 cells (Fig. 3A) (34, 45, 46). This genetically engineered probe has Ca2+-responsive elements between an enhanced CFP (eCFP) and a YFP (citrine). FRET between these fluorescent proteins changes on Ca2+ binding to the probe (45). We found that acute palmitate treatment in a subset (31.3 ± 4.7%) of MIN6 cells caused a reversible decrease in ER Ca2+ (Fig. 3B). These Ca2+ signals were also repeatable within individual cells. The decrease in ER Ca2+ caused by palmitate was 45 ± 3.7% of that seen with 1 μM thapsigargin, a drug that directly inhibits the SERCA Ca2+ uptake pumps on the ER (Fig. 3, C and D). Furthermore, treatment of MIN6 cells with 2-bromopalmitate did not consistently stimulate a robust release of ER Ca2+ (n = 83 cells; data not shown). In a limited number of preliminary experiments, we found that blocking IP3R with 1 μM xestospongin C did not block palmitate-induced ER Ca2+ release in MIN6 cells (data not shown). Together, these data provide the first direct demonstration that intracellular Ca2+ store mobilization plays a role in palmitate-induced Ca2+ signaling.
Measurements of cytosolic Ca2+ with D3cpv cameleon.
In addition to its ability to acutely stimulate β-cells, palmitate has potent long-term effects on β-cell function and survival. We sought to determine the effects of sustained palmitate exposure on β-cell calcium homeostasis. One limitation of conventional Ca2+ dyes, such as fura-2, is that they perform poorly over long periods of time (>1 h) because of dye leakage, dye compartmentalization into organelles, and photobleaching. To examine the long-term effects of palmitate on β-cell cytosolic Ca2+ homeostasis, we employed another recently developed FRET-based Ca2+ indicator, D3cpv, which is targeted to the cytosol (Fig. 4A) (44, 46). As expected, the D3cpv FRET probe recorded the characteristic cytosolic Ca2+ signals evoked by acute exposures to 20 mM glucose (Fig. 4B) as well as 100 μM palmitate and 25 mM KCl in MIN6 cells (Fig. 4C). We also confirmed the involvement of both intracellular stores and extracellular Ca2+ in palmitate signaling (Fig. 4, D and E). Together, these experiments indicate that D3cpv is an effective Ca2+ probe for β-cells and confirm the role of both extracellular influx and intracellular stores in palmitate-induced Ca2+ signals.
Long-term analysis of Ca2+ signals with cameleons.
A major advantage of the fluorescent protein-based Ca2+ indicators is their utility in long-term experiments. We examined cytosolic and ER Ca2+ responses to palmitate in MIN6 cells over multiple hours. Interestingly, palmitate-induced cytosolic Ca2+ oscillations were sustained and did not appear to diminish appreciably with time in a large proportion of cells (Fig. 5A). This indicates that in the continuous presence of palmitate the fatty acid generates signals using components that do not desensitize within the time frame of these studies. We observed distinct groups of cellular responses to palmitate. About one-half of the cells responded immediately, and ∼75% of these exhibited sustained cytosolic Ca2+ oscillations (Fig. 5A). The other half of the cells did not begin responding with sustained cytosolic Ca2+ oscillations until they had been exposed to palmitate for an average of ∼50 min. Additionally, we examined ER Ca2+ on long-term palmitate stimulation (Fig. 5B). Typically, palmitate induced a release of ER luminal Ca2+ that included a gradual and oscillatory refilling of the organelle (Fig. 5B). Furthermore, these cells remained responsive to 100 μM carbachol, an IP3 agonist, after 100 min of palmitate treatment, indicating that SERCA pumps were still active and the ER Ca2+ store was not completely and permanently depleted. To the best of our knowledge, this is the first use of a protein-based indicator to examine extended Ca2+ signals in β-cells.
Effect of chronic treatment of palmitate on Ca2+ dynamics.
In previous studies of β-cells chronically treated with palmitate, indirect measurements of Ca2+ store loading have yielded conflicting results (10, 27). To directly test the hypothesis that chronic palmitate treatment alters ER Ca2+, we measured ER Ca2+ in D1ER-transfected MIN6 cells exposed to palmitate for 24 h. ER Ca2+ levels were significantly lower in the palmitate-treated cells compared with the vehicle control-treated cells (Fig. 6). The amount of stored Ca2+ that could be released on thapsigargin exposure was not significantly different, indicating that chronic palmitate treatment only caused a partial depletion. These results suggest that chronically high levels of palmitate may affect the ER Ca2+ “set point” and contribute to deregulation of β-cell Ca2+ homeostasis.
Effects of palmitate on ER stress and cell death.
Sustained elevations in intracellular Ca2+ are known to predispose β-cells to programmed cell death. Moreover, extended periods of lowered ER Ca2+ would be expected to promote the induction of ER stress and to promote cell death as well (34). Given the rapid effects of palmitate on luminal ER Ca2+, we tested the hypothesis that components of the ER stress response might be activated rapidly. Indeed, after just 5 min, palmitate caused a significant increase in the phosphorylation of the luminal calcium sensor PERK (Fig. 7A). Subsequently, palmitate caused a significant increase in both the activated, spliced (54 kDa) form of Xbp-1 (32) and the proapoptotic transcription factor Chop (Fig. 7, B and C), consistent with previous reports (8, 27, 28, 32). This was correlated with the kinetics of palmitate-induced propidium iodide incorporation, an indicator of β-cell death (Fig. 7D). Interestingly, the effect of 2-bromopalmitate on death was slower than that of palmitate. Together, these results suggest that both Ca2+ influx downstream of palmitate metabolism and ER Ca2+ store mobilization may contribute to ER stress and β-cell apoptosis induced by long-term palmitate treatment.
Effects of palmitate on insulin secretion.
Next, we examined the effect of palmitate on insulin secretion at multiple time points. Some previous studies have indicated that palmitate can significantly increase insulin release and that this effect may be glucose dependent (41). However, in our hands we found palmitate to be relatively ineffective at stimulating sustained insulin secretion from human islets or MIN6 cells (Fig. 8, A and C). This did not depend on whether human islets or MIN6 cells were cultured in 5 mM glucose or 25 mM glucose. Similarly, we did not observe significant effects of palmitate on insulin content in human islets (Fig. 8B), in contrast to previous studies employing rodent cells or cell lines. Together, these data indicate that palmitate-induced Ca2+ signals are not well coupled to insulin exocytosis, similar to our previous observations with insulin (35).
Effects of oleate on ER Ca2+ dynamics.
The monounsaturated fat oleate has been shown to affect β-cells in a number of studies, although it is generally considered to be less toxic (12, 15, 20, 27, 37, 38). Oleate is thought to activate GPR40 in β-cells (19, 54). Using D1ER-transfected MIN6 cells, we found that oleate caused a decrease in ER luminal Ca2+ that was similar to that of palmitate (Fig. 9A). We noted a trend that a higher percentage of cells responded to oleate (45 ± 8% for oleate, 25 ± 5% for palmitate, not statistically significant; Fig. 9B). This may be a dose-dependent effect, because the free concentration of oleate in equilibrium with its BSA-bound form is twice that of palmitate (51). We performed further experiments in which these fatty acids were added sequentially and again noted no difference in response to either fat after treatment (Fig. 9, C–E). These results indicate that palmitate does not block the cell's ability to respond to oleate, and that oleate also does not seem to impair the ability of palmitate to signal.
The present study was undertaken to elucidate the effects of palmitate on pancreatic β-cell Ca2+ homeostasis. Our study has three main findings. First, we have shown that palmitate causes cytosolic Ca2+ signals in primary human islet β-cells and MIN6 cells, in agreement with previous findings in other β-cell lines and primary mouse islets (19, 22). Second, we have demonstrated that these Ca2+ signals are generated by a combination of extracellular Ca2+ influx and Ca2+ release from intracellular stores. We provide the first direct evidence that palmitate and oleate induce Ca2+ release from the ER lumen. Third, we have characterized the time course of the actions of palmitate on Ca2+, ER stress, programmed cell death, and insulin secretion. Together, these findings contribute to our understanding of the mechanisms by which palmitate alters β-cell function and survival.
In this report, we have characterized the nature of the palmitate-induced Ca2+ signals in human β-cells and MIN6 cells. The Ca2+ responses were qualitatively and quantitatively similar between the two cell types, indicating that the MIN6 cell line may be a suitable model for studying some aspects of β-cell Ca2+ homeostasis. A recent study also described the similarities between MIN6 cells and human β-cells in terms of their susceptibility to apoptosis in response to palmitate (31), but to our knowledge no study has directly compared the two cell types with respect to their Ca2+ responses. Using MIN6 cells, we demonstrate that even low doses of palmitate (e.g., 25 μM) can effectively stimulate Ca2+ signals. Palmitate is bound to carrier proteins (serum albumin), but it is the free, unbound form of palmitate that is active (23). Our solutions containing 0.1% BSA would be expected to have nanomolar concentrations of unbound palmitate (51), which is within the normal physiological range for plasma free fatty acids (57). Using the D3cpv cameleon, we also provide the first evidence for sustained, oscillatory cytosolic Ca2+ signals generated by fatty acids. These experiments also validate this probe as a reliable method for measuring β-cell Ca2+, especially in longer experiments.
Our work sheds new light on the mechanisms behind the palmitate-induced Ca2+ signals. Although some investigators have reported that fatty acids do not reduce Ca2+ stores (19, 27), other reports have suggested the possible role of both extracellular Ca2+ influx and intracellular store release (6, 49). Our data from both primary human islet β-cells and MIN6 cells support the latter model. Importantly, we imaged palmitate-induced ER depletion directly, using the D1ER cameleon probe. GPR40 might be expected to mobilize ER Ca2+ stores via Gqα (3, 23), phospholipase C (19), and IP3R-gated ER Ca2+ release. Our experiments were not able to define a major role for IP3Rs, and they leave open the possibility that palmitate may affect ryanodine receptors or the SERCA pumps.
Interestingly, our study found that oleate also transiently depleted ER Ca2+, in agreement with previous studies showing effects of this fatty acid on β-cell calcium homeostasis (19, 54). Oleate treatment did not block the ability of subsequent palmitate to induce ER Ca2+ release, suggesting that the interaction of each fatty acid with its receptor may be transient. A study in rat α-cells, which also express GPR40 (17), demonstrated that blocking ER stores abolished oleate-induced cytosolic Ca2+ signals (18). RNA interference-mediated GPR40 knockdown in rat islets (19) and INS-1 cells (54) also abolished oleate-stimulated Ca2+ signals. However, oleate-stimulated Ca2+ signals in cells expressing L-type calcium channels but not GPR40 were recently reported (58). Palmitate's effects on L-type calcium channels in mouse islets have also been documented (42). These studies, in concert with ours, suggest that fatty acids, as well as their metabolites, may interact with more than one Ca2+ signaling pathway.
The idea that palmitate can target intracellular Ca2+ stores is supported by the results with the nonmetabolizable palmitate analog 2-bromopalmitate, which is esterified to 2-bromopalmitate CoA but cannot undergo β-oxidation. It is thought that 2-bromopalmitate prevents other fatty acids from undergoing β-oxidation by blocking carnitine palmitoyltransferase I (CPT I), which transports long-chain free fatty acids into the mitochondria (50). Treatment of human islets and MIN6 cells with 2-bromopalmitate induced an attenuated Ca2+ signal. Because 2-bromopalmitate cannot be fully metabolized, this small Ca2+ signal may be the result of partial activation of GPR40, which binds a variety of fatty acids (3, 23, 30). Nevertheless, the effects of 2-bromopalmitate on GPR40 activation remain controversial (47, 53). We did not consistently detect marked decreases in ER luminal Ca2+ on pretreatment with 2-bromopalmitate in a limited number of preliminary experiments. A previous report found that concurrent treatment of palmitate with 2-bromopalmitate abolished palmitate-induced Ca2+ signals (49), indicating that 2-bromopalmitate may inhibit the action of palmitate. The mechanism of this inhibition remains to be resolved. Another report noted that treatment of mouse islets with palmitoyl CoA, the esterified metabolite of palmitate, did not mimic palmitate's actions on islet capacitance and Ca2+ currents (42), although direct Ca2+ signaling was not measured.
In the present study, we observed substantial heterogeneity in Ca2+ signals induced by palmitate. This is similar to the well-described response heterogeneity seen with glucose (29, 59–61) and other agonists (26). It is possible that a higher percentage of MIN6 cells might have responded at increased glucose levels. Previous reports have also suggested a glucose dependence of fatty acid-induced Ca2+ signals in mouse (49) and rat (19) β-cells. Remizov et al. (49) suggested that this may be related to increased mitochondrial activity and inhibition of CPT I at stimulatory glucose levels, creating a backlog of cytosolic esterified fatty acids that might then become available for signaling.
We also examined the effects of palmitate on ER stress, cell death, and insulin secretion in MIN6 cells. As dedicated, high-capacity secretory cells, pancreatic β-cells are particularly susceptible to ER stress (39). Previous studies have documented an effect of palmitate on ER stress in β-cells (6, 27, 28, 31), and we confirm those findings here. Furthermore, ER Ca2+ levels are suggested to be critical in the ER stress response (34, 64), and a recent report demonstrated that palmitate-induced CHOP induction was partially mediated by extracellular Ca2+ influx (6). We recently showed (24) that a component of the palmitate-induced β-cell death is due to the rapid Ca2+-dependent degradation of carboxypeptidase E. The present study adds direct ER Ca2+ depletion to the many insults caused by palmitate. The extent of this ER Ca2+ release was less than half of that caused by the ER stress inducer thapsigargin (34). However, since high luminal Ca2+ must be maintained for proper protein folding in the ER, the palmitate-induced depletion could be expected to exacerbate ER stress. ER Ca2+ levels are important in proinsulin processing and transport (21). Indeed, we observed phosphorylation of the ER Ca2+ sensor PERK after 5 min, indicating a rapid response to the palmitate-induced Ca2+ release. PERK is also rapidly activated when the MIN6 cell ER is directly depleted by blocking SERCA pumps with thapsigargin (34). Our observation that chronic palmitate treatment decreased in ER Ca2+ levels is consistent with a recent report and may provide an additional link between ER Ca2+ deregulation and cell survival (10). However, it is clear that a modest Ca2+ depletion alone cannot account for β-cell ER stress and cell death. Indeed, chronic palmitate treatment has deleterious effects on β-cell function (8, 27, 31, 32, 48), including the impairment of oscillatory fuel-induced Ca2+ responses (22). We hypothesize that altered Ca2+ homeostasis is an important part of the lipotoxic pathway, but that modest ER Ca2+ depletion is not sufficient to cause death on its own. Blocking Ca2+ influx protects rat islets and cultured cells from palmitate-induced cell death (6, 24), but blocking IP3Rs did not exert a protective effect (6). Furthermore, we have demonstrated that oleate stimulation creates similar ER Ca2+ responses, and several groups have shown that this unsaturated fatty acid is indeed less toxic than palmitate (12, 15, 24, 27, 38). The observation that 2-bromopalmitate induced a less rapid rise in cell death compared with palmitate suggests that multiple aspects of palmitate signaling are required for the full cytotoxic effects of this fatty acid. Together, these findings indicate that palmitate has multiple deleterious effects on pancreatic β-cells. We suggest a model in which chronic palmitate-induced ER Ca2+ depletion perturbs ER Ca2+ homeostasis, contributing to defects in proinsulin processing (24), and eventually ER stress. This model can help explain β-cell apoptosis in the context of hyperlipidemia, in cases of peripheral insulin resistance or hyperglycemia where there is an increased demand on insulin biosynthesis.
Fatty acids acutely stimulate insulin secretion in several model systems (41). However, in contrast to previous reports in rodent islets and cultured mouse β-cells (1, 11, 62), palmitate was not a robust insulin secretagogue in our hands. Similarly, palmitate treatment stimulated only a modest acute insulin response in perifused human islets, regardless of whether they were examined in high or low glucose (24). Thus we were unable to confirm studies demonstrating either a palmitate-induced decrease in insulin content or a robust glucose-dependent increase in insulin secretion (47, 63). Together, these data imply that the Ca2+ signals generated by palmitate are only weakly coupled to the insulin secretory apparatus in human β-cells and MIN6 cells. Thus human islets appear to be more similar to MIN6 cells than to mouse islets in this regard.
In summary, we have examined the effects of palmitate on Ca2+ signaling in human islet β-cells and MIN6 cells, using a combination of conventional and new live cell imaging approaches. We directly measured Ca2+ release from the β-cell ER in response to palmitate. This rapid depletion of ER Ca2+ may contribute to palmitate-induced ER-stress in β-cells, but it is unlikely to play a major role in insulin exocytosis. Together, these data help define the mechanisms linking hyperlipidemia in obesity and β-cell failure in type 2 diabetes.
This research was supported by operating grants to J. D. Johnson from the Canadian Institutes of Health Research (CIHR), the Canadian Diabetes Association (CDA), and the Natural Sciences and Engineering Research Council (NSERC). J. D. Johnson is a CIHR New Investigator, a CDA Scholar, and a Michael Smith Foundation for Health Research (MSFHR) Scholar. K. S. Gwiazda is supported by training awards from CIHR and MSFHR.
We thank Dr. Dan Luciani for advice and expertise throughout the project. We thank Drs. Roger Tsien and Amy Palmer for the gift of the Ca2+-sensing cameleons D3cpv and D1ER. The technical assistance of Xiaoke Hu is acknowledged.
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 © 2009 the American Physiological Society