Group VIA Phospholipase A2 Mitigates Palmitate-induced β-Cell Mitochondrial Injury and Apoptosis*

Background: Lipid-induced β-cell loss contributes to type 2 diabetes mellitus (T2DM). Results: Palmitate-induced β-cell lipid oxidation, mitochondrial dysfunction, and apoptosis correlate inversely with expression of iPLA2β, which associates with mitochondria, generates monolysocardiolipin, and lowers oxidized phospholipid content. Conclusion: iPLA2β mitigates palmitate-induced β-cell mitochondrial injury and apoptosis and may facilitate repair of oxidized lipids. Significance: Understanding lipid-induced β-cell loss could lead to T2DM therapies. Palmitate (C16:0) induces apoptosis of insulin-secreting β-cells by processes that involve generation of reactive oxygen species, and chronically elevated blood long chain free fatty acid levels are thought to contribute to β-cell lipotoxicity and the development of diabetes mellitus. Group VIA phospholipase A2 (iPLA2β) affects β-cell sensitivity to apoptosis, and here we examined iPLA2β effects on events that occur in β-cells incubated with C16:0. Such events in INS-1 insulinoma cells were found to include activation of caspase-3, expression of stress response genes (C/EBP homologous protein and activating transcription factor 4), accumulation of ceramide, loss of mitochondrial membrane potential, and apoptosis. All of these responses were blunted in INS-1 cells that overexpress iPLA2β, which has been proposed to facilitate repair of oxidized mitochondrial phospholipids, e.g. cardiolipin (CL), by excising oxidized polyunsaturated fatty acid residues, e.g. linoleate (C18:2), to yield lysophospholipids, e.g. monolysocardiolipin (MLCL), that can be reacylated to regenerate the native phospholipid structures. Here the MLCL content of mouse pancreatic islets was found to rise with increasing iPLA2β expression, and recombinant iPLA2β hydrolyzed CL to MLCL and released oxygenated C18:2 residues from oxidized CL in preference to native C18:2. C16:0 induced accumulation of oxidized CL species and of the oxidized phospholipid (C18:0/hydroxyeicosatetraenoic acid)-glycerophosphoethanolamine, and these effects were blunted in INS-1 cells that overexpress iPLA2β, consistent with iPLA2β-mediated removal of oxidized phospholipids. C16:0 also induced iPLA2β association with INS-1 cell mitochondria, consistent with a role in mitochondrial repair. These findings indicate that iPLA2β confers significant protection of β-cells against C16:0-induced injury.


Palmitate (C16:0) induces apoptosis of insulin-secreting
␤-cells by processes that involve generation of reactive oxygen species, and chronically elevated blood long chain free fatty acid levels are thought to contribute to ␤-cell lipotoxicity and the development of diabetes mellitus. Group VIA phospholipase A 2 (iPLA 2 ␤) affects ␤-cell sensitivity to apoptosis, and here we examined iPLA 2 ␤ effects on events that occur in ␤-cells incubated with C16:0. Such events in INS-1 insulinoma cells were found to include activation of caspase-3, expression of stress response genes (C/EBP homologous protein and activating transcription factor 4), accumulation of ceramide, loss of mitochondrial membrane potential, and apoptosis. All of these responses were blunted in INS-1 cells that overexpress iPLA 2 ␤, which has been proposed to facilitate repair of oxidized mitochondrial phospholipids, e.g. cardiolipin (CL), by excising oxidized polyunsaturated fatty acid residues, e.g. linoleate (C18:2), to yield lysophospholipids, e.g. monolysocardiolipin (MLCL), that can be reacylated to regenerate the native phospholipid structures. Here the MLCL content of mouse pancreatic islets was found to rise with increasing iPLA 2 ␤ expression, and recombinant iPLA 2 ␤ hydrolyzed CL to MLCL and released oxygenated C18:2 residues from oxidized CL in preference to native C18:2. C16:0 induced accumulation of oxidized CL species and of the oxidized phospholipid (C18:0/hydroxyeicosatetraenoic acid)-glycerophosphoethanolamine, and these effects were blunted in INS-1 cells that overexpress iPLA 2 ␤, consistent with iPLA 2 ␤mediated removal of oxidized phospholipids. C16:0 also induced iPLA 2 ␤ association with INS-1 cell mitochondria, consistent with a role in mitochondrial repair. These findings indicate that iPLA 2 ␤ confers significant protection of ␤-cells against C16:0-induced injury.
Chronic elevation of free fatty acids (FFAs) 2 in blood and tissues, alone or combined with hyperglycemia, is associated with both insulin resistance and type 2 diabetes mellitus (1)(2)(3). Results from several laboratories suggest that islet accumulation of lipid is deleterious and eventuates in ␤-cell failure and death in a process designated "lipotoxicity," and FFAs have been shown to cause ␤-cell death by both apoptosis and necrosis (4 -7). Although the molecular and cellular mechanisms underlying FFA-induced ␤-cell apoptosis are not fully understood, participating processes include generation of reactive oxygen species (ROS) and mitochondrial dysfunction (8 -11), production of ceramide and nitric oxide (NO) (4,12), and induction of endoplasmic reticulum stress (13)(14)(15)(16)(17)(18).
The lipid-metabolizing enzyme Group VIA phospholipase A 2 (iPLA 2 ␤) plays signaling roles in insulin secretion, promotes ␤-cell proliferation, and affects responses to stimuli that induce apoptosis (19 -23). Here we examined palmitate-induced apoptosis of INS-1 insulinoma cells and of native pancreatic islet ␤-cells and found that ␤-cells with increased or reduced iPLA 2 ␤ activity have blunted or enhanced sensitivity, respectively, to palmitate-induced injury. The reduction in palmitateinduced apoptosis of INS-1 cells that overexpress iPLA 2 ␤ is associated with attenuation of the effects of palmitate to activate capase-3 and to cause collapse of the mitochondrial membrane potential (⌬⌿ m ). Incubation of ␤-cells with palmitate was found to result in subcellular redistribution of iPLA 2 ␤ and its association with mitochondria where it appears to participate in remodeling oxidized phospholipid species, including cardiolipin.
Generation of Genetically Modified Mice and Wild-type Littermates-All animal protocols were approved by the Washington University Animal Studies Committee. Preparation and characterization of global iPLA 2 ␤ knock-out (KO) mice (25,26), transgenic (TG) mice that overexpress iPLA 2 ␤ in pancreatic islet ␤-cells (27), and their wild-type littermates on a C57BL/6 genetic background have been described previously as have genotyping procedures for these mice (25)(26)(27).
Islet Isolation-Islets were isolated from minced pancreata of male mice by collagenase digestion followed by Ficoll step density gradient separation and stereomicroscopic manual selection to exclude contaminating tissues as described (26 -28). Mouse islets were counted, and aliquots of homogenate were used for Coomassie protein determinations and other measurements.
Cell Culture-Preparation and properties of stably transfected iPLA 2 ␤-OE INS-1 rat insulinoma cells that overexpress iPLA 2 ␤, control INS-1 cells stably transfected with empty vector only, and iPLA 2 ␤ knockdown INS-1 cells in which iPLA 2 ␤ expression is knocked down by siRNA have been described previously (24,29,30). INS-1 cell lines were cultured as described (24) in RPMI 1640 medium containing 11 mM glucose, 10% fetal calf serum, 10 mM HEPES buffer, 2 mM glutamine, 1 mM sodium pyruvate, 50 mM ␤-mercaptoethanol, 100 units/ml penicillin, and 100 g/ml streptomycin. Medium was replaced with fresh medium every 2 days, and cell cultures were divided once weekly. Cells were grown to 80% confluence and harvested after treatment described in the figure legends. All incubations were performed at 37°C under 95% air, 5% CO 2 .
Detection of Apoptosis by Annexin V-FLUOS Staining-INS-1 cell apoptosis was determined as described (20,22,23) by measuring phosphatidylserine externalization in early apoptosis using an Annexin V-FLUOS staining kit (Roche Applied Science) to stain cells with fluorescein isothiocyanate-conjugated Annexin V according to the manufacturer's protocol in medium that also contained propidium iodide, which stains late stage apoptotic and necrotic cells. Briefly, about 10 6 cells were harvested, washed with PBS by centrifugation (200 ϫ g, 5 min), and resuspended in Annexin-V-FLUOS labeling solution (100 l). Cells were incubated (15 min, 20°C) and then analyzed by flow cytometry on a BD FACSCalibur (BD Biosciences) instrument at an excitation wavelength of 488 nm, and data were processed with WinMDI 2.9 software. Cells in early stage apoptosis were annexin V-positive and propidium iodide-negative, and those in late stage apoptosis were both annexin V-and propidium iodide-positive.
Assessment of Mitochondrial Membrane Potential by Flow Cytometry-Loss of ⌬⌿ m was measured as described (19,20) in INS-1 cells with a commercial kit according to the manufacturer's instructions (Cell Technology, Inc.). Briefly, cells were washed twice with phosphate-buffered saline, resuspended in JC-1 reagent solution (0.5 ml), incubated (37°C, 15 min), washed twice with PBS (1 ml), reconstituted in assay buffer (0.5 ml), and transferred to fluorescence-activated cell sorting tubes. Cellular fluorescence was analyzed with a BD FACSCalibur (BD Biosciences) flow cytometer in the FL2 channel.
Subcellular Fractionation to Isolate Mitochondria and Determine iPLA 2 ␤ Association-INS-1 cell mitochondria were separated from cytosol as described (22) with minor modifications. Briefly, isolation buffer (5 volumes; 20 mM HEPES-KOH, 100 mM KCl, 1.5 mM MgCl 2 , 1 mM EGTA, 250 mM sucrose) plus the protease inhibitor phenylmethylsulfonyl fluoride (1 mM) and protease inhibitor mixture (50 l/ml) were added to the cell pellet (20 min, on ice). Cells were then homogenized (Dounce apparatus, 20 strokes), and the homogenate was centrifuged (750 ϫ g, 5 min). The pellet containing any remaining intact cells and nuclei was discarded. Supernatant was centrifuged (10 5 ϫ g, 15 min) to remove mitochondria (pellet), and that supernatant was ultracentrifuged (10 6 ϫ g, 1 h). Alternatively, mitochondrial and cytosolic fractions were isolated from cells with a Mitochondria/Cytosol Fractionation kit (BioVsion Research Products) and centrifugation as described (19). Isolated mitochondria were sonicated (200 l of PBS, on ice), and aliquots of mitochondrial and cytosolic proteins were analyzed by SDS-PAGE and immunoblotted with antibodies to iPLA 2 ␤ (T-14, Santa Cruz Biotechnology) and the mitochondrial marker COX IV subunit II (Molecular Probes, Eugene, OR). Densitometric ratios of bands from immunoblots were determined with AlphaEaseFC software.
Immunostaining and Fluorescence Microscopic Analyses-As described (34,35), stably transfected INS-1 cells that expressed an iPLA 2 ␤-GFP construct were cultured on glass coverslips (18-mm diameter) in 6-well plates. After treatment, cells were stained with MitoTracker Red (Invitrogen, M7512) according to the manufacturer's protocol. Coverslips were then removed from the plate and sealed with a drop of ProLong Gold antifade reagent on glass slides that were examined with a Nikon TE300 microscope.
In Situ Detection of DNA Cleavage by TUNEL and DAPI Staining-TUNEL assays were performed essentially as described (21,36). In brief, after treatment, INS-1 cells were washed twice with ice-cold PBS, immobilized on slides by cytospin, fixed (4% paraformaldehyde in PBS (pH 7.4) (1 h, room temperature), washed with PBS, and incubated in permeabilization solution (0.1% Triton-X-100 in 0.1% sodium citrate, PBS; 30 min, room temperature). That solution was then removed, TUNEL reaction mixture (50 l) was added, and cells were incubated (1 h, 37°C) in a humidified chamber, washed again with PBS, and counterstained with DAPI (1 g/ml in PBS, 10 min) to identify nuclei. The incidence of apoptosis was assessed under a fluorescence microscope (Nikon Eclipse TE300) using a fluorescein isothiocyanate filter. Cells with TUNEL-positive nuclei were considered apoptotic. DAPI staining was used to determine the total number of cells in a field.
Preparation of Recombinant Group VIA PLA 2 with an N-terminal Polyhistidine Tag and a C-terminal FLAG Tag-Recombinant lentivirus containing cDNA encoding rat pancreatic islet iPLA 2 ␤ with an N-terminal polyhistidine tag and a C-terminal FLAG tag were prepared, and the recombinant virus was used to achieve stable transfection of INS-1 cells that expressed the fusion protein as described (37). Recombinant His-iPLA 2 ␤-FLAG was purified by immobilized metal affinity chromatography on cobalt-based TALON columns as described (34,38).

Measurement of ROS Production by INS-1 Insulinoma
Cells-Intracellular ROS production was measured with an OxiSelect TM Intracellular ROS Assay kit according to the manufacturer's instructions (Cell Biolabs, Inc., San Diego, CA). The assay principle is that ROS oxidize 2Ј,7Ј-dichlorodihydrofluorescin to fluorescent 2Ј,7Ј-dichlorofluorescein. Fluorescence is then measured with a plate reader using excitation and emission wavelengths of 480 and 530 nm, respectively, and quantitation is performed relative to an eight-point 2Ј,7Ј-dichlorofluorescein standard curve (0 -10,000 nM).

HPLC/ESI/MS/MS Analysis of 4-HNE from INS-1 Cells-
Analysis of 4-HNE was performed essentially as described (39). Briefly, extracted INS-1 cell phospholipids were mixed with d 3 -4-HNE internal standard and concentrated to dryness under N 2 . Saturated DNPH solution (0.5 ml) containing 1 N HCl was added to the residue and incubated in the dark (2 h, room temperature). DNPH derivatives were extracted twice with CH 2 Cl 2 and concentrated to dryness under N 2 . Samples were reconstituted with isopropanol/acetonitrile/water (65:50:5, v/v/v) and analyzed by LC/ESI/MS/MS on a Thermo Finnegan TSQ Quantum Vantage mass spectrometer equipped with a Finnigan Surveyor Plus pump. Reverse-phase HPLC was performed on a Sigma Acentis Express C 8 column (150 ϫ 2.1 mm, 5 m) with a solvent gradient over 30 min from 32 to 97% solvent B (90% isopropanol, 10% acetonitrile, 10 mM ammonium formate) and from 68 to 3% solvent A (60% acetonitrile, 10 mM ammonium formate). Selected reaction monitoring was performed in negative ion mode. Collisionally activated dissociation of 4-HNE-DNPH was optimized at 24 eV in argon (1.0 millitorr). Transitions m/z 335 to m/z 182 and m/z 338 to m/z 182 were monitored for 4-HNE-DNPH and d 3 -4HNE-DNPH, respectively.

Cardiolipin and Monolysocardiolipin Analyses by ESI/MS/MS-
was added to extracted lipids, and the mixture was concentrated, reconstituted, and infused into the ion source of an LTQ-Orbitrap Velos mass spectrometer (ThermoElectron) operated at a resolution of 30,000 with a maximum injection time of 50 ms (40). Alternatively, lipid extracts containing cardiolipin and/or monolysocardiolipin were analyzed by  ). An asterisk (*) denotes p Ͻ 0.05 for the difference between the BSA and palmitate conditions, and an x denotes a significant difference between wild type and knock-out. In A, islets from wild type, iPLA 2 ␤-null (KNOCKOUT), and transgenic mice that overexpress iPLA 2 ␤ in ␤-cells were incubated (24 h) in buffer that contained 1% BSA without (CONTROL; lightly stippled bars) or with 1 mM palmitate (wild type, finely cross-hatched bars; knock-out, solid black bars; or transgenic, coarsely cross-hatched bars). RNA was then isolated, and cDNA was generated by RT-PCR. CHOP (left set of bars) and ATF4 (right set of bars) mRNA levels were measured by quantitative PCR and expressed as the ratio of the value for palmitate-treated cells divided by that for cells incubated in BSA buffer alone. Mean values ϮS.E. (error bars) are indicated (n ϭ 4). In A, an asterisk (*) denotes a significant (p Ͻ 0.05) increase over control, and a plus (ϩ) denotes a significant difference from wild type. In B, iPLA 2 ␤-OE or vector-only INS-1 cells were incubated (6 or 16 h) in buffer that contained 1% BSA without (CONTROL) or with 1 mM palmitate (solid black bars for vector-only and crosshatched bars for iPLA 2 ␤-OE cells). CHOP and ATF4 mRNA levels were measured by quantitative PCR and expressed as a ratio of values for palmitatetreated cells divided by that for cells incubated in BSA-buffer alone. Mean values ϮS.E. (error bars) are indicated (n ϭ 4). An asterisk (*) denotes a significant (p Ͻ 0.05) increase over control, and a plus (ϩ) denotes a significant difference between vector-only and iPLA 2 ␤-OE cells. Inset immunoblots illustrate relative iPLA 2 ␤ expression levels in WT, KO, and TG islets (A) and vectoronly and iPLA 2 ␤-OE INS-1 cells (B), respectively.
LC/MS(/MS) in a manner similar to that described previously (41) on a Surveyor HPLC (ThermoElectron) using a modified gradient (42) on a C 8 column (15 cm ϫ 2.1 mm; Sigma) interfaced with the ion source of a ThermoElectron Vantage triple quadruple mass spectrometer with extended mass range operated in negative ion mode.
Preparation of Oxidized Cardiolipin-As described (39), standard (C18:2) 4 -CL was dissolved in chloroform in a glass vial and concentrated to dryness under nitrogen. PBS (100 l; 50 mM, pH 7.4) with 100 M diethylene triamine pentaacetic acid was then added, and the lipid mixture was Vortex-mixed and sonicated (10 min, under N 2 , in water). Cytochrome c (20 l; 200 M) and H 2 O 2 (20 l; 250 M) were then added, and the mixture was incubated (37°C, under air, 1 h). During the incubation, H 2 O 2 was added at 15-min intervals (final concentration, 100 M). Lipid extraction was performed as above, and concentrated extracts were reconstituted (chloroform/methanol, 1:1, v/v; 200 l) and analyzed by ESI/MS on a ThermoElectron TSQ Vantage triple quadrupole mass spectrometer in negative ion mode.
Cardiolipin Hydrolysis by iPLA 2 ␤-Purified recombinant iPLA 2 ␤ (2 g) was added to hydrolysis buffer (50 l; 200 mM Tris (pH 7.5) 5ϫ, 20 mM EGTA) and diluted with homogenization buffer (0.25 M sucrose, 40 mM Tris (pH 7.5)) to achieve a 200-l final volume. Standard (C18:2) 4 -CL or oxidized CL in ethanol (5 l) was then added, and the mixture was Vortex- Statistical Analyses-Results are presented as mean Ϯ S.E. Data were evaluated by unpaired, two-tailed Student's t test for differences between two conditions or by analysis of variance with appropriate post hoc tests for larger data sets (22). Significance levels are described in figure legends, and a p value Ͻ0.05 was considered significant.

Palmitate Induces Caspase-3 Activation and Other Markers of Cellular Injury in INS-1 Cells and in Mouse Pancreatic Islets, and the Magnitudes of These Effects Vary Inversely with iPLA 2 ␤
Expression Level-Proteolytic activation of caspase-3, which is a key protease in the execution of apoptosis via the mitochondrial pathway (43), is reflected by generation of the active p17 product from its p32 precursor, and incubation with palmitate induced time-dependent activation of caspase-3 in INS-1 insulinoma cells stably transfected with vector only (Fig. 1A). Such activation was not observed with iPLA 2 ␤-OE INS-1 cells that overexpress iPLA 2 ␤ (Fig. 1B)  described previously (24,29), and the differences in expression levels are illustrated in Fig. 2B, inset.) Caspase-3 activation was also examined by a more sensitive bioluminescence assay (31) in mouse pancreatic islets, which can be obtained in only limited quantities, and incubating iPLA 2 ␤-KO islets with palmitate resulted in caspase-3 activation that was significantly greater than that for wild-type (WT) islets (Fig. 1C). (Generation and properties of KO mice have been described previously (25)(26)(27), and lack of iPLA 2 ␤ expression is illustrated in Fig. 2A, inset.) Both INS-1 cell and islet data thus indicate that higher iPLA 2 ␤ expression tends to confer protection against palmitate toxicity as reflected by caspase-3 activation. Consistent with previous reports (11, 13, 44 -49), incubation with palmitate also caused other manifestations of cellular injury in mouse pancreatic islets and INS-1 insulinoma cells (Fig. 2). These included accumulation of mRNA encoding ATF4 and CHOP (Fig. 2), which are transcription factors involved in endoplasmic reticulum stress-induced apoptosis, and accumulation of ceramide (supplemental Fig. S1), which is a mediator of fatty acid toxicity that inflicts mitochondrial injury and accumulates in ␤-cells undergoing apoptosis (4, 21, 50 -52). Each of these responses was attenuated by overexpression of iPLA 2 ␤ in INS-1 cells and amplified by iPLA 2 ␤ deletion ( Fig. 2 and supplemental Fig. S1), consistent with the protective effect of iPLA 2 ␤ in palmitate-induced ␤-cell injury suggested by caspase-3 activation data (Fig. 1).
Overexpression of iPLA 2 ␤ Reduces the Sensitivity of INS-1 Cells to Palmitate-induced Apoptosis-Caspase-3 activation is a harbinger of apoptosis, and to determine whether iPLA 2 ␤ expression level affects palmitate-induced apoptosis of ␤-cells, the extent of apoptosis of INS-1 insulinoma cells incubated with palmitate was assessed by measuring phosphatidylserine externalization by apoptotic cells. iPLA 2 ␤-OE INS-1 cells that overexpress iPLA 2 ␤ were compared with cells transfected with empty vector only that express the lower levels of iPLA 2 ␤ of the parental cell line. Cells were incubated with palmitate or vehicle and then treated with Annexin-FITC to impart fluorescence to cells that had externalized phosphatidylserine. The extent of apoptosis was then determined by flow cytometry as illustrated in Fig. 3A where the apoptotic cell populations are represented in the regions labeled M1. Incubation with palmitate induced a time-dependent increase in the percentage of cells that had undergone apoptosis, and this was significantly lower for iPLA 2 ␤-OE INS-1 cells than for control INS-1 cells after both 6 and 16 h of incubation with palmitate (Fig. 3B). These data indicate that increased expression of iPLA 2 ␤ by INS-1 cells confers a significant degree of protection against palmitate-induced apoptosis. (Similar results were obtained upon examination of cells that both bound Annexin and exhibited propidium iodide uptake (not shown), which reflects a late stage of apoptosis.)

Overexpression of iPLA 2 ␤ Attenuates Palmitate-induced INS-1 Cell Mitochondrial Membrane Potential
Loss-⌬⌿ m collapse is an early event in the apoptosis pathway that precedes phosphatidylserine externalization and coincides with caspase activation (53,54), and the percentage of INS-1 cells with ⌬⌿ m collapse increased upon incubation with palmitate as reflected by fluorescence-activated cell sorting (FACS) of cells loaded with potential-sensitive indicator JC-1 (Fig. 4). Loss of ⌬⌿ m is reflected by signal in the region designated M1 in Fig. 4, A-D, and after 16 h, that percentage was significantly lower for iPLA 2 ␤-OE INS-1 cells than for control INS-1 cells, indicating that increased iPLA 2 ␤ expression confers protection against palmitate-induced mitochondrial injury and loss of ⌬⌿ m .

iPLA 2 ␤ Undergoes Subcellular Redistribution upon Incubation of INS-1 Cells with Palmitate and Associates with
Mitochondria-If protection of ␤-cells from palmitate toxicity reflects mitigation of mitochondrial injury as suggested by Fig.  4, then association of iPLA 2 ␤ with mitochondria might be expected in cells incubated with palmitate. To examine this possibility, mitochondria were isolated from INS-1 cells, and their iPLA 2 ␤ content was determined by immunoblotting and compared with that of the mitochondrial protein COX IV (22). Fig. 5A illustrates that incubating INS-1 cells with palmitate induced a marked increase in immunoreactive iPLA 2 ␤ in the mitochondrial fraction relative to the BSA control, and there was a corresponding increase in Ca 2ϩ -independent PLA 2 enzymatic activity associated with mitochondria (Fig. 5B). (Note that much of the iPLA 2 activity associated with mitochondria  (Fig. 5C, second lane, lower panel). The fourth lane in Fig. 5C represents the merge of the first three lanes. It illustrates that the uniform olive hue of the extranuclear portion of cells incubated with BSA vehicle (Fig. 5C, fourth  lane, upper panel) is replaced upon incubation with palmitate (Fig. 5C, fourth lane, lower panel) by an image with a punctate distribution of yellow spots from colocalized green and red signals from GFP-iPLA 2 ␤ and MitoTracker, respectively, reflecting palmitate-induced iPLA 2 ␤ redistribution of from cytosol to mitochondria.
Because iPLA 2 ␤ undergoes proteolytic processing that affects its organellar association (35,38,56,57), a His-iPLA 2 ␤-FLAG fusion protein with N-terminal polyhistidine and C-terminal FLAG tags was expressed in INS-1 cells that were incubated with palmitate and then disrupted. Mitochondria were isolated, and their proteins were analyzed by SDS-PAGE and immunoblotting with antibodies against polyhistidine or FLAG or against mitochondrial COX IV (Fig. 6A, left panel). Cell homogenate was similarly analyzed with antibodies against polyhistidine or FLAG, iPLA 2 ␤ internal sequence, or actin as controls (Fig. 6A, right panel). Mitochondrial immunoreactivity with anti-FLAG antibody at iPLA 2 ␤-FLAG fusion protein molecular weight increased substantially in INS-1 cells incubated with palmitate (Fig. 6A, left panel, middle blot), but little such signal was obtained with anti-polyhistidine antibody (Fig.  6A, left panel, top blot), although signal was obtained with fusion protein-expressing INS-1 cell homogenates (Fig. 6A,  right panel, top blot). Mitochondrial iPLA 2 ␤-FLAG immunoreactivity increased with palmitate incubation interval (Fig. 6B). This suggests that the form of iPLA 2 ␤ that C16:0 causes to associate with mitochondria is processed at the N terminus.
Palmitate Induces Phospholipid Oxidation in INS-1 Cells-Consistent with reports that palmitate toxicity in ␤-cells and other cells results from injury by ROS (8, 58 -60), incubating INS-1 cells with palmitate was found to result in increased ROS generation as reflected by 2Ј,7Ј-dichlorodihydrofluorescin oxidation to the fluorescent 2Ј,7Ј-dichlorofluorescein, and there was no significant difference between control and iPLA 2 ␤-OE INS-1 cells in that regard (Fig. 7A). Similarly, incubating INS-1 cells with palmitate induced a time-dependent rise in iNOS mRNA levels, consistent with reports that palmitate causes increased iNOS expression in ␤-cells and other cells (12,58,(61)(62)(63)(64), and there was no significant difference between control and iPLA 2 ␤-OE INS-1 cells in the magnitude of that effect (Fig. 7B). Both ROS and NO can induce lipid oxidation in mitochondrial and other cellular membranes (65)(66)(67)(68), and incubation with palmitate also induced a rise in INS-1 cell content of the lipid peroxidation product 4-HNE as demonstrated by a multiple reaction monitoring LS/MS/MS assay (Fig. 8). The magnitude of the rise in 4-HNE was inversely related to iPLA 2 ␤ expression level (Fig. 8B), consistent with the possibility that iPLA 2 ␤ excises oxidized linoleate residues from CL and that the yield of 4-HNE from linoleate oxidation is much greater for residues within tetralinoleoyl-CL ((18:2) 4 -CL) species because intramolecular radical addition between neighboring linoleate chains amplifies 4-HNE formation (39). ). An asterisk (*) denotes a significant (p Ͻ 0.05) difference between BSA and palmitate conditions, and an x denotes a significant difference between iPLA 2 ␤-OE and iPLA 2 ␤ knockdown cells.

DISCUSSION
Apoptosis contributes to ␤-cell loss in the development of type 2 diabetes mellitus, and fatty acid toxicity participates in these processes by incompletely understood mechanisms. Generation of ROS by mitochondria has also been proposed to contribute to palmitate toxicity to cells (59), including ␤-cells (8 -10), and prolonged ␤-cell exposure to high FFA concentrations does cause ROS production (8,49,77). Palmitate is reported to cause ␤-cell apoptosis by affecting fission and fusion of mitochondrial membranes (78,79), and iPLA 2 ␤ also affects mitochondrial membranes (19,20,22) and is proposed to participate in their repair from oxidative injury inflicted by ROS (73,74). Our findings indicate that palmitate-induced mitochondrial injury in ␤-cells as reflected by collapse of ⌬⌿ m is mitigated by overexpressing iPLA 2 ␤ in INS-1 cells, and iPLA 2 ␤ may thus act to restrain the mitochondrial pathway of apoptosis.
Association of iPLA 2 ␤ with mitochondria could mitigate injury from ROS by repairing oxidized mitochondrial phospholipids, such as CL. CL is critical for mitochondrial function and retention of cytochrome c, a mobile carrier required for electron transfer chain function (80 -87). Interaction with cytochrome c depends upon CL C18:2 substituents, and a decline in inner mitochondrial membrane CL content or alteration of its C18:2 substituents resulting from defective CL remodeling or oxidative modification (84, 87) diminishes cytochrome c mem-brane affinity (87,88). The resultant release of cytochrome c into the cytosol can precipitate apoptosis (86). iPLA 2 ␤ may participate in generating and maintaining the C18:2-rich composition of CL under physiologic and pathophysiologic conditions. Newly synthesized CL must be remodeled to produce mature C18:2-rich CL, and the mitochondrial enzymes MLCL acyltransferase and tafazzin appear to cooperate with a PLA 2 in this process. To remodel nascent CL, the substrate MLCL must be generated so that reacylation with linoleate can occur, and a Group VI PLA 2 , e.g. iPLA 2 ␤, iPLA 2 ␥, or both (55, 89 -92), may catalyze MLCL formation. iPLA 2 ␤ may play a role in repairing oxidized CL (19, 20, 70 -72, 75) under pathophysiological circumstances that is similar to its proposed role in physiological CL remodeling. C18:2 residues are especially susceptible to oxidation because they contain bisallylic methylene moieties with a labile hydrogen atom that can be abstracted to yield a carbon-centered radical that readily reacts with molecular oxygen to form a fatty acid hydroperoxide (93). Oxidization reduces hydrophobicity of the fatty acid substituent and allows it to approach the hydrophilic phospholipid headgroup more closely (93). This increases separation between headgroups, causing the ester bond to be more accessible to PLA 2 .
Group VI PLA 2 enzymes may participate in such repair of oxidized mitochondrial membrane phospholipids (19, 20, 73, 94 -96). iPLA 2 ␤ localizes to mitochondria in insulinoma cells and protects against oxidant-induced apoptosis, and pancreatic islets from iPLA 2 ␤-null mice exhibit increased susceptibility to oxidant-induced apoptosis (19,20,73). Oxidant-induced lipid peroxidation and death of renal proximal tubule cells are potentiated by bromoenol lactone (95), which inhibits Group VI PLA 2 enzymes (97,98). This may reflect iPLA 2 ␤-catalyzed removal of oxidized PUFA residues from mitochondrial glycerophospholipids formed during oxidative stress. This would permit the resultant lysophospholipid to be reacylated with an unoxidized PUFA residue to restore functions impaired as a result of membrane oxidation. In the absence of iPLA 2 ␤ or when its activity is reduced, this repair mechanism would not be fully operative, and this could result in progressive mitochondrial injury that eventually triggers the mitochondrial pathway of apoptosis (90 -92). Conversely, increased iPLA 2 ␤ activity might confer increased resistance to oxidative injury that would otherwise result in apoptosis and that is consistent with the protection against palmitate-induced apoptosis of INS-1 cells conferred by overexpression of iPLA 2 ␤ demonstrated here.
Our findings indicate that ␤-cell MLCL content rises with increasing iPLA 2 ␤ expression level, which is compatible with a role for iPLA 2 ␤ in CL remodeling by excising oxidized PUFA residues from CL to yield MLCL species for reacylation with unoxidized C18:2-CoA to regenerate the native CL structure and function. This would stabilize the association of cytochrome c with mitochondrial membranes and mitigate ROS injury that would otherwise induce apoptosis (87). Our observations that the ␤-cell content of oxidized lipids rises after incubation with palmitate is consistent with the proposal that palmitate toxicity involves generation of ROS (8,12,49,59,63), and the correlation of these effects with iPLA 2 ␤ expression level is consistent with the possibility that iPLA 2 ␤ participates in an excision-reacylation repair mechanism for reducing membrane oxidized lipid content.
This proposed role for iPLA 2 ␤ in repair of oxidized phospholipids represents a special case of the originally proposed function of the enzyme in phospholipid remodeling (99 -102) and is consistent with the observations that oxidation of membranes accelerates iPLA 2 ␤-catalyzed fatty acid release from membranes and that iPLA 2 ␤ mediates oxidant-induced arachidonic acid release from cells (103)(104)(105). Moreover, iPLA 2 ␤ is active against phospholipids with short chain sn-2 substituents (106), such as those produced from polyunsaturated fatty acids by oxidation reactions (107). These are also properties of the Group VII platelet-activating factor acetylhydrolase enzymes (101,108), some of which have been have been proposed to function physiologically in clearance and/or repair of oxidized phospholipids (109). A similar role of a Group VI PLA 2 , such as iPLA 2 ␤, is plausible (109), and it is of interest in that regard that a plant analog of the mammalian Group VI PLA 2 enzymes designated patatin-containing phospholipase A (pPLAII␣) has been proposed to negatively regulate oxylipin production and to effect removal of oxidized fatty acids from the membranes of Arabidopsis thaliana (110).