Group VIB Phospholipase A2 Promotes Proliferation of INS-1 Insulinoma Cells and Attenuates Lipid Peroxidation and Apoptosis Induced by Inflammatory Cytokines and Oxidant Agents

Group VIB Phospholipase A2 (iPLA2γ) is distributed in membranous organelles in which β-oxidation occurs, that is, mitochondria and peroxisomes, and is expressed by insulin-secreting pancreatic islet β-cells and INS-1 insulinoma cells, which can be injured by inflammatory cytokines, for example, IL-1β and IFN-γ, and by oxidants, for example, streptozotocin (STZ) or t-butyl-hydroperoxide (TBHP), via processes pertinent to mechanisms of β-cell loss in types 1 and 2 diabetes mellitus. We find that incubating INS-1 cells with IL-1β and IFN-γ, with STZ, or with TBHP causes increased expression of iPLA2γ mRNA and protein. We prepared INS-1 knockdown (KD) cell lines with reduced iPLA2γ expression, and they proliferate more slowly than control INS-1 cells and undergo increased membrane peroxidation in response to cytokines or oxidants. Accumulation of oxidized phospholipid molecular species in STZ-treated INS-1 cells was demonstrated by LC/MS/MS scanning, and the levels in iPLA2γ-KD cells exceeded those in control cells. iPLA2γ-KD INS-1 cells also exhibited higher levels of apoptosis than control cells when incubated with STZ or with IL-1β and IFN-γ. These findings suggest that iPLA2γ promotes β-cell proliferation and that its expression is increased during inflammation or oxidative stress as a mechanism to mitigate membrane injury that may enhance β-cell survival.


Introduction
Diabetes mellitus (DM) is the most common human endocrine disease and is reaching pandemic proportions in the US and elsewhere [1]. DM represents a constellation of disorders that are grouped into the major categories types 1 and 2 (T1DM and T2DM). T1DM is caused by autoimmune destruction of insulin-secreting pancreatic islet β-cells [2,3], and inflammatory cytokines released by invading leukocytes during insulitis are believed to participate in these processes [4,5]. Among them are IL-1β, which impairs insulin secretion and inflicts islet injury [6], and IFN-γ, which greatly potentiates the destructive effects of IL-1β [7]. These effects are mediated in part by induction of nitric oxide (NO) synthase expression and overproduction of NO [8][9][10][11][12][13], which can induce apoptosis of cells by mechanisms that involve generation of reactive oxygen species that cause oxidative stress [13][14][15].
T2DM is thought to evolve after a period of initial insulin resistance in which nearly normal glucose tolerance is maintained by compensatory hypersecretion of insulin by βcells [16,17]. At some point there is failure to sustain insulin secretion at sufficiently high levels and glucose intolerance and then overt DM ensue [18]. One contributor to the eventual failure of β-cell compensation is a reduction in β-cell mass by 50% or more, and this occurs at least in part by apoptotic β-cell death [16,19]. Although many mechanisms probably participate in these processes, production of reactive oxygen species induced by metabolic stress has been proposed to represent a final common pathway of injury that ultimately results in β-cell failure [17,[20][21][22][23][24][25][26][27][28][29]. Among supporting observations are that β-cells express low levels of antioxidant defense enzymes compared to other tissues [30][31][32][33] and that antioxidant compounds confer protection from glucose toxicity to islets in vitro and against development of T2DM in animal models in vivo [34,35].
Beta cells must sustain a high level of metabolic activity, which provides critical signals in the coupling of nutrient sensing to insulin secretion [36][37][38][39][40][41][42][43][44][45][46][47][48][49][50], in order to meet the unceasing demand for insulin biosynthesis and processing. Prolonged overstimulation of β-cells may eventually contribute to their failure as a consequence of stresses imposed on the endoplasmic reticulum (ER) and mitochondria [51][52][53]. Protein synthesis, including that of proinsulin, occurs in ER, and nascent proteins must be properly folded, which involves formation of disulfide bonds. This is an oxidative reaction that requires a prooxidant environment to be maintained in ER. Sustained hyperstimulation can result in ER stress and overproduction of reactive oxygen species (ROS) that exceed ER reductive capacity, resulting in ROS leakage from ER and cellular oxidative stress [54][55][56][57].
A PLA 2 is suited for such a role because oxidized fatty acid substituents usually occur at the sn-2 position of phospholipids where most polyunsaturated fatty acid (PUFA) substituents, such as linoleate (C18:2) and arachidonate (C20:4), are esterified [73,74]. PUFA are especially susceptible to oxidation because they contain bis-allylic methylene moieties with a labile H atom that can be abstracted to yield a carbon-centered radical that readily reacts with molecular oxygen to form a fatty acid hydroperoxide [73]. Oxidization reduces hydrophobicity of the sn-2 fatty acid substituent and allows it to approach the hydrophilic phospholipid headgroup more closely [73]. This increases separation between head groups, which causes the sn-2 ester bond to be more accessible to PLA 2 . Liberated peroxy fatty acids can then be reduced to alcohols by glutathione peroxidases after release from phospholipids by PLA 2 enzymes [75,76].
The latter observations [70][71][72] suggest that iPLA 2 γ acts to reduce lipid peroxidation and to protect against oxidantinduced apoptosis in renal proximal tubule cells, and this may reflect iPLA 2 γ-catalyzed removal of oxidized PUFA residues from glycerophospholipids that are formed in mitochondria under conditions of oxidative stress. This could permit the resultant lysophospholipid to be reacylated with an unoxidized PUFA residue, which would restore functions that are impaired as a result of membrane oxidation. In the absence of iPLA 2 γ or when its activity is impaired, this repair mechanism cannot operate fully, and this could result in progressive mitochondrial injury that eventually triggers the mitochondrial pathway of apoptosis [70][71][72].

Cell
Culture. INS-1 rat insulinoma cells that had been stably transfected and mock-transfected INS-1 cells were generated and cultured 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 β mercaptoethanol, 100 units/mL penicillin, and 100 μg/mL streptomycin, essentially as previously described [80]. The medium was exchanged every 2 days, and the cell cultures were split once a week. Cells were grown to 80% confluence and harvested after treatment as indicated in the figure legends or the text of the Results section. All incubations were performed at 37 • C under an atmosphere of 95% air/5% CO 2 .

Determination of INS-1 Cell Proliferation
Rate. INS-1 cell proliferation rates were measured by two approaches, as previously described [80]. One assay is based on fluorescence enhancement when CyQuant GR binds to nucleic acids, which reflects the amount of cell DNA [82]. Cells were seeded onto 96-well plates (3 × 10 3 cells/well). Medium was removed after 1 or 3 days, and cells were frozen (−20 • C). DNA was measured with a CyQuant assay kit (Molecular Probes, Inc., Eugene, OR, USA) with reference to a standard curve. CyQuant GR solution (200 μL) was added to each well and incubated (5 min, room temperature). Fluorescence was measured on a microplate fluorimeter (excitation, 480 nm; emission, 538 nm). A second assay is based on incorporation of thymidine analog 5-bromo-2 -deoxyurindine (BrdU) into DNA in proliferating cells [83]. Cells were seeded (10 4 cells/ well) and cultured (3 days) before assay with an enzymelinked immunoassay detection kit III (Roche Applied Science) after BrdU labeling.
2.6. Lipid Peroxidation. Lipid peroxidation was quantitated using a Cayman TBARS assay kit (Cayman Chemical, Ann Arbor, MI, USA) according to the manufacturer's instructions, as previously described [64,84]. Lipid peroxides derived from polyunsaturated fatty acids decompose to form a complex series of compounds that include reactive carbonyl species, such as MDA. Measurement of thiobarbituric acid reactive substances (TBARS) by determining absorbance at 530 nm is used to assess the extent of lipid peroxidation [84]. Results are expressed as μmol/μg protein.

HPLC-ESI-MS/MS Analysis of Oxidized Lipids.
Lipids extracted from INS-1 cells were stored in sealed vials (under N 2 at −20 • C) to suppress artifactual oxidation, and extracts were then analyzed by LC/MS/MS in a manner similar to that previously described [85] on a Surveyor HPLC (Thermo-Electron, San Jose, CA, USA) using a modified gradient [86] on a C8 column (15 cm × 2.1 mm, Sigma Chemical Co., St. Louis, MO, USA) interfaced with the ion source of a Thermo-Electron Vantage triple quadruple mass spectrometer with extended mass range operated in negative ion mode. Tandem MS scans for precursors of m/z 295, m/z 319, and m/z 343 were performed to identify glycerolipid molecular species that contained singly oxygenated forms of the polyunsaturated fatty acids (PUFA) linoleate (C18:2), arachidonate (C20:4), or docosahexaenoate (C22:6), respectively. The major oxylipid species identified was (1-stearoyl, 2-hydroxyeicosatetraenoyl)-sn-glycerophospho-ethanolamine [(C18:0/HETE)-GPE], and it was quantified by MRM of 782.76 → 319.3, which is a transition that corresponds to production of the HETE carboxylate anion from the [M-H] − ion of the parent oxy-phospholipid species.

Assessment of Apoptosis by Flow Cytometry.
INS-1 cell apoptosis was determined using an Annexin-VFLUOS Staining Kit (Roche Applied Science, Indianapolis, IN, USA) according to the manufacturer's instructions, essentially as previously described [64,87]. Briefly, harvested cells were washed with PBS and resuspended in Annexin-VFLUOS labeling solution (100 μL). After incubation (10-15 min, 15-25 • C), cells were transferred to fluorescence-activated cell sorting (FACS) tubes and diluted 1 : 5 with buffer provided in the kit. Fluorescence in cells was analyzed with a FACscan flow cytometer (BD Biosciences, Sparks, MD, USA) at an excitation wavelength of 488 nm, and data were processed with WinMDI 2.9 software.

Statistical Analyses.
Results are expressed as mean ± SEM. Data were evaluated by unpaired, two-tailed Student's t-test for differences between two conditions or by analysis of variance with appropriate posthoc tests for larger sets, as previously described [80,81,87]. Significance levels are described in the figure legends, and a P value <0.05 was considered to reflect a significant difference.

INS-1 Cell iPLA 2 γ Expression and the Influence of Inflammatory Cytokines and Oxidative
Agents. INS-1 insulinoma cells were found to express iPLA 2 γ mRNA and iPLA 2 γimmunoreactive protein by quantitative PCR and by Western blotting, respectively (Figure 1), and also to exhibit iPLA 2 activity (not shown). Incubation with the inflammatory cytokines IL-1β and IFN-γ resulted in increased INS-1 cell expression of iPLA 2 γ mRNA in a concentration-dependent manner ( Figure 1(a)), and expression of iPLA 2 γ immunoreactive protein exhibited a similar pattern (Figure 1(b)).
The major iPLA 2 γ-immunoreactive band ( Figure 1(b)) migrated with an apparent MW of about 74 kDa on SDS-PAGE and Western Blotting analyses of INS-1 cell lysates, but other bands of variable intensity were observed with apparent MW between 40 and 88 kDa (not shown), as previously reported [88,89]. Although full-length iPLA 2 γ cDNA encodes an 88 kDa protein, transcriptional and translational regulatory mechanisms result in production of multiple gene products of various sizes, and the major species often migrate with apparent MW of 74-77 kDa [88,89].
The increased expression of iPLA 2 γ mRNA induced by IL-1β and IFN-γ was dependent on the incubation time ( Figure 1(c)). Incubation of INS-1 cells with the oxidative agents streptozotocin (STZ) [79] (Figure 2 These findings suggested the possibility that upregulation of iPLA 2 γ expression might represent a compensatory response to injurious agents in order to enhance β-cell survival in the settings of inflammation or oxidative stress. To examine this possibility, effects of suppressing INS-1 cell iPLA 2 γ expression were examined.

Establishing iPLA 2 γ-Knockdown INS-1 Cell
Lines. INS-1 cells were infected with FIV constructs containing inserts that produced either scrambled RNA (control) or shRNA directed against sequences in iPLA 2 γ mRNA. Selection of neomycin-resistant cells resulted in isolation of two clones that had stably incorporated knockdown constructs and expressed less than 20% of the control cell iPLA 2 γ mRNA content when analyzed by real-time PCR (Figure 3(a)) or Northern blots (not shown) and reduced amounts of iPLA 2 γ immunoreactive protein on Western blots (Figure 3(b)). The iPLA 2 γ-knockdown (iPLA 2 γ-KD) cell lines also exhibited a reduction in iPLA 2 activity (not shown) that was comparable in magnitude to the reduction in mRNA levels (Figure 3(a)). The level of iPLA 2 γ expression was a stable property of control and iPLA 2 γ-KD INS-1 cell lines that persisted on serial passage in culture.

INS-1 Cell Line Proliferation
Rates. Cell proliferation was measured using an indicator that exhibits strong fluorescence enhancement upon association with nucleic acids [82]. Identical numbers of cells of each INS-1 cell line were seeded at time 0, and their growth rates were monitored for 72 hr. INS-1 iPLA 2 γ-KD lines proliferated at rates that were significantly lower than those for control INS-1 cells (Figure 4). Similar results were obtained when proliferation was measured by BrdU incorporation into DNA [83] and when seeding was performed at different initial cell densities (not shown).

HPLC-ESI-MS/MS Analysis of Oxidized Lipid Molecular Species That Accumulate in INS-1 Cells Incubated with
Streptozotocin. To examine oxidized lipid molecular species in INS-1 cells, LC/ESI/tandem mass spectrometric scanning was used to detect parent ions that liberate an oxidized polyunsaturated fatty acid carboxylate anion (Figure 6(a)) upon collisionally activated dissociation (CAD) [85]. Hydroxyeicosatetraenoate (HETE) (m/z 319.3) arising from the oxidized analog of the glycerophosphoethanolamine (GPE) species 18 : 0/20 : 4-GPE (oxy-analog m/z 782.76) was found to represent the most abundant of the oxidized lipid species in INS-1 cells (Figure 6(b)), which is consistent with the facts that this is also the most abundant oxidized GPE lipid in activated platelets [85] and that 18 : 0/20 : 4-GPE is the most abundant GPE lipid in INS-1 cells [90] and rat islets [91].   (Figures 7(f) and 7(g)). Both basal levels of (C18:0/HETE)-GPE and those achieved after incubation with STZ for the iPLA 2 γ-knockdown INS-1 cells exceeded those for control INS-1 cells (Figure 7(h)), and this is consistent with the proposal that iPLA 2 β acts to excise oxidized PUFA residues from phospholipids so that the resultant lysophospholipid can be reacylated with an unoxidized fatty acid substituent to restore the structure and function of the parent phospholipid [63,64].
Oxidized cardiolipin species were not observed directly due to the relatively low abundance of the parent lipid among all cellular lipids and the tendency of linoleate residues, which are the principal fatty acid substituents of cardiolipin, upon oxidization to undergo chain scission reactions that yield a variety of truncated sn-2 substituents rather than a single-predominant species [92]. Similar behavior has been observed for other polyunsaturated fatty acids [93]. Mitochondria contain substantial amounts of GPE lipids, however, and they undergo the largest fractional modification of measured mitochondrial lipid classes upon induction of apoptosis [94], which suggests that GPE lipid oxidation represents a surrogate marker for mitochondrial phospholipid oxidation. [64,87] with INS-1 cell lines incubated with IL-1β and IFN-γ under conditions similar to those in Figure 1. The inflammatory cytokine mixture induced a robust increase in apoptosis of both control INS-1 cells and iPLA 2 γ-KD cells, and the percentage of apoptotic cells for the cytokine-treated condition was significantly higher for the latter (30.2 ± 0.5%) than for the former (25.7 ± 0.9%) (Figure 7(a)). Incubation of the INS-1 cell lines with the oxidant agent STZ also induced a concentration-dependent increase in the percentage of apoptotic cells that was significantly higher for iPLA 2 γ-KD cells (12.1 ± 0.9) than for control INS-1 cells (4.5 ± 0.9) at the highest STZ concentration (7.5 mM) tested (Figure 7(b)).
Understanding the mechanisms that the β-cell uses to protect its mitochondrial membranes from oxidative injury could yield insight into the pathogenesis of β-cell loss and development of means to treat or prevent T1DM and T2DM. Phospholipases A 2 (PLA 2 ) hydrolyze glycerophospholipids to yield a free fatty acid and a 2-lysophospholipid [77,78], and PLA 2 are thought to participate in signaling and membrane-remodeling processes that include repairing of oxidative damage to membranes in order to preserve their functional integrity [70][71][72][73][74]. When lipid peroxidation occurs, the oxidized sn-2 fatty acid substituent of phospholipids becomes less hydrophobic and more accessible to phospholipases [73]. The lysophospholipid that results from PLA 2 -catalyzed removal of oxidized fatty acid substituents can be reacylated with an unoxidized fatty acid to restore the native structure and function of the parent phospholipid.
Two members of a lipase family [95] that has been designated Group VI PLA 2 [96] or patatin-like phospholipase domain-containing (PNPLA) proteins [97] may play such a role in remodeling mitochondrial cardiolipin, and neither enzyme requires Ca 2+ for catalytic activity. Group VIA PLA 2 (iPLA 2 β) was the first member of this family to be recognized [98,99] and is also designated PNPLA9. Group VIB PLA 2 (iPLA 2 γ) was recognized thereafter [88,100] and is also designated PNPLA8 [96,97]. iPLA 2 γ, is expressed in mitochondria and peroxisomes [89,101,102], which are both membranous organelles that produce reactive oxygen species, and iPLA 2 γ cooperates with iPLA 2 β in stimulated phospholipid hydrolysis in some circumstances [103].
Mitochondria also contain iPLA 2  interest in the possibility that iPLA 2 β participates in cardiolipin remodeling [104]. Barth syndrome results from mutations in the tafazzin gene, which encodes a mitochondrial phospholipid-lysophospholipid transacylase, and the disorder is characterized by severe cardioskeletal myopathy, low-cardiolipin content, and abnormal cardiolipin fatty acyl composition [105]. Tafazzin-deficient Drosophila have similar abnormalities in cardiolipin content and mitochondrial function associated with monolysocardiolipin accumulation, and this phenotype is suppressed by inactivation of the iPLA 2 β gene, suggesting that iPLA 2 β contributes to monolysocardiolipin formation [104].
Several observations indicate that iPLA 2 γ is also involved in cardiolipin remodeling. Selective overexpression of iPLA 2 γ in mouse myocardium results in altered mitochondrial function associated with cardiac dysfunction [106,107], and genetic ablation of iPLA 2 γ produces a deficient mitochondrial bioenergetic phenotype [108] associated with cognitive dysfunction and hippocampal abnormalities that include mitochondrial degeneration and alterations in cardiolipin content and molecular species distribution [109]. iPLA 2 γ-null mice also exhibit exaggerated high-fat dietinduced changes in tissue cardiolipin content and composition and altered patterns of mitochondrial fatty acid oxidation [110,111]. Cardiolipin remodeling in myocardial mitochondria that occurs during heart failure in rats also appears to involve iPLA 2 γ [112]. iPLA 2 β and iPLA 2 γ cooperate in effecting certain cell fate decisions [113], and both enzymes may participate in determining whether a cell survives or succumbs to oxidative injury via their roles in cardiolipin metabolism. In β-cells, stimuli that induce apoptosis cause iPLA 2 β to redistribute from cytosol to mitochondria [63-65, 92, 114-119]. Staurosporine, for example, stimulates INS-1 cell mitochondrial superoxide production, and this results in mitochondrial membrane peroxidation, cytochrome c release, and apoptosis [63][64][65]. Staurosporine-induced membrane peroxidation and apoptosis in β-cells are attenuated by overexpressing iPLA 2 β and amplified by its pharmacologic inhibition or genetic ablation [63][64][65]. This may reflect a role for iPLA 2 β to excise oxidized cardiolipin fatty acid residues to generate monolysocardiolipin species that can be reacylated to restore the native structure [63][64][65].
Our findings indicate that iPLA 2 γ could participate in maintaining the aggregate mass of β-cells by promoting their proliferation and by protecting them from oxidative membrane injury induced by inflammatory cytokines or by oxidant agents that leads to apoptosis. Our iPLA 2 γknockdown (KD) INS-1 cell lines exhibited significantly lower growth rates than control INS-1 cells did. The inflammatory cytokines IL-1β and IFN-γ increased INS-1 cell iPLA 2 γ expression, and a similar response occurred when INS-1 cells were incubated with the oxidant agents STZ and TBHP. Those findings suggest that iPLA 2 γ may be upregulated as a compensatory repair mechanism in response to agents that injure β-cells, and this is consistent with the observations that iPLA 2 γ-KD INS-1 cells were also more sensitive than control cells to injury from inflammatory cytokines and oxidative agents. These findings in β-cell lines are consistent with the increased sensitivity to oxidant-induced lipid peroxidation and apoptosis of renal proximal tubular cells with reduced iPLA 2 γ expression [70] and suggest that iPLA 2 γ plays a role in repairing oxidized membranes and mitigating oxidant-induced cellular injury.
The mechanisms by which cytokines and oxidant agents increase the expression of iPLA 2 γ have not yet been determined experimentally, but the accumulation of iPLA 2 γ mRNA suggests that increased transcription is involved. Current experiments to examine potential mechanisms are focused on three possibilities. One is that the redox sensitive transcription factor NF κ B is activated via ROS-mediated inactivation of its inhibitory subunit I κ B [121][122][123][124] and that NF κ B stimulates transcription of the iPLA 2 γ gene directly or indirectly. NF κ B activation is known to contribute to β-cell injury induced by cytokines [125] under conditions similar to those employed in the studies described here. Two is that transcriptional activation of the iPLA 2 γ gene might occur via p38 MAPK-dependent pathways, since stimuli that induce βcell ER stress and apoptosis result in p38 MAPK activation [87], and ROS-induced p38 MAPK activation contributes to apoptosis in other cells [126,127].
Three is that ROS-induced oxidation of cellular phospholipids yields agonistic ligands for the transcription factor PPARγ, as previously reported [128], which then activates transcription of the iPLA 2 γ gene. It is of interest in this regard that conditions that result in differentiation of 3T3L1 fibroblasts to adipocytes lead to increased expression of PPARγ and in transcriptional upregulation of iPLA 2 γ and iPLA 2 β and that siRNA directed against either enzyme blocks differentiation [113].
The presence of oxidized phospholipids in INS-1 cells treated with oxidant agents in the studies described here was determined by performing LC/MS/MS scans of lipid extracts for precursors of m/z 295, m/z 319, and m/z 343 in order to identify glycerolipid molecular species that contained singly oxygenated forms of the polyunsaturated fatty acids (PUFA) linoleate (C18:2), arachidonate (C20:4), or docosahexaenoate (C22:6), respectively. The major oxylipid species identified was (stearoyl, hydroxyeicosatetraenoyl)-glycerophosphoethanol-amine [(C18:0/HETE)-GPE], and it was quantified by MRM of the transition 782.6 → 319.3, which corresponds to production of the HETE carboxylate anion from the [M-H] − ion of the parent oxy-phospholipid species. Minor species were observed at other m/z values but were not further characterized because of the low signal obtained from the limited amount of lipid contained in the quantities of INS-1 cells with which it was practical to work.
Although C18:0/HETE-PE is the most abundant oxidized phospholipid observed here, it is probably not the only oxidized species formed under these conditions. Oxidized lipids represent only a tiny fraction (substantially below 1%) of their unoxidized precursors, and all but the most abundant species will likely fall below the limit of quantitation, even if present in the mixtures, when the amount of membrane lipid available for analysis is limiting. In addition, other phospholipid oxidation products, for example, those that contain     esterified hydroperoxy-or ketofatty acid derivatives [129], would not have been detected by the approach used here, which would also have failed to detect esterified short chain substituents arising from PUFA oxidation [130]. Moreover, neither the regio-nor the stereospecificity of oxygenation was determined in our studies because of the limited amount of oxidized lipid available for characterization, and it is possible that C18:0/HETE-PE consisted of several distinct isomers, as reported for other cells [85].
Nonetheless, we think that C18:0/HETE-PE represents a reasonable marker for phospholipid oxidation in our experiments for several reasons. First, oxidized species of PE are much more abundant in stimulated monocytes and platelets than are oxidized species of other phospholipid head group classes, including PC, PI, PS, or PG [131], and C18:0/HETE-PE is the most abundant oxidized diacyl-phospholipid under those conditions [85,131,132]. Second, the precursor C18:0/ C20:4-PE is the most abundant PE species in INS-1 cells [90,91]. Moreover, mitochondria contain substantial amounts of PE lipids, and they undergo the largest fractional modification of measured mitochondrial lipid classes upon induction of apoptosis [94], which suggests that PE lipid oxidation serves as a surrogate marker for mitochondrial phospholipid oxidation.
The LC/MS/MS measurements reported here indicate that INS-1 cell C18:0/HETE-PE content rises upon incubation with an oxidant agent and is higher in cells in which iPLA 2 γ expression level has been knocked down compared to control cells. This is compatible with a role for iPLA 2 γ in remodeling of oxidized phospholipids that involves excision of oxidized PUFA residues to yield lysophospholipid species that can be reacylated with unoxidized fatty acyl-CoA molecules. This would regenerate the native phospholipid structure and restore its normal function, thereby mitigating the effects of oxidative insults that might otherwise induce apoptosis.

Conclusions
Group VIB Phospholipase A 2 (iPLA 2 γ) is distributed in mitochondria and expressed by insulin-secreting pancreatic islet β-cells and INS-1 insulinoma cells that are susceptible to oxidative injury by inflammatory cytokines, for example, IL-1β and IFN-γ, and by oxidizing toxins, for example, streptozotocin (STZ) or t-butyl-hydroperoxide (TBHP), via processes relevant to β-cell loss in types 1 and 2 diabetes mellitus. We demonstrate here that INS-1 cells incubated with IL-1β and IFN-γ, with STZ, or with TBHP increase their expression of iPLA 2 γ mRNA and protein and that INS-1 knockdown (KD) cell lines with reduced iPLA 2 γ expression proliferate more slowly than control INS-1 cells and undergo increased membrane peroxidation when incubated with cytokines or oxidants. Accumulation of the oxidized phospholipid species (1-stearoyl, 2-hydroxyeicosatetraenoyl)-snglycerophosphocholine was demonstrated in STZ-treated INS-1 cells by LC/MS/MS scanning, and the levels in iPLA 2 γ-KD cells exceeded those in control cells. iPLA 2 γ-KD INS-1 cells also exhibited higher levels of apoptosis than control cells when incubated with STZ or with IL-1β and IFN-γ. Together, these observations suggest that iPLA 2 γ promotes β-cell proliferation and that its increased expression during inflammation or oxidative stress may serve to mitigate membrane injury and thereby to enhance β-cell survival under these conditions.