Effects of Stable Suppression of Group VIA Phospholipase A2 Expression on Phospholipid Content and Composition, Insulin Secretion, and Proliferation of INS-1 Insulinoma Cells*

Studies involving pharmacologic inhibition or transient reduction of Group VIA phospholipase A2 (iPLA2β) expression have suggested that it is a housekeeping enzyme that regulates cell 2-lysophosphatidylcholine (LPC) levels, rates of arachidonate incorporation into phospholipids, and degradation of excess phosphatidylcholine (PC). In insulin-secreting islet β-cells and some other cells, in contrast, iPLA2β signaling functions have been proposed. Using retroviral vectors, we prepared clonal INS-1 β-cell lines in which iPLA2β expression is stably suppressed by small interfering RNA. Two such iPLA2β knockdown (iPLA2β-KD) cell lines express less than 20% of the iPLA2β of control INS-1 cell lines. The iPLA2β-KD INS-1 cells exhibit impaired insulin secretory responses and reduced proliferation rates. Electrospray ionization mass spectrometric analyses of PC and LPC species that accumulate in INS-1 cells cultured with arachidonic acid suggest that 18:0/20:4-glycerophosphocholine (GPC) synthesis involves sn-2 remodeling to yield 16:0/20:4-GPC and then sn-1 remodeling via a 1-lyso/20:4-GPC intermediate. Electrospray ionization mass spectrometric analyses also indicate that the PC and LPC content and composition of iPLA2β-KD and control INS-1 cells are nearly identical, as are the rates of arachidonate incorporation into PC and the composition and remodeling of other phospholipid classes. These findings indicate that iPLA2β plays signaling or effector roles in β-cell secretion and proliferation but that stable suppression of its expression does not affect β-cell GPC lipid content or composition even under conditions in which LPC is being actively consumed by conversion to PC. This calls into question the generality of proposed housekeeping functions for iPLA2β in PC homeostasis and remodeling.

Secretory PLA 2 (sPLA 2 ) are low molecular weight enzymes that require millimolar [Ca 2ϩ ] for catalysis and affect inflammation and other processes, and the PAF-acetylhydrolase PLA 2 family exhibits substrate specificity for PAF and oxidized phospholipids (1). Of Group IV cytosolic PLA 2 (cPLA 2 ) family members (1), cPLA 2 ␣ was the first identified and prefers substrates with sn-2 arachidonoyl residues, catalyzes arachidonate release for subsequent metabolism, associates with its substrates in membranes upon rises in cytosolic [Ca 2ϩ ], and is also regulated by phosphorylation (6). There are additional members of the cPLA 2 family that arise from separate genes (7)(8)(9)(10).
The Group VI PLA 2 (iPLA 2 ) enzymes (11)(12)(13) do not require Ca 2ϩ for catalysis and are inhibited by a bromoenol lactone (BEL) suicide substrate (14) that does not inhibit sPLA 2 or cPLA 2 at similar concentrations (14 -17). The Group VIA PLA 2 (iPLA 2 ␤) resides in the cytoplasm of resting cells, but Group VIB PLA 2 contains a peroxisomal targeting sequence and is membrane-associated (18,19). These enzymes belong to a larger class of serine lipases that are encoded by multiple genes (20,21). The iPLA 2 ␤ enzymes cloned from various species are 84 -88-kDa proteins that contain a GXSXG lipase consensus sequence and eight stretches of a repetitive motif homologous to that in the protein-binding domain of ankyrin (11)(12)(13).
It has been proposed that iPLA 2 ␤ plays housekeeping roles in phospholipid metabolism (22,23), such as generating lysophospholipid acceptors for incorporating arachidonic acid into phosphatidylcholine (PC) of murine P388D1 macrophage-like cells, based on studies involving reducing iPLA 2 activity with BEL or an antisense oligonucleotide, which suppresses [ 3 H]arachidonate incorporation into PC and reduces [ 3 H]lysophosphatidylcholine (LPC) levels (24,25). Arachidonate incorporation involves a deacylation/reacylation cycle of phospholipid remodeling (26,27), and the level of LPC is thought to limit the [ 3 H]arachidonic acid incorporation rate into P388D1 cell PC (24,25).
Another housekeeping function for iPLA 2 ␤ in PC homeostasis has been proposed from studies of overexpression of CTP:phosphocholine cytidylyltransferase (28,29), which catalyzes the rate-limiting step in PC synthesis. Cells that overexpress CTP:phosphocholine cytidylyltransferase exhibit increased rates of PC biosynthesis and degradation and little net change in PC levels, suggesting that PC degradation is up-regulated to prevent excess PC accumulation. Increased PC degradation in CTP:phosphocholine cytidylyltransferase-overexpressing cells is prevented by BEL, and iPLA 2 ␤ protein and activity increase, suggesting that iPLA 2 ␤ is up-regulated (28,29).
If general, this could be important, because PC synthesis is involved in regulating the cell cycle and apoptosis (30,31), but studies involving iPLA 2 ␤ overexpression give a different perspective on this issue (32,33).
The activity of iPLA 2 ␤ has been reported to vary with the cell cycle, to be required for proliferative responses in lymphocyte cell lines (34,35), and to play a role in membrane biogenesis (36).
The ambiguity of pharmacologic studies with BEL makes manipulating iPLA 2 ␤ expression by molecular biologic means an attractive alternative to study iPLA 2 ␤ functions. Transient suppression of iPLA 2 ␤ activity with antisense oligonucleotides has been useful in monocytemacrophages and vascular myocytes (25,37,41,42,47), but substantial suppression of expression is not readily achieved in all cells in this way. Insulinoma cell lines experience toxicity from antisense oligonucleotides at concentrations that fail to reduce iPLA 2 ␤ activity (66).
Establishing iPLA 2 ␤ Knockdown INS-1 Insulinoma Cell Lines Using siRNA (70)-Two hairpin-forming oligonucleotides directed against iPLA 2 ␤ mRNA were cloned into RNAi-Ready pSIREN Retro-Q as per the manufacturer's protocol (BD Biosciences Clontech). Targeting sequences within the synthetic oligonuceotides are italicized and underlined below. The sequence of the first was gatccAACAGCACAGAGA-ATGAGGAGTTCAAGAGACTCCTCATTCTCTGTGCTGTTTT-TTTTg. The second oligonucleotide was gatccGCCTGAACCAGGT-GAACAATTCAAGAGATTGTTACCTGGTTCAGGCTTTTTTg. Constructs that express the siRNAs are pSIREN-iPLA 2 -1 and pSIREN-iPLA 2 -2. Retroviruses were packaged in PT67 cells and used to infect INS-1 cells. Cells were selected with 0.4 g/ml puromycin. A construct that encoded scrambled RNA was used to prepare control INS-1 cell lines.
Analyses of INS-1 Cell iPLA 2 ␤ mRNA-Northern blots of iPLA 2 ␤ mRNA were performed as described (71). For quantitative real time reverse transcription-PCR, total RNA was isolated with an RNeasy kit (Qiagen Inc.). The SuperScript First Strand Synthesis System (Invitrogen) was used to synthesize cDNA in 20-l reactions that contained DNase I-treated total RNA (2 g). The cDNA product was treated (20 min, 37°C) with RNase H (2 units; Invitrogen) and heat-inactivated (70°C for 15 min). A reaction without reverse transcriptase was performed to verify the absence of genomic DNA. PCR amplifications were performed using SYBR Green dye in an ABI 7000 detection system (Applied Biosystems). Product sizes were determined on 3% (w/v) agarose-TAE gels.
Determination of Insulin Secretion by INS-1 Cells-At confluence, INS-1 cells were detached from T75 flasks, and aliquots (10 5 cells) were added to 24-well plates. Culture medium was removed the next day, and cells were washed twice in KRB containing 1 mM glucose and 0.1% bovine serum albumin. Cells were preincubated (1 h, 37°C, under 95% air, 5% CO 2 ), and medium was replaced with KRB containing glucose (3 or 20 mM) with or without forskolin (2.5 M). Cells were incubated (1 h, 37°C, under 95% air, 5% CO 2 ), and medium was then removed to measure insulin by radioimmunoassay. Secretion was normalized to cell protein measured with Coomassie reagent (Pierce) (72).
Determination of INS-1 Cell Proliferation Rate-One assay used to measure INS-1 cell proliferation rates is based on fluorescence enhancement when CyQuant GR binds to nucleic acids, which reflects the amount of cell DNA (73). Cells were seeded onto 96-well plates (3 ϫ 10 3 cells/well). Medium was removed after 1 or 3 days, and cells were frozen (Ϫ20°C). DNA was measured with a CyQuant assay kit (Molecular Probes, Inc., Eugene, OR) with reference to a standard curve. CyQuant GR solution (200 l) 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-2Ј-deoxyurindine (BrdUrd) into DNA in proliferating cells (74). Cells were seeded (10 4 cells/well) and cultured (3 days) before assay with BrdUrd labeling and an enzyme-linked immunoassay detection kit III (Roche Applied Science).
Extraction INS-1 Cell Phospholipids and Quantitation of Phosphorus-Lipids were extracted (75), and their lipid phosphorus content was measured (76), as described.
Electrospray Ionization Mass Spectrometric Analyses of Glycerophosphocholine Lipids-PC and LPC were analyzed as Li ϩ adducts by positive ion ESI/MS on a Finnigan (San Jose, CA) TSQ-7000 triple stage quadrupole mass spectrometer with an ESI source controlled by Finnigan ICIS software. Phospholipids were dissolved in methanol/chloroform (2/1, v/v) containing LiOH (10 pmol/l), infused (1 l/min) with a Harvard syringe pump, and analyzed as described (77)(78)(79). For tandem MS, precursor ions selected in the first quadrupole were accelerated (32-36-eV collision energy) into a chamber containing argon (2.3-2.5 millitorrs) to induce collisionally activated dissociation (CAD), and product ions were analyzed in the final quadrupole. The identities of GPC species were determined from their tandem spectra (77)(78)(79), and their quantities were determined relative to the internal standards 14:0/ 14:0-GPC and 18:0/22:6-GPC by interpolation from a standard curve (66,80). To quantitate LPC species, constant neutral loss of 59 scanning was performed. The intensity of the ion for the 14 (80), and their tandem spectra were obtained as described (69,72).
Statistical Analyses-Two groups were compared by Student's t test, and multiple groups were compared by one-way analysis of variance with post hoc Newman-Keul's analyses.

RESULTS
Establishing iPLA 2 ␤ Knockdown Cell Lines-INS-1 insulinoma cells were infected with retroviral constructs containing inserts that produced either scrambled RNA (control) or siRNA directed against sequences in iPLA 2 ␤ mRNA. Selection of puromycin-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 ␤ mRNA content when analyzed by Northern blots (Fig. 1A) or by real time PCR (Fig. 1B). The iPLA 2 ␤ activity in control INS-1 cells was stim- ulated by 1 mM ATP and inhibited by 10 M BEL (Fig. 1C), as is iPLA 2 ␤ activity in islets and other insulinoma cells (13). The iPLA 2 ␤ knockdown (iPLA 2 ␤-KD) cell lines exhibited reduced iPLA 2 ␤ activity, and the reduction of activity was comparable with that of iPLA 2 ␤ mRNA. The iPLA 2 ␤ expression level was a stable property of control and iPLA 2 ␤-KD INS-1 cells that persisted on serial passage in culture.
Insulin Secretion Is Attenuated in iPLA 2 ␤ Knockdown Cells-Glucose is the dominant insulin secretagogue, and pancreatic islets and insulinoma cells exhibit greater insulin secretory responses in the presence of forskolin (32,68). When treated with forskolin (2.5 M), control INS-1 cells secreted more insulin than either iPLA 2 ␤-KD INS-1 cell line (Fig.  2). This effect increased with the medium glucose concentration over the range of 3-20 mM. Reduced insulin secretory responses were observed with both iPLA 2 ␤-KD cell lines under all conditions, supporting the proposal that iPLA 2 ␤ participates in insulin secretion (2,32,68).
Cell Proliferation Rate in iPLA 2 Knockdown Cells-Pharmacologic inhibition of iPLA 2 ␤ reduces cell proliferation rates, and iPLA 2 ␤ overexpression results in increased proliferation (32,34,35). It might thus be predicted that iPLA 2 ␤-KD cells would proliferate less rapidly than control INS-1 cells. To test this possibility, INS-1 cell proliferation was measured using an indicator that exhibits strong fluorescence enhancement upon association with nucleic acids (73). Identical numbers of INS-1 cells were seeded at time 0, and their growth rates were monitored for 1-3 days. INS-1 iPLA 2 ␤-KD lines proliferated at rates that were significantly lower than those for control INS-1 cells (Fig. 3A). Proliferation was also measured by BrdUrd incorporation into DNA (74). INS-1 cells were again seeded in identical numbers at time 0, and the increase in BrdUrd signal was monitored for 3 days. INS-1 iPLA 2 ␤-KD lines were again found to proliferate more slowly than control INS-1 cells, and similar results were obtained when seeding was performed at either of two different initial cell densities at time 0 (Fig. 3B).
INS-1 cell LPC levels were thus measured with an internal standard by ESI/MS/MS scanning for loss of trimethylamine, which greatly increases the signal/noise ratio (66,78,79) and permits LPC measure- ment under conditions where total ion current ESI/MS tracings are overwhelmed by chemical noise (66). PC species also yield signal in such scans, and the 14:0/14:0-GPC internal standard can be used to measure LPC (Fig. 9).
The most abundant ions in such scans of LPC Li ϩ adducts from resting control (Fig. 9A) (72,82). Fig. 9C is the tandem spectrum of m/z 528 in Fig. 9, A or B, and shows that the ion represents 18:1/2-lyso-GPC-Li ϩ rather than the regioisomer 1-lyso/18:1-GPC or the isobar 1-O-hexadecenyl/2-acetyl-GPC (16: 1-PAF). That the ion represents 2-LPC rather than 1-LPC is reflected by the relative abundance of m/z 469 (loss of trimethylamine) and m/z 339 (net loss of Li ϩ phosphocholine via initial loss of trimethylamine) (78,79). The latter loss is favored by an sn-2 ␣-hydrogen atom, which is present in 1-LPC but not 2-LPC (78,79). This also accounts for the greater abundance in 2-LPC spectra of the head group-derived ion m/z 104 compared with ions from losses of phosphocholine (m/z 345 and 339), whereas the converse is true for 1-LPC (78,79). That 16:1-PAF does not account for m/z 528 in Fig. 9, A-C, is reflected by the absence of an ion for loss of Li ϩ -acetate and of a charge-remote ion series for 1-O-alkyl chain fragmentation, which are strong ions in PAF-Li ϩ adduct spectra (79).
Little difference in the LPC content of resting control and iPLA 2 ␤-KD INS-1 cells was observed (Fig. 9, A and B), indicating that iPLA 2 ␤ expression level is not the major regulator of their LPC content. Interestingly, incubating either control or iPLA 2 ␤-KD INS-1 cells with arachidonic acid caused accumulation of material represented by an ion of m/z 550 (Fig. 9D).
That m/z value is consistent with 20:4-LPC-Li ϩ , but, in attempts to obtain a CAD spectrum, the ion of m/z 550 was resistant to fragmentation at a collision energy of 32 eV, which produced ready fragmentation of other LPC species (Fig. 9C), or at higher values. There was modest attenuation of the parent ion and limited production of fragment ions (Fig. 10A). When this material was stored for several days or heated for 2 h, the CAD spectrum in Fig. 10B was obtained, which is similar to that reported for 1-20:4/2-lyso-GPC (79) and to that obtained from standard 1-20:4/2-lyso-GPC prepared from 20:4/20:4-GPC with an sPLA 2 from N. naja venom (Fig. 10C). Fig. 10, B and C, identifies the parent ions as 1-20:4/2-lyso-GPC-Li ϩ , because they contain the Li ϩ -arachidonic acid ion (m/z 311) that identifies the fatty acid substituent; ions reflecting losses of trimethylamine (m/z 491) and of phosphocholine with H ϩ or Li ϩ (m/z 367 and 361, respectively) that identify the head group; and ions characteristic of  (Fig. 11A).
Continued culture of INS-1 cells with arachidonic acid for 24 h causes some further rise in 20:4-LPC levels, but other LPC species remain nearly constant (Fig. 11B), despite the fact that continued accumulation of 20:4-GPC species, and presumably consumption of LPC, occurs between 6 and 24 h (Figs. 5-8). The increasing abundance of 20:4-LPC (Fig. 11) must reflect its more rapid generation than consumption in synthetic or degradative reactions. No accumulation of 20:4/20:4-GPC (m/z 836 for MLi ϩ ) was observed.
ESI/MS Analyses of GPE, GPG, and GPI Lipids in INS-1 Cells Incubated with Arachidonic Acid-Arachidonic acid first incorporated into PC is then transferred to other phospholipids (66,84). When cultured with arachidonic acid, both control and iPLA 2 ␤-KD INS-1 cells replaced GPE, GPG, and GPI species that contained sn-2 18:1 with species that contained sn-2 20:4 without obvious differences between the cell lines (Fig. 12).

DISCUSSION
We have generated clonal INS-1 iPLA 2 ␤-KD cell lines that stably express siRNA that reduces iPLA 2 ␤ expression to less than 20% of control INS-1 cell levels, and this property is stable on serial passage. This provides a tool to study ␤-cell iPLA 2 ␤ function, which is important because the best iPLA 2 ␤ pharmacologic inhibitor (BEL) also inhibits other enzymes (17,18,20,21,64,65) and because antisense oligonucleotides are toxic to ␤-cells at concentrations that fail to reduce iPLA 2 ␤ expression (66).
Although iPLA 2 ␤ has been proposed to play the housekeeping role of maintaining membrane phospholipid homeostasis by degrading excess PC (28,29), stable suppression of iPLA 2 ␤ expression in INS-1 cells does not result in a significant change in their PC content or composition. Another proposed iPLA 2 ␤ housekeeping role is to provide LPC acceptors for incorporating arachidonic acid into PC (22)(23)(24)(25), but we also observe no change in LPC content or composition or in arachidonate incorporation rates into iPLA 2 ␤-KD INS-1 cells.
Rather, we find that INS-1 iPLA 2 ␤-KD cells incorporate arachidonate into GPC lipids as readily as control cells when cultured with arachidonic acid. In both control and iPLA 2 ␤-KD INS-1 cells cultured in standard medium, GPC lipids contain primarily sn-2 oleate or palmitoleate, and such species are also abundant in anionic phospholipids. Upon culture with arachidonic acid, all of these lipids are remodeled, and species with sn-2 arachidonate become their most abundant components.
Our ESI/MS findings that the PC and LPC content and composition and the arachidonate incorporation rates into phospholipids of iPLA 2 ␤-KD and control INS-1 cells are virtually identical are consistent with effects of pharmacologic inhibition (57,66) and molecular biologic overexpression (32,68) of iPLA 2 ␤ on these parameters in INS-1 cells and islets.
Our findings also contribute to evidence that iPLA 2 ␤, the activity of which varies with the cell cycle (34,35), is involved in cell proliferation. Phytohemagglutinin-induced lymphocyte proliferation (35) and thrombin-induced vascular myocyte proliferation (42) are suppressed by BEL or by reducing iPLA 2 ␤ expression with antisense oligonucleotides (35). The reduced proliferation rates reported here for INS-1 cells in which iPLA 2 ␤ expression is stably suppressed complement the find-ing that stable iPLA 2 ␤ overexpression in INS-1 cells increases the proliferation rate (32).
Although we find no evidence that iPLA 2 ␤ plays the proposed housekeeping roles in PC homeostasis (28,29) and remodeling (22)(23)(24)(25) in ␤-cells, iPLA 2 ␤ is widely expressed and might have multiple functions that vary among tissues and cell types, perhaps dependent in part on which splice variants (49 -52) and proteolytic processing products (38,57) of iPLA 2 ␤ are expressed in a given cell and on what interacting proteins (102) are present in the cell compartment (68,103) in which the iPLA 2 ␤ isoform resides. Our findings do indicate that iPLA 2 ␤ plays signaling or effector role(s) in stimulated insulin secretion from ␤-cells, and these results with molecular biologic suppression of iPLA 2 ␤ expression provide an important independent test of inferences from studies with pharmacologic inhibition of iPLA 2 ␤ activity (54 -58, 72, 88).