Islet Complex Lipids

Abstract

The β-isoform of group VIA calcium-independent phospholipase A₂ (iPLA₂β) does not require calcium for activation, is stimulated by ATP, and is sensitive to inhibition by a bromoenol lactone suicide substrate. Several potential functions have been proposed for iPLA₂β. Our studies indicate that iPLA₂β is expressed in β-cells and participates in glucose-stimulated insulin secretion but is not involved in membrane phospholipid remodeling. If iPLA₂β plays a signaling role in glucose-stimulated insulin secretion, then conditions that impair iPLA₂β functions might contribute to the diminished capacity of β-cells to secrete insulin in response to glucose, which is a prominent characteristic of type 2 diabetes. Our recent studies suggest that iPLA₂β might also participate in β-cell proliferation and apoptosis and that various phospholipid-derived mediators are involved in these processes. Detailed characterization of the iPLA₂β protein level reveals that β-cells express multiple isoforms of the enzyme, and our studies involve the hypothesis that different isoforms have different functions.

CLASSIFICATION OF PHOSPHOLIPASE A₂ AND FEATURES OF A GROUP VIA PHOSPHOLIPASE A₂ (iPLA₂β)

Phospholipase A₂ (PLA₂) (1) is a diverse group of enzymes that catalyze hydrolysis of the sn-2 substituent from glycerophospholipid substrates to yield a free fatty acid and a 2-lysophospholipid (1). At present, the recognized PLA₂s are classified into 14 different groups, based on their calcium requirement for activation and sequence homology (2). These include the low–molecular weight secretory PLA₂s (groups IB, IIA, IID, IIE, IIF, III, V, X, and XII) and the higher–molecular weight Ca²⁺-dependent cytosolic PLA₂s (groups IVA, IVB, and IVC) and the Ca²⁺-independent PLA₂s (groups VIA and VIB).

Among the PLA₂s is an 84-kDa (752 amino acid residues) cytosolic PLA₂ that does not require Ca²⁺ for catalysis. This PLA₂ has now been cloned from several sources (3–5), including rat and human pancreatic islet β-cells (4,6), is classified as group VIA PLA₂, and is designated the β-isoform of group VIA calcium-independent phospholipase A₂ (iPLA₂β) (7–9). Salient features (10,11) of the iPLA₂β amino acid sequence (Fig. 1) include eight NH_2-terminal ankyrin repeats, a caspase-3 cleavage site, an ATP-binding domain, a serine lipase consensus sequence (GXSXG), a bipartite nuclear localization sequence, and a COOH-terminal calmodulin-binding domain(s). An 88-kDa iPLA₂β isoform is also expressed in human pancreatic islets and is the product of an mRNA species that arises from alternate splicing (6). This isoform contains a 54–amino acid sequence that interrupts the eighth ankyrin repeat.

FIG. 1.

Rat pancreatic islet iPLA₂β-deduced amino acid sequence. The deduced 752–amino acid sequence of rat pancreatic islet iPLA₂β, updated from our earlier illustration (4), is shown to contain an underlined region of amino acid sequences homologous to a repetitive motif in ankyrin, in addition to a caspase-3 cleavage site, an ATP-binding domain, a catalytic serine lipase consensus sequence, a bipartite nuclear localization consensus sequence, and calmodulin-binding domain(s), which are underlined and numbered 1–5, respectively.

PROPOSED FUNCTIONS FOR iPLA₂β

Phospholipid remodeling

A housekeeping role involving generation of lysophospholipid acceptors for incorporation of arachidonic acid into phospholipids has been proposed for iPLA₂β, based on experiments involving inhibition of iPLA₂β activity in P388D1 cells with the bromoenol lactone (BEL) suicide substrate or with an antisense oligonucleotide (10,11). Inhibition of iPLA₂β activity in P388D1 cells suppressed (~60%) incorporation of [³H]-arachidonic acid into phospholipids while reducing (~60%) [³H]-lysophosphatidylcholine levels in [³H]-choline–labeled P388D1 cells. However, [³H]-palmitic acid incorporation was reduced only slightly. This is thought to represent the mechanism whereby iPLA₂β inhibition reduces incorporation of [³H]-arachidonic acid into P388D1 cell phospholipids. Such incorporation reflects a deacylation/reacylation cycle (12) of phospholipid remodeling rather than de novo synthesis (13), and the level of lysophosphatidylcholine acceptors is thought to limit the rate of [³H]-arachidonic acid incorporation into P388D1 cell phosphatidylcholine (10,11).

A second housekeeping function for iPLA₂β is suggested from studies with CTP:phosphocholine cytidyltransferase (CT)-overexpressing cells (14). CT catalyzes the rate-limiting step in phosphatidylcholine biosynthesis via the Kennedy pathway, and cells overexpressing CT exhibit increased rates of phosphatidylcholine biosynthesis and degradation and little net change in phosphatidylcholine accumulation (14). The increased phosphatidylcholine degradation observed in CT-overexpressing cells is prevented by BEL, and immunoreactive iPLA₂β protein and activity increase in such cells, suggesting that iPLA₂β is upregulated in response to CT overexpression (14). If general, this could represent an important role for iPLA₂β in cell biology because phosphatidylcholine biosynthesis is involved in regulation of the cell cycle and apoptosis (15).

Cell proliferation

Inhibition of iPLA₂β with BEL is reported to reduce arachidonic acid release, [³H]-thymidine incorporation, and rates of lymphocyte (16) and Caco-2 (17) cell proliferation. In view of findings that arachidonic acid and/or its metabolites can stimulate c-fos, c-jun, or mitogen-activated protein kinase (18), it is possible that iPLA₂β might affect nuclear events involved in cell division.

Apoptosis

Involvement of iPLA₂β in processes leading to apoptotic cell death is suggested by several lines of evidence. Induction of apoptosis of U927 cells by anti-fas antibody is associated with hydrolysis of arachidonic acid from membrane phospholipids that is not catalyzed by group IV cytosolic phospholipase A₂, which is inactivated by caspases during apoptosis. The release of arachidonic acid is not inhibited by inhibitors of group II secretory phospholipase A₂, but the release is suppressed by the iPLA₂β inhibitor BEL (19). Both arachidonate 12-lipoxygenase and inducible nitric oxide synthase (iNOS) knockout mice are resistant to the diabetogenic effects of low doses of streptozotocin (20,21). Interleukin-1β–induced generation of nitric oxide in pancreatic islets causes accumulation of arachidonic acid and augments production of the arachidonic acid metabolite 12-hydroxy-(5,8,10,14)-eicosatetraenoic acid (12-HETE), and these effects are prevented by BEL (22). Induction of apoptosis in human promonocytic U937 cells is associated with activation of iPLA₂β after proteolysis (23). Collectively, these observations suggest that iPLA₂β is involved in releasing arachidonic acid from membrane phospholipids in apoptosis and that arachidonic acid and/or its metabolites serve mediator functions in apoptosis.

Signal transduction

Studies in various cellular systems suggest that iPLA₂β might also participate in various signaling pathways. Evidence in support of this role include the observations that BEL, at concentrations that inhibit iPLA₂β, suppresses 1) parathyroid-induced generation of arachidonic acid in rat proximal tubules (24), 2) stimulated superoxide generation in neutrophils (25,26), 3) interleukin-1β–stimulated increases in iNOS protein and nitric oxide generation in cardiac myocytes (27), 4) cAMP response element binding protein phosphorylation during the period of myocardial ischemia (28), and 5) virus-induced or double-stranded RNA-induced activation of iNOS expression by macrophages (29). Because pancreatic β-cells also express iPLA₂β, we have examined the potential role(s) of this group VIA Ca²⁺-independent iPLA₂β in β-cells, and our findings are discussed below.

POTENTIAL ROLES OF iPLA₂β IN β-CELLS

Evidence for a role in signal transduction

Glucose and other fuel secretagogues induce hydrolysis of phospholipids in β-cell membranes, and this is reflected by accumulation of phospholipid-derived mediators including inositol 1,4,5-triphosphate, free arachidonic acid, and arachidonate metabolites (30). Metabolism of fuel secretagogues to yield ATP is an obligatory event in their induction of hydrolysis of arachidonate from β-cell membrane phospholipids (30,31) just as it is for insulin secretion (32). The fact that fuel secretagogue-induced hydrolysis of membrane phospholipids is, in part, independent of Ca²⁺ influx (33) suggests that a Ca²⁺-independent phospholipase such as a iPLA₂β might be involved in this process.

Pancreatic islets, islet β-cells, and glucose-responsive insulinoma cells all express iPLA₂β mRNA and iPLA₂β enzymatic activity that is stimulated by ATP and inhibited by BEL (4,34–37). Inhibition of β-cell iPLA₂β activity with BEL suppresses glucose-stimulated hydrolysis of arachidonic acid from membrane phospholipids, the rise in cytosolic [Ca²⁺]_i, and insulin secretion (34–37). However, BEL does not inhibit incorporation of labeled fatty acids into either pancreatic islet or insulinoma cell membrane phospholipids (9,38). Whereas BEL, at concentrations that inhibit iPLA₂β, does not inhibit group IV cytosolic PLA₂ (37) or glucose oxidation (35), which requires both glycolytic metabolism and mitochondrial oxidation of glucose, in pancreatic islets, it has been recognized to decrease the ATP/ADP ratio in mouse islets (39) and also inhibit phosphatidate phosphohydrolase (40). In view of such reports, nonpharmacological approaches involving molecular biological manipulations were used to explore involvement of the iPLA₂β protein in β-cell function. Attempts to suppress β-cell iPLA₂β activity with antisense oligonucleotides were ineffective (38); therefore, effects of an alternate approach of overexpressing iPLA₂β in insulinoma cells were examined.

INS-1 cells were stably transfected with a retroviral vector containing an iPLA₂β cDNA insert (9). Two stably transfected lines were isolated that overexpressed iPLA₂β activity and protein by 10-fold compared with the parental cell line. These were designated overexpressing (OE) cells. Insulin secretion and phospholipid remodeling in OE cells were then examined. Relative to the secretory responses observed with INS-1 cells transfected with empty retroviral vector that did not contain iPLA₂β cDNA, INS-1 cells that overexpressed iPLA₂β exhibited a much greater insulin secretory response to glucose alone and to glucose in combination with the cAMP-elevating agents forskolin or isobutylmethylxanthine (9,41). Pretreatment of INS-1 cells with BEL suppressed stimulated insulin secretion, but neither overexpression of iPLA₂β nor inhibition of iPLA₂β enzymatic activity in the iPLA₂β-overexpressing INS-1 cells with BEL affected the rate or extent of arachidonic acid incorporation into INS-1 cell phosphatidylcholine.

Overexpression of iPLA₂β as a fusion protein with enhanced green fluorescence protein in INS-1 cells permitted examination of changes in its intracellular location after stimulation (Fig. 2), and glucose plus cAMP-elevating agents were found to induce accumulation of iPLA₂β in the perinuclear region. Such intracellular translocation appears to require cAMP-dependent protein kinase A–mediated phosphorylation events because inhibition of protein kinase A with H89 prevented stimulated perinuclear accumulation of iPLA₂β (41). Nuclear association of iPLA₂β induced by cAMP-elevating agents in INS-1 cells is of interest because glucose promotes both β-cell insulin secretion and proliferation, and glucose-induced INS-1 cell mitogenesis is cAMP-dependent (42). Because membranes of the nucleus and endoplasmic reticulum (ER) are contiguous (43), perinuclear accumulation of iPLA₂β is consistent with its association with a subcellular compartment that is likely to include the ER. Using organelle-specific trackers, association of iPLA₂β with β-cell ER and Golgi compartments upon stimulation has been observed (44). The likelihood that iPLA₂β might associate with the nucleus is also supported by the presence of a bipartite nuclear localization consensus sequence (⁵¹¹KREFGEHTKMTDVKKPK⁵²⁷) in the deduced amino acid sequence of iPLA₂β (45). Taken together, the above findings suggest that iPLA₂β has a signaling role in the β-cell, although iPLA₂β does not appear to participate in β-cell phospholipid remodeling or in phosphatidylcholine homeostasis.

FIG. 2.

Stimulated translocation of iPLA₂β to the perinuclear region of INS-1 cells. INS-1 cells overexpressing iPLA₂β as a fusion protein with enhanced green fluorescence protein (EGFP) were stimulated with glucose in the absence and presence of forskolin, as described (41). After a stimulation period of 30 min, the cells were examined by fluorescence microscopy to monitor green fluorescence associated with EGFP. A: Control cells. B: Cells treated with glucose (2 mmol/l) alone. C: Cells treated with both glucose (2 mmol/l) and forskolin (2.5 μmol/l). After stimulation, punctate perinuclear accumulation of fluorescence that reflects the location of iPLA₂β-EGFP is evident.

iPLA₂β and β-cell proliferation

The iPLA₂β-overexpressing INS-1 cell line also proliferates more rapidly than nontransfected parental or empty-vector transfected INS-1 cells. This is reflected by a greater rate of increase in cell number in cultures of iPLA₂β-overexpressing INS-1 cells (46). The bases for these phenomena have not been determined, but the presence of a bipartite nuclear localization consensus sequence in iPLA₂β raises the possibility that iPLA₂β might affect nuclear events involved in cell division (18). Further, iPLA₂β activation converts phosphatidic acid to lysophosphatidic acid (LPA) (5), and LPA is a potent mitogen (16,47). Enhanced proliferation might result from a rise in cellular LPA levels that occurs as a consequence of iPLA₂β overexpression, and OE cells have been demonstrated to contain higher levels of lysophospholipid species than control cells.

iPLA₂β and ER stress–mediated β-cell apoptosis

Apoptosis is involved in β-cell death in type 1 diabetes (48) and might also contribute to β-cell death in type 2 diabetes (49). At present, three apoptotic signaling pathways are recognized (50). These are the extrinsic death receptor pathway involving adaptor molecules, the intrinsic mitochondrial pathway, and the ER stress pathway.

Agents that deplete ER Ca²⁺ stores, such as thapsigargin (51), induce apoptosis of MIN-6 insulinoma cells by a pathway that does not require increases in [Ca²]_i but that does require generation of the arachidonic acid metabolite 12-HETE (52). Thapsigargin has been demonstrated to induce hydrolysis of arachidonic acid from islet membrane phospholipids, and this is suppressed by the iPLA₂β inhibitor BEL (53). These findings suggest that ER stress induced by Ca²⁺ store depletion activates iPLA₂β, which then hydrolyzes membrane phospholipids to yield products that promote β-cell death.

In support of this hypothesis are the findings that thapsigargin induces a threefold increase in apoptosis of parental INS-1 cells (control, 4.2 ± 0.2% vs. plus thapsigargin, 12.3 ± 0.7%, P < 0.05), and apoptosis is suppressed by BEL. Whereas the spontaneous incidence of apoptosis in control OE cells (3.7 ± 0.5%) is similar to that in parental cells, the apoptotic effect of thapsigargin is greatly amplified in the OE cells (Fig. 3A), and apoptosis is significantly suppressed by BEL. Thapsigargin also induces higher iPLA₂β activity and the generation of a 62-kDa iPLA₂β-immunoreactive protein (Fig. 3B). Inhibition of caspase-3 prevents both thapsigargin-induced apoptosis and perinu-clear accumulation of iPLA₂β (Fig. 3C). Immunoblotting and immunofluorescence analyses reveal that a 62-kDa iPLA₂β-immunoreactive protein accumulates in the perinuclear region of thapsigargin-treated INS-1 cells. These findings suggest that caspase-3 that is activated during apoptosis cleaves the 84-kDa iPLA₂β at its consensus sequence site to generate a 62-kDa product in INS-1 cells.

FIG. 3.

ER stress–induced apoptosis of INS-1 cells. A: Suppression of thapsigargin-induced apoptosis of iPLA₂β-overexpressing INS-1 cells by BEL. INS-1 cells stably transfected with a vector containing iPLA₂β cDNA (OE) construct were treated with thapsigargin (T) (1 μmol/l) for 24 h in the absence or presence of BEL (10 μmol/l). The cells were then harvested for TUNEL (Tdt-mediated dUTP nick end labeling) analyses to assess the magnitude of cell apoptosis (*T + BEL–treated group significantly different from T alone–treated group, P < 0.05). B: Immunoblotting analyses. Aliquots of INS-1 cell protein (50 μg), prepared from OE cells treated with dimethylsulfoxide (lane 1) or thapsigargin (1 μmol/l) (lane 2), were analyzed by SDS-PAGE (7.5%) and transferred onto an Immobolin-P polyvinylidene difluoride membrane. The electroblot was probed with piPLA₂β antibodies, and immunoreactive protein bands were visualized by enhanced chemiluminescence, as described (41). C, control. C: Effects of caspase-3 inhibition on iPLA₂β subcellular distribution. iPLA₂β OE INS-1 cells seeded in glass chambered slides were pretreated with caspase-3 inhibitor (C3-I) (500 nmol/l) or vehicle for 24 h. The cells were then treated with thapsigargin in the absence and presence of C3-I for 24 h and subsequently processed for iPLA₂β immunofluorescence analyses by confocal microscopy. Con, control.

Caspase-3–catalyzed cleavage of iPLA₂β has been reported to occur in U937 promonocytes induced to undergo apoptosis, and the COOH-terminal product appears to be constitutively activated (23). Overexpression of the full-length iPLA₂β or the COOH-terminal caspase-3 proteolysis product of iPLA₂β (¹⁸⁴aa → COOH-terminal) in human embryonic kidney cells resulted in a higher incidence of apoptotic cell death and greater induction of arachidonic acid release from cells overexpressing the truncated iPLA₂β compared with those overexpressing full-length iPLA₂β. These findings suggest that the iPLA₂β product of a caspase-3–mediated cleavage was more active than the full-length iPLA₂β. It was proposed that the more active shorter iPLA₂β isoform could stimulate excessive turnover of nuclear membrane phospholipids, disrupt membrane fluidity, and promote apoptotic cell death.

Our findings suggest that induction of ER stress in INS-1 cells by depleting Ca²⁺ stores with thapsigargin stimulates iPLA₂β activity and promotes caspase-3–mediated cleavage of the INS-1 cell iPLA₂β. The resultant accumulation of a 62-kDa iPLA₂β-immunoreactive product in the perinuclear region of INS-1 cells raises the possibility that this might be one mechanism by which iPLA₂β participates in ER stress–induced apoptosis in β-cells.

Our studies also reveal that thapsigargin induces increased generation of ceramides in the OE cells compared with control cells. Ceramides (a family of 2-N-acylsphingosines) are important lipid second messengers that have been implicated as suppressors of cell growth and inducers of apoptosis (54). They can be generated by sphingomyelinase-catalyzed sphingomyelin hydrolysis (55), de novo synthesis (56), or as a consequence of decreased ceramidase activity (57), raising the possibility that any one of these ceramide-generating pathways might also be affected by iPLA₂β.

β-Cell iPLA₂β is a candidate for posttranslational modification

To examine the possibility that iPLA₂β undergoes posttranslational modifications that affect signaling, recombinant 84-kDa iPLA₂β was purified from sf9 cells after a sequential anion exchange and ATP affinity and calmodulin affinity chromatography protocol (58). NH₂-terminal amino acid sequencing analyses of the purified material yields a sequence that begins with residue 11 of the iPLA₂β-deduced amino acid sequence, reflecting loss of the first 10 amino acid residues. Liquid chromatography/electrospray ionization/mass spectrometry (Fig. 4A) analyses of a tryptic digest of the purified iPLA₂β reveal a peptide corresponding to residues 12–23 of the deduced amino acid sequence that would not be expected to arise from trypsin digestion of full-length iPLA₂β but that would be expected from an iPLA₂β variant lacking the first 11 amino acid residues. In the full-length iPLA₂β sequence, the first tryptic cleavage site occurs between residues 6 and 7 and the second between residues 23 and 24. Trypsin digestion thus yields peptides 1–6 and 7–23. In an iPLA₂β variant that lacks residues 1–11, trypsin digestion yields peptide 12–23. Peptide 12–23 is thus a signature peptide for this NH₂-terminally processed iPLA₂β variant. The presence of this peptide in the tryptic digest of purified iPLA₂β is reflected by the presence of [M + H]⁺¹ and [M + 2H]⁺² ions at m/z 1368.6 and 685.3, respectively. In addition, the expected aa^7–23 peptide from intact iPLA₂β is also present in the tryptic digest (Fig. 4B), as is the aa^1–6 peptide modified at the NH₂-terminal by acetylation. This NH₂-blocked peptide would not yield NH₂-terminal sequence upon Edman analyses of the intact iPLA₂β protein. Although the mechanisms responsible for generating an iPLA₂β protein lacking the first 11 amino acids have not yet been determined, these findings indicate that the β-cell iPLA₂β undergoes posttranslational modification by NH₂-terminal processing, and this might represent a means to regulate the activity, subcellular location, or protein–protein interactions of iPLA₂β.

FIG. 4.

Liquid chromatography/electrospray ionization/mass spectrometry (LC/ESI/MS) of tryptic peptide of recombinant iPLA₂β. Recombinant iPLA₂β purified by sequential chromatography was digested with trypsin, and an aliquot was analyzed by LC/ESI/MS. A: Evidence for acetylated NH₂-terminus and NH₂-terminally truncated variant of iPLA₂β. B: Evidence for sequence proximal to residue 12.

Multiple isoforms of iPLA₂β

The iPLA₂β cDNA first cloned from a rat pancreatic library encodes an 84-kDa protein, and human islets and lymphocytes were subsequently shown to express two iPLA₂β mRNA species that arise by alternate splicing and encode 84- and 88-kDa proteins (6,59). Recent studies reveal predominant expression of an iPLA₂β-immunoreactive protein with an apparent molecular mass of 70 kDa in robustly glucose-responsive 832/13 INS-1 cells (60). These cells continue to express cytosolic iPLA₂β activity that is stimulated by ATP and inhibited by BEL. Additionally, inhibition of 832/13 iPLA₂β activity with BEL suppresses glucose-stimulated insulin secretion but does not affect arachidonate incorporation into phosphatidylcholine. Therefore, the BEL-sensitive catalytic activity expressed in 832/13 INS-1 cells is most likely attributable to a 70-kDa iPLA₂β-immunoreactive protein because it is essentially the only iPLA₂β-immunoreactive protein present in the cytosol of these cells.

CONCLUSIONS

A diminished capacity of β-cells to secrete insulin in response to glucose is a prominent characteristic of type 2 diabetes, and this motivates studies to achieve a fuller understanding of glucose-sensing mechanisms within the β-cell. Our findings suggest that iPLA₂β participates in β-cell signal transduction. The expression level of iPLA₂β appears to affect β-cell proliferation and apoptosis, and the enzyme may thus be an important participant in the life cycle of the β-cell. iPLA₂β does not appear to participate in arachidonate acid incorporation into phospholipids or phosphatidylcholine homeostasis in β-cells, although such housekeeping functions of the enzyme have been proposed for other cells. The phospholipase activity of iPLA₂β would result in the production of phospholipid-derived mediators including arachidonic acid and arachidonate metabolites, and lysophospholipids and the iPLA₂β expression level also affect cellular content of the lipid second messenger ceramide. The finding that multiple iPLA₂β-immunoreactive isoforms are expressed in β-cells raises the possibility that different isoforms serve different functions that vary with the stimulation condition, subcellular compartment, phase of cell cycle, levels of interacting proteins, or other factors.

Glossary

12-HETE

12-hydroxy-(5,8,10,14)-eicosatetraenoic acid

BEL

bromoenol lactone

CT, CTP

phosphocholine cytidyltransferase

ER

endoplasmic reticulum

iNOS

inducible nitric oxide synthase

iPLA₂β

β-isoform of group VIA calcium-independent phospholipase A₂

LPA

lysophosphatidic acid

OE

over-expressing

PLA₂

phospholipase A₂