The Role of Cytosolic Phospholipase A₂ in Insulin Secretion

Human islets of Langerhans were obtained with appropriate consent from the Human Islet Transplantation Unit at King’s College London (via Dr. G.-C. Huang and Prof. S. Amiel). Rat islets were isolated by collagenase digestion of whole rat pancreas. MIN6 cells were obtained from Dr. Y. Oka and Prof. J.-I. Miyazaki (University of Tokyo) and stably transfected with vectors encoding the full-length mouse cPLA₂α cDNA (a gift from Dr. B. Kennedy, Merck Frosst, PQ, Canada) in the pcDNA3.1 vector (Stratagene Europe, Amsterdam, the Netherlands) as described (24). cPLA₂ immunoreactivity was detected by Western blotting (24) or fluorescence immunocytochemistry (25) using an anti-cPLA₂ antibody (Genetics Institute, Cambridge, MA). PLA₂ enzyme activity was measured by quantifying the release of [³H]-AA from the sn-2 position of phosphatidylcholine Lα-stearoyl-2-[³H]-arachidonyl (³H-PC). Briefly, ³H-PC (2 μmol/l, 4 Ci/mmol) was incubated (10 min, 37°C) with tissue extracts in a final volume of 25 μl Tris buffer (pH 8.0) containing 50% glycerol (vol/vol), 1 mmol/l dithiothreitol, 0.5 mmol/l NaVO₄, 0.25 mmol/l NaF, 0.1 mmol/l EDTA, 0.1 mmol/l EGTA, 20 μmol/l E64, 20 μmol/l tosyl-l-lysine chloromethyl ketone, 10 μmol/l okadaic acid, 0.1% β-mercaptoethanol (vol/vol), and 50 μg/ml leupeptin and supplemented with either 0.8 mmol/l CaCl₂ or 1 mmol/l EGTA. Reactions were terminated, and products were extracted by the addition of butanol (40 μl, 4°C). [³H]-AA was separated from an uncleaved substrate by thin-layer chromatography on silica thin-layer gels using petroleum ether:diethyl ether:glacial acetic acid 70:30:1 (vol/vol/vol) and quantified by liquid scintillation counting. Changes in [Ca²⁺]_i in MIN6 cells were assessed by single-cell microfluorimetry of Fura-2 loaded cells (26). Changes in intracellular NAD(P)H were assessed by measuring NAD(P)H autofluorescence in MIN6 cell suspensions. The rate and pattern of insulin release was assessed using a multi-chamber perifusion system at 37°C in a temperature-controlled environment (24). To maximize secretory responses, MIN6 cells were configured as pseudoislets for secretion experiments (27). Insulin content was measured by radioimmunoassay (28), and quantitative measurements of (pre)proinsulin mRNA were obtained using real-time quantitative PCR with GAPDH as an internal standard (29). Electron micrographs were prepared by the Electron Microscope Unit at King’s College London using standard techniques. Data are expressed as means ± SE and were analyzed using ANOVA, Student’s t test, and Bonferroni’s multiple comparison test as appropriate. Differences between treatments were considered significant at P < 0.05.

As has been reported previously (5,7,8), islets and β-cell lines express cPLA₂ immunoreactivity and enzyme activity. Figure 1A shows immunoreactive cPLA₂ in extracts of the mouse MIN6 insulin-secreting cell line (lanes 1 and 3) and human islets (lane 2). Whereas MIN6 cells contain only one immunoreactive protein migrating as cPLA₂α, human islets also express another lower molecular weight immunoreactivity that may represent the recently identified cPLA₂γ isoform (20). At present, it is not known whether this lower molecular weight immunoreactivity is expressed in human β-cells or in another islet cell type. The cPLA₂ enzymes are so named because, in many cell types, they are primarily localized in the cytosol, and enzyme activity is regulated through Ca²⁺-dependent translocation to membrane compartments enriched in the phosphatidylcholine substrate (18,30). Figure 1B shows Ca²⁺-dependent and Ca²⁺-independent PLA₂ enzyme activities in fractions prepared from unstimulated rat islets, where the EGTA-inhibited activity most likely reflects the Ca²⁺-dependent type IV cPLA₂ activity. The Ca²⁺-insensitive PLA₂ activity, detected in the presence of EGTA, was largely confined to the cytosol, whereas the Ca²⁺-sensitive cPLA₂ activity was found in both the cytosolic and particulate fractions.

These biochemical measurements are supported by our fluorescence immunocytochemical localization of cPLA₂ in monolayer MIN6 cells, as shown in Fig. 1C. Under unstimulated conditions (2 mmol/l glucose), cPLA₂ immunoreactivity was distributed throughout the cell, including the nuclear compartment, as shown in the left panel of Fig. 1C, and elevations in [Ca²⁺]_i caused a redistribution of the cPLA₂ immunoreactivity in MIN6 β-cells. In these experiments, exposure to tolbutamide (100 μmol/l) caused a time-dependent redistribution of cPLA₂ immunoreactivity from the nuclei of MIN6 cells to the cytoplasm (Fig. 1C, right panel), and a similar response was caused by exposure to a depolarizing concentration of KCl and, to a lesser extent, to a stimulatory concentration of glucose (25).

Although unusual, this distribution of cPLA₂ is not unique to β-cells because bovine endothelial cells are reported to exhibit a basal nuclear localization of cPLA₂, which translocates from the nucleus in response to Ca²⁺-mobilizing stimuli (31). We suggest that the redistribution of cPLA₂ in stimulated β-cells may allow agents that increase [Ca²⁺]_i to rapidly increase cPLA₂ activity in the cytoplasm by promoting its translocation from the nucleus, in a similar manner to the regulation of hepatic glucokinase by its export from the nucleus to the cytoplasm (32,33). Whatever the functional significance of the Ca²⁺-induced translocation of β-cell cPLA₂ from the nucleus, it is consistent with a stimulus-dependent activation of the enzyme, and the subsequent generation of AA, being involved in stimulus-secretion coupling. In support of this chain of events, Fig. 1D shows that exogenously applied AA (50 μmol/l) stimulated insulin secretion from human islets in the presence of a substimulatory concentration of glucose (2 mmol/l), suggesting that elevated levels of AA are sufficient to initiate a secretory response.

MIN6 cells deficient in cPLA₂α were generated by transfection with a vector coding cPLA₂ in the antisense orientation and selection for G418-resistant clones. A number of stably transfected clones showing a loss of cPLA₂ expression by immunoblotting were expanded and used for functional studies, showing a stable cPLA₂-deficient phenotype for a least 1 year in continuous culture (24). The cPLA₂-deficient cells had a much lower basal (2 mmol/l glucose) rate of insulin secretion per cell (8.6 ± 2% of passage-matched controls, P < 0.01, n = 4). However, the ability of the cells to recognize and respond to elevations in the extracellular glucose concentration was not markedly affected by the loss of cPLA₂, as shown in Fig. 2A. In these perifusion experiments, the secretory responses of cPLA₂-deficient cells to increased glucose were similar to those of control cells when expressed relative to the basal secretory rate, showing a rapid increase in insulin release upon elevation of glucose from 2 to 20 mmol/l and maintaining an elevated rate of secretion thereafter. The cPLA₂-deficient cells also showed normal patterns of insulin secretion in response to nonnutrient stimuli that elevate [Ca²⁺]_i, such as tolbutamide and KCl (24). These data imply that cPLA₂α is not required for pancreatic β-cells to show an initial secretory response to nutrient and nonnutrient secretagogues, suggesting that it is not a pivotal Ca²⁺-sensitive sensor in β-cell stimulus-secretion coupling. It has been reported previously that the amplifying (ATP-dependent K⁺ channel-independent) effect of glucose on insulin secretion from mouse islets does not require PLA₂ activation and AA generation (23); therefore, it seems unlikely that cPLA₂ plays an important role in glucose-induced insulin secretion.

Although they appeared able to respond normally to nutrient and nonnutrient stimuli, the cPLA₂-deficient cells were unable to mount prolonged secretory responses to maintained stimulation, as shown in Fig. 2B. Thus, a subsequent exposure to the adenylate cyclase activator forskolin in the presence of 20 mmol/l glucose caused a further increase in the rate of insulin secretion that was maintained in control cells, but was both lesser (∼8-fold vs. 18-fold) and transient in cPLA₂-deficient cells, returning to preforskolin rates within ∼10 min. The inability to maintain a high rate of secretion was independent of the stimulus, and similar results were obtained with the protein kinase C activator phorbol myristate acetate in combination with glucose, tolbutamide, or KCl (24). The lower basal rate of insulin secretion and the inability to maintain secretion correlated well with the severely reduced insulin content of the cPLA₂-deficient MIN6 cells: the clone used for the secretion experiments in Fig. 2 contained only 8.1 ± 1.6% (n = 5) of the insulin content of control MIN6 cells transfected with an empty vector. We therefore suggest that the cPLA₂ underexpressing β-cells have sufficient intracellular stores of insulin for an initial secretory response (albeit reduced in mass in accordance with the reduced insulin content) but insufficient to maintain secretion in response to a strong prolonged stimulus.

It is not immediately apparent why cPLA₂ deficiency should cause such marked reductions in insulin content, but it cannot be accounted for by reduced insulin gene transcription because quantitative RT-PCR demonstrated similar levels of preproinsulin mRNA in cPLA₂-deficient and control cells (cPLA₂-deficient 84 ± 16% control, n = 3), suggesting a defect in translation, processing, or packaging of (pre)proinsulin. In other tissues, cPLA₂ has been implicated in maintaining the stability of the Golgi apparatus to enable appropriate intracellular vesicle trafficking (34), and this would be consistent with reported changes in ultrastructure of cPLA₂-deficient MIN6 cells (24). Thus, in accordance with their reduced insulin content, these cells contained very few typical dense-cored insulin-containing secretory vesicles and large numbers of nonelectron dense vesicles (24). Further studies are required to determine whether the reduced content of insulin-containing dense-cored secretory vesicles is due to defective packaging of the vesicles or to defective processing and crystallization of insulin.

In summary, the functional phenotype of β-cells deficient in cPLA₂α does not support a vital role for cPLA₂ as a Ca²⁺-sensitive transduction element in the β-cell stimulus-secretion coupling pathway, but implicates cPLA₂ in the maintenance of insulin stores in dense-cored secretory vesicles. These conclusions are not entirely inconsistent with some of the earlier studies that provided the rationale for the current work. Thus, physiological stimuli that enhance insulin secretion must also enhance the production and storage of more insulin to maintain the β-cell secretory capacity. The activation/redistribution of cPLA₂ in response to insulin secretagogues may be involved in this element of stimulus-response coupling, rather than in the immediate regulation of the secretory process.

To complement our studies on cPLA₂-depleted β-cells, we have started to investigate the functional consequences of constitutive overexpression of cPLA₂α by generating stably transfected MIN6 cells expressing the full-length mouse cDNA inserted into the pcDNA3.1 expression vector in the sense orientation. Overexpression of cPLA₂α was driven by the cytomegalovirus promoter and produced high and stable levels of expression of immunoreactive cPLA₂, as assessed by Western blotting. As for the cPLA₂-deficient MIN6 cells, the functional phenotype of the overexpressing clones was not what we had expected. Thus, overexpression of cPLA₂ had no marked effect on insulin content nor on the basal rate of insulin secretion at a substimulatory glucose concentration (2 mmol/l). However, the cPLA₂-overexpressing clones failed to show insulin secretory responses to glucose, as shown in Fig. 3A. The lack of secretory responses to glucose was accompanied by a loss of the normal rapid increases in [Ca²⁺]_i induced by glucose (data not shown) and by mitochondrial substrates such as ketoisocaproic acid, as shown in Fig. 3B, perhaps suggesting a mitochondrial defect. The cells retained normal [Ca²⁺]_i responses to depolarizing concentrations of KCl (Fig. 3B) and tolbutamide (data not shown), localizing the defect in the stimulus-response coupling pathway at a stage beyond glucose entry and phosphorylation but before depolarization of the β-cells and the subsequent influx of extracellular Ca²⁺. This conclusion was supported by the demonstration that cPLA₂-overexpressing cells had relatively normal, although slightly delayed, onset and secretory responses to depolarizing [Ca²⁺]_i-elevating stimuli, as shown in Fig. 3C. Incidentally, the normal pattern of both basal secretion and of secretion in response to elevated [Ca²⁺]_i in cells that greatly overexpress cPLA₂α further reinforces the conclusion from the studies with cPLA₂-deficient MIN6 cells that cPLA₂α is not an important transducer of Ca²⁺-induced secretory responses in β-cells.

Measurements of NAD(P)H autofluorescence as an indirect assessment of oxidative phosphorylation revealed a major metabolic defect in the cPLA₂-overexpressing cells. Thus, exposure to 25 mmol/l glucose caused a rapid and reversible increase in NAD(P)H autofluorescence in control MIN6 cells that was almost completely absent in the cPLA₂-overexpressing cells (data not shown). These observations may explain the defective secretory phenotype of the cPLA₂-overexpressing cells, since uncoupling nutrient metabolism from ATP production will prevent the ATP-dependent depolarization that initiates the insulin secretory response. The underlying cause of this defective metabolism is likely to be the chronic overproduction of the long-chain unsaturated fatty acid AA by the overexpressed cPLA₂α in the sense transfects. There is a growing body of evidence that prolonged exposure to fatty acids can compromise β-cell function (35,36), with inhibition of glucose-induced insulin secretion being caused by sustained exposure to high levels of free fatty acids (37,38). The inhibitory effects on secretion appear to be confined to nutrient stimuli, with normal responses being maintained to nonnutrients (37,38), suggesting that excessive exposure to fatty acids perturbs β-cell metabolic pathways proximal to elevated [Ca²⁺]_i, in accordance with the secretory phenotype of our cPLA₂-overexpressing MIN6 cells. One link between fatty acid excess and mitochondrial dysfunction is uncoupling protein 2 (UCP2), which dissipates mitochondrial proton gradients, thus increasing the activity of respiratory chain complex enzymes and the oxidation of NADH. High concentrations of fatty acids, including AA (39,40), can upregulate expression of UCP2 in a number of tissues, including β-cells (41,42), and UCP2 upregulation is associated with reduced nutrient-induced insulin secretion (43–45) Our measurements of UCP2 mRNA demonstrate an upregulation of UCP2 expression in the cPLA₂-overexpressing cells, with mRNA levels of up to 256 ± 28% control values (P < 0.001), presumably in response to the high local concentrations of AA in these cells. These results suggest that the uncoupling of oxidative phosphorylation from mitochondrial ATP generation is responsible for the defective secretory phenotype of the cPLA₂-overexpressing cells and are consistent with UCP2 upregulation being involved in β-cell dysfunction in type 2 diabetes (44,45). The cPLA₂-overexpressing MIN6 cells may offer a useful experimental model in which to study the effects of chronic exposure to fatty acids on β-cell (dys)function.

In conclusion, our expectation at the outset of these studies was to confirm an important role for cPLA₂α as a Ca²⁺ sensor in β-cell stimulus-secretion coupling. However, our results did not support our preconceptions and suggest that cPLA₂α is not required for Ca²⁺-dependent insulin secretion, although it may play an important, but poorly defined, role in the maintenance of insulin stores in dense-cored secretory vesicles. The unexpected phenotype associated with excessive cPLA₂ activity may give further insights into fatty acids, mitochondrial dysfunction, and the failure of β-cell secretory responses.

We are grateful to Diabetes UK for project grant support (RD97/0001525). C.J.B. is a Diabetes Research and Wellness Foundation fellow. V.D.B. was funded by the Eli Lilly International Foundation. H.M.R.-M. was funded by a Biotechnology and Biological Sciences Research Counsel Animal Sciences Committee postgraduate studentship.

We are grateful to Dr. Paul Squires (Warwick University, U.K.) for measurements of [Ca²⁺]_i, to Dr. K.J. Parker (King’s College London) for the PLA₂ enzyme assay, and to Nicholas Evans (King’s College London) for measurements of NAD(P)H autofluorescence in MIN6 cells. We thank Dr. Y. Oka and Prof. J.I. Miyazaki (University of Tokyo) for provision of the MIN6 cells, Dr. B. Kennedy (Merck Frosst, PQ, Canada) for the cPLA₂ cDNA, Dr. R. Kramer (Eli Lilly, Indianapolis, IL) for the cPLA₂ antiserum, and Dr. G.-C. Huang and Prof. S. Amiel (King’s College London) for the human islets.