INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THROMBOTIC CORONARY ARTERY occlusion has been demonstrated to contribute directly to arrhythmogenesis during myocardial ischemia, suggesting that products released from or associated with an intracoronary thrombus may directly or indirectly influence the electrophysiological properties of ischemic cardiac myocytes (12). In isolated ventricular myocytes, thrombin has been shown to stimulate phospholipase A₂ (PLA₂) activity (25) and generate lysophospholipids (24), which are potent arrhythmogenic amphiphilic lipid metabolites (29). Thus one of the potential factors contributing to arrhythmogenesis in the setting of myocardial ischemia may be the direct action of thrombin on ischemic myocytes, resulting in PLA₂ activation and amphiphilic lipid metabolite generation.

Numerous types of mammalian PLA₂s have been identified and classified into several groups, each of which demonstrates unique characteristics (11). Secretory PLA₂s (sPLA₂s) require millimolar concentrations of Ca²⁺ for activity and consequently are only active extracellularly. At least two types of PLA₂ are active intracellularly, a Ca²⁺-dependent cytosolic PLA₂ (cPLA₂) that requires micromolar concentrations of Ca²⁺ for translocation to intracellular membranes and a Ca²⁺-independent PLA₂ (iPLA₂). The three types of PLA₂ have been shown to coexist in mammalian cells and may interact with each other (2). To date, published evidence suggests that iPLA₂ may be involved in cell signaling and/or phospholipid remodeling, which appears to depend upon both the stimulus and the cell type involved (1, 2, 16, 24-26, 32).

The phospholipid composition of isolated cardiac myocytes is unique in being composed predominantly of plasmalogen molecular species (7, 28). Furthermore, plasmalogens are enriched in the electrically active membranes, the sarcolemma and sarcoplasmic reticulum (13, 14), and are important in modulating the kinetic properties of ion channels and ion transporters (10, 37). Previous studies suggest also that plasmalogen phospholipids may be the preferred substrates for ischemia-activated phospholipases (17). A plasmalogen-selective PLA₂ has been described in several tissues, including myocardium (17, 25, 26), brain (18), and kidney (31), and may be involved in accelerated plasmalogen hydrolysis in response to agonist stimulation or ischemia. Accelerated plasmalogen hydrolysis in the ischemic myocardium may lead to the accumulation of lysoplasmenylcholine, a metabolite that has been shown to have profound effects on the electrophysiological properties of isolated cardiac myocytes (17, 28), which could lead to arrhythmogenesis in the ischemic heart.

The present study was undertaken to determine which individual phospholipid molecular species serve as endogenous substrates for PLA₂ and to determine whether increased phospholipid hydrolysis leads to bioactive metabolite accumulation in thrombin-stimulated rabbit ventricular myocytes that may directly or indirectly contribute to the production of malignant ventricular arrhythmias in the ischemic heart.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Materials. Methyl arachidonyl fluorophosphonate (MAFP) and arachidonyltrifluoromethyl ketone (AACOCF₃) were purchased from Calbiochem (La Jolla, CA). Bromoenol lactone (BEL) was a gift from Hoffmann-La Roche (Nutley, NJ). Collagenase (type II) was purchased from Worthington Biochemical (Lakewood, NJ). [³H]acetic acid, [³H]acetic anhydride, and [¹⁴C]acetic anhydride were purchased from NEN (Boston, MA); [¹⁴C]lysophosphatidylcholine was purchased from Amersham (Arlington Heights, IL). Thrombin (isolated from human plasma and essentially free of other known clotting factors, plasminogen and plasmin) and 17:0 L--lysophosphatidylcholine were purchased from Sigma Chemical (St. Louis, MO). Lysoplasmenylcholine was prepared by alkaline hydrolysis of bovine heart choline glycerophospholipids, as described previously (25). The majority of miscellaneous reagents were purchased from Sigma.

Isolation of ventricular myocytes. Adult rabbits of either sex weighing 2-3 kg were anesthetized with intravenous pentobarbitone sodium (50 mg/kg), and the heart was rapidly removed. The heart was mounted on a Langendorff perfusion apparatus and perfused for 5.5 min with a Tyrode solution containing (in mmol/l) NaCl 118, KCl 4.8, CaCl₂ 1.2, MgCl₂ 1.2, NaHCO₃ 24, KH₂PO₄ 1.2, and glucose 11; the Tyrode solution was saturated with 95% O₂-5% CO₂ to yield a pH of 7.4. This was followed by a 3.5-min perfusion with a Ca²⁺-free Tyrode solution containing EGTA (100 µM) and a final perfusion for 20 min with the Tyrode solution containing 100 µM Ca²⁺ and 0.033% collagenase. The left and right ventricles were cut into small pieces and placed in two Erlenmeyer flasks containing fresh enzyme solution, which were then shaken in a Dubnoff metabolic shaker at 37°C for 15 min, with 95% O₂-5% CO₂ blowing into each flask. The first harvest of myocytes was discarded. Cells from the next three harvests were combined and washed with a HEPES buffer containing (in mmol/l) NaCl 133.5, KCl 4.8, MgCl₂ 1.2, KH₂PO₄ 1.2, HEPES 10, and glucose 10 plus 300 µM CaCl₂, with pH adjusted to 7.4 with 10 N NaOH. Extracellular Ca²⁺ concentration was increased to 1.2 mM in three stages at intervals of 20 min. Elongated myocytes were separated from rounded nonviable cells by repeated differential sedimentation.

Stimulation of myocytes with thrombin. Aliquots of myocytes (2 × 10⁶ in 2 ml of 1.2 mM Ca²⁺ HEPES buffer) were incubated at 37°C. AACOCF₃, BEL, or MAFP was dissolved in DMSO and diluted at least 1 in 1,000 in HEPES buffer before addition to the cells. Thrombin was dissolved directly in HEPES buffer and added to the myocytes with or without prior PLA₂ inhibition. At the end of the stimulation period, chloroform and methanol were added directly to the myocyte suspension for measurement of phospholipids, lysophospholipids, and platelet-activation factor (PAF). For measurement of PLA₂ activity, the surrounding buffer was removed and immediately replaced with ice-cold buffer containing (in mmol/l) sucrose 250, KCl 10, imidazole 10, EDTA 5, and dithiothreitol (DTT) 2 with 10% glycerol, pH = 7.8 (PLA₂ assay buffer). For measurement of 6-keto-PGF₁, the myocyte suspension was centrifuged and the supernatant was removed for assay of 6-keto-PGF₁ content.

Extraction, separation, and analysis of phospholipid classes. Cellular phospholipids were extracted from isolated adult rabbit ventricular myocytes by the method of Bligh and Dyer (4) at 0-4°C. The chloroform layer was dried under N₂, and the lipid residue was resuspended in 1 ml chloroform-methanol (1:1 vol/vol). Three 5-µl aliquots were removed for measurement of total lipid phosphorus, and 200-µl aliquots were injected onto an Ultrasphere-Si (5 µm silica), 4.6 × 250-mm HPLC column (Beckmann Instruments). Phospholipids were separated into different classes on the basis of differences in polar headgroup composition with the use of gradient elution with a mobile phase composed of hexane-isopropanol-water as described previously (7). Phospholipid classes were quantified in the isolated fractions by measurement of lipid phosphorus by microphosphate assay (6). The fatty acid composition of the isolated glycerophospholipid classes was determined by gas chromatographic (GC) analysis of the fatty acid methyl ester (FAME) and dimethylacetal (DMA) derivatives produced after acid-catalyzed methanolysis. Identification of individual FAME species was established by comparison of their GC retention times with commercial standards (Alltech, Deerfield, IL). Individual DMA species were identified by comparison of their GC retention times with the DMA derivatives produced after acid-catalyzed methanolysis of lysoplasmenylcholine derived from bovine heart choline glycerophospholipids (7). The alkylacyl glycerophospholipid content of phosphatidylcholine and phosphatidylethanolamine was determined by quantification of lipid phosphorus in the lysophospholipid fraction remaining after sequential, exhaustive base- and acid-catalyzed hydrolysis of the diradylphospholipids (13).

Separation and quantification of individual choline and ethanolamine glycerophospholipid molecular species. Individual choline and ethanolamine glycerophospholipid molecular species were isolated by reverse-phase HPLC with the use of an Ultrasphere ODS (5 µm, C-18) column, 4.6 × 250 mm (Beckmann Instruments). Individual molecular species were separated with the use of a gradient elution system with a mobile phase composed of acetonitrile-methanol-water with 20 mM choline chloride (27). The molecular identity of individual molecular species was established by GC characterization of the FAME and DMA derivatives produced after acid-catalyzed methanolysis of the phospholipid species recovered in column effluents and by comparison of absolute retention time, relative retention time, and order of elution of individual species with previously injected synthetic phospholipids of known composition (27). Quantification of individual phospholipid molecular species was achieved by determination of lipid phosphorus in reverse-phase HPLC column effluents by the method of Itaya and Ui (19).

PLA₂ activity. Myocytes suspended in ice-cold PLA₂ assay buffer were sonicated on ice six times for 10 s, and the sonicate was centrifuged at 14,000 g for 10 min. The resultant supernatant fraction was centrifuged at 100,000 g for 60 min to separate the membrane fraction (pellet) from the cytosolic fraction (supernatant). The membranes were washed twice and resuspended in PLA₂ assay buffer. PLA₂ activity in subcellular fractions was assessed by incubating enzyme (8 µg membrane protein or 200 µg cytosolic protein) with 100 µM (16:0, [³H]18:1) plasmenylcholine, phosphatidylcholine, or alkylacyl glycerophosphorylcholine (24) in assay buffer containing 100 mM Tris, 4 mM EGTA, and 10% glycerol, pH = 7.0, at 37°C for 5 min in a total volume of 200 µl. Reactions were terminated by the addition of 100 µl butanol, and then tubes were vortexed and centrifuged at 2,000 g for 5 min. Released radiolabeled fatty acid was isolated by application of 25 µl of the butanol phase to channeled silica gel G plates, development in petroleum ether-diethyl ether-acetic acid (70:30:1 vol/vol/vol), and subsequent quantification by liquid scintillation spectrometry. The reaction conditions selected resulted in linear reaction velocities with respect to both time and total protein concentration for each substrate examined. Protein content of each sample was determined by the Lowry method with the use of freeze-dried bovine serum albumin as the protein standard, as described previously (23).

Measurement of choline lysophospholipids. Lysophosphatidylcholine and lysoplasmenylcholine measurements were made using a modification of a radiometric assay method described previously (28). Lipids were extracted from the myocyte suspension by the method of Bligh and Dyer (4), and lysophospholipids were separated from other phospholipids by HPLC. The purified lysophosphatidylcholine and lysoplasmenylcholine fractions as well as known amounts of lysophosphatidylcholine and lysoplasmenylcholine standards were then acetylated with [³H]acetic anhydride with the use of 0.33 M dimethylaminopyridine as a catalyst. The acetylated lysophospholipid was then separated by TLC, and radioactivity was quantified by liquid scintillation spectrometry. Standard curves were constructed, and lysophosphatidylcholine and lysoplasmenylcholine content were derived for all samples and normalized according to the protein content of the myocytes as described previously (23). [¹⁴C]lysophosphatidylcholine was added as an internal standard to all samples to correct for loss of sample that occurred during extraction, purification, and acetylation.

PGI₂ formation. After thrombin stimulation, the myocyte suspension was rapidly centrifuged, and the supernatant was removed. PGI₂ released from ventricular myocytes was measured in surrounding medium as its stable metabolite 6-keto-PGF₁. The 6-keto-PGF₁ in the sample was measured using a commercially available immunoassay kit (R&D Systems, Minneapolis, MN). The myocyte cell pellet was resuspended in distilled water, and the amount of protein was determined as described previously (23).

Production of PAF. Myocytes (10⁵ in 1 ml) were incubated with 50 µCi [³H]acetic acid for 20 min. After thrombin stimulation for the selected time interval, lipids were extracted from the cell suspension by the method of Bligh and Dyer (4). The chloroform layer was concentrated by evaporation under N₂, applied to a silica gel 60 TLC plate, and developed in chloroform-methanol-acetic acid-water (50:25:8:4 vol/vol/vol/vol). The region corresponding to PAF was scraped and radioactivity was quantified with the use of liquid scintillation spectrometry. Loss of PAF during extraction and chromatography was corrected by adding a known amount of [¹⁴C]PAF as an internal standard. [¹⁴C]PAF was synthesized by acetylating the sn-2 position of lyso-PAF with [¹⁴C]acetic anhydride (Amersham, Arlington Heights, IL) with the use of 0.33 M dimethylaminopyridine as a catalyst. The synthesized [¹⁴C]PAF was purified by HPLC. The specific activity of the final [¹⁴C]PAF product was 300 µCi/µmol. The amount of protein in each sample was measured as described previously (23).

Statistics. Statistical comparison of values was performed by Student's t-test or analysis of variance with the Fisher multiple-comparison test as appropriate. All results are expressed as means ± SE. Statistical significance was considered to be P < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

We have demonstrated previously that the majority of PLA₂ activity in isolated rabbit cardiac myocytes is Ca²⁺ independent and membrane associated and exhibits a preference for arachidonylated phospholipid substrates (25, 26). Thrombin stimulation resulted in an increase in membrane-associated PLA₂ activity measured in the absence of Ca²⁺ with the use of either plasmenylcholine, phosphatidylcholine, or alkylacyl glycerophosphorylcholine substrates (Fig. 1). Thrombin-stimulated PLA₂ activity was inhibited by pretreatment with BEL (a specific inhibitor of myocardial iPLA₂) but was unaffected by pretreatment with either AACOCF₃ or MAFP (inhibitors of both Ca²⁺-dependent and -independent cytosolic PLA₂s) (Fig. 1).

View larger version (32K): [in this window] [in a new window]   Fig. 1.   Effect of pretreatment of rabbit ventricular myocytes with specific phospholipase A2 (PLA2) inhibitors (10-min incubation) on membrane-associated PLA2 under basal conditions or after thrombin stimulation (0.05 U/ml, 1 min) with use of (16:0, [3H]18:1) plasmenylcholine, (16:0, [3H]18:1) phosphatidylcholine, or (16:0, [3H]18:1) alkylacyl glycerophosphorylcholine in absence of Ca2+ (4 mM EGTA). Values are means ± SE of independent results derived from 6 separate animals. MAFP, methyl arachidonyl fluorophosphonate; AACOCF3, arachidonyltrifluoromethyl ketone; BEL, bromoenol lactone. ** P < 0.01 compared with corresponding PLA2 activity in absence of inhibitor.

To determine which phospholipids serve as endogenous substrates for the membrane-associated iPLA₂, we subsequently characterized the phospholipid molecular species composition of control and thrombin-stimulated ventricular myocytes. The effect of inhibition of iPLA₂ with BEL on phospholipid hydrolysis was also investigated.

The total phospholipid phosphorus in isolated rabbit ventricular myocytes was 141 ± 8 nmol/mg protein. The major phospholipid classes were choline and ethanolamine glycerophospholipids (Table 1). These were also the only classes found to contain plasmalogens or alkylacyl glycerophospholipids (Table 1). Choline glycerophospholipids were composed of 45% plasmenylcholine and 55% phosphatidylcholine, whereas ethanolamine glycerophospholipids were composed of 50% plasmenylethanolamine (Table 1 and Fig. 2). Thrombin stimulation with or without pretreatment with BEL had no effect on the total mass of choline or ethanolamine phospholipids.

View larger version (27K): [in this window] [in a new window]   Fig. 2.   Changes in mass of plasmenylcholine molecular species after thrombin stimulation (0.05 U/ml, 2 min) of isolated rabbit ventricular myocytes. Pretreatment with bromoenol lactone (BEL; 10 µM, 10 min) completely blocked catabolism of arachidonylated plasmenylcholine. Values are means ± SE of duplicate measurements for independent results obtained from 6 separate animals. * P < 0.05 and ** P < 0.01 compared with untreated myocytes.

In choline glycerophospholipids, arachidonylated species were predominantly found in plasmalogens; 55% of plasmenylcholine contained arachidonic acid at the sn-2 position, but only 24% of phosphatidylcholine species were arachidonylated (Table 2). Thrombin stimulation of ventricular myocytes resulted in a significant decrease in the mass of arachidonylated plasmenylcholine (Fig. 3A) that reflected decreases in both (16:0, 20:4) and (18:1, 20:4) plasmenylcholine (Table 2). Thrombin-stimulated ventricular myocytes demonstrated no corresponding change in arachidonylated phosphatidylcholine (Table 2). The decrease in arachidonylated plasmenylcholine was accompanied by an increase in plasmenylcholine species containing linoleic (18:2) and linolenic (18:3) acids at the sn-2 position (Table 2 and Fig. 2), suggesting that the majority of lysoplasmenylcholine produced by PLA₂ in response to thrombin stimulation in ventricular myocytes may be rapidly reacylated by CoA-dependent acyltransferase and/or CoA-independent transacylase enzymes. Pretreatment with BEL before thrombin stimulation abolished completely any thrombin-induced alterations in phospholipid molecular species (Fig. 2), demonstrating that these alterations are the direct result of PLA₂ activation and not the result of metabolite flux through alternate, competing pathways.

View larger version (52K): [in this window] [in a new window]   Fig. 3.   Changes in mass of plasmenylethanolamine (A) and phosphatidylethanolamine (B) molecular species after thrombin stimulation (0.05 U/ml, 2 min) of isolated rabbit ventricular myocytes. Pretreatment with BEL (10 µM, 10 min) completely blocked catabolism of arachidonylated plasmenylethanolamine (A). Values are means ± SE of duplicate measurements for independent results obtained from 6 separate animals. * P < 0.05 compared with untreated myocytes.

The majority of ethanolamine glycerophospholipid was found to be arachidonylated, with 62% of plasmenylethanolamine and 73% of phosphatidylethanolamine containing arachidonic acid at the sn-2 position (Fig. 3). Thrombin stimulation of ventricular myocytes resulted in a decrease in total arachidonylated plasmenylethanolamine (Fig. 3A) that represented a decrease in the mass of each individual arachidonylated molecular species (Table 2). No change in phosphatidylethanolamine species was observed in thrombin-stimulated myocytes (Fig. 3B). The decrease in arachidonylated plasmenylethanolamine in thrombin-stimulated myocytes was significantly less than that seen in the choline phospholipid class and was accompanied by an increase in (18:1, 18:2) and (16:0, 18:3) plasmenylethanolamine (Table 2, Fig. 3A). Pretreatment with BEL blocked any thrombin-stimulated changes in ethanolamine glycerophospholipids (Fig. 3A), again confirming that these alterations are mediated by PLA₂ activation. Thus thrombin stimulation of isolated rabbit ventricular myocytes results in the selective hydrolysis of both choline and ethanolamine plasmalogen phospholipids by iPLA₂ with a preference for those plasmalogen species possessing a phosphocholine headgroup.

View larger version (29K): [in this window] [in a new window]   Fig. 4.   Lysoplasmenylcholine and lysophosphatidylcholine content in control and thrombin-stimulated (0.05 U/ml, 1 min) ventricular myocytes. Pretreatment with BEL (10 µM, 10 min) significantly inhibited basal lysoplasmenylcholine and blocked completely increase in lysoplasmenylcholine produced by thrombin stimulation. Values are means ± SE for independent results from 6 separate animals. ** P < 0.01 compared with unstimulated myocytes.

The selective hydrolysis of arachidonic acid from the sn-2 position of plasmalogen phospholipids in thrombin-stimulated myocytes suggests the selective concomitant production of lysoplasmalogens, which may or may not accumulate depending on the presence and activity of other enzymes catalyzing lysoplasmalogen catabolism or reacylation. Measurement of choline lysophospholipid content in thrombin-stimulated ventricular myocytes demonstrated a significant increase in lysoplasmenylcholine that was measured after 1 min of thrombin stimulation with no corresponding increase in lysophosphatidylcholine (Fig. 4). The increase in lysoplasmenylcholine was inhibited completely by BEL pretreatment (Fig. 4). Incubation of ventricular myocytes with thrombin for increasing time intervals resulted in a significant increase in lysoplasmenylcholine content for up to 5 min, after which time, lysoplasmenylcholine returned to basal values. The increase in lysoplasmenylcholine content was ~0.4 nmol/mg protein (Fig. 4), and there was an increase of 7-8 nmol/mg protein in choline glycerophospholipids containing either linoleic or linolenic acid at the sn-2 position (Fig. 2). Because PLA₂-catalyzed hydrolysis of plasmenylcholine resulted in a decrease of 8-10 nmol/mg protein of arachidonylated molecular species (Fig. 2), the majority of arachidonylated plasmalogen phospholipid hydrolyzed by thrombin-activated PLA₂ was reacylated, thereby limiting the accumulation of lysoplasmenylcholine within the cell.

The products formed after PLA₂ hydrolysis of membrane phospholipids, namely arachidonic acid and lysophospholipids, may affect membrane function directly (28, 29) or serve as precursors for biologically active metabolites. Lysophospholipids can serve as PAF precursors, whereas arachidonic acid is the precursor for eicosanoid production. To determine whether there is further metabolism of lysoplasmenylcholine and arachidonic acid to other bioactive products after thrombin stimulation of ventricular myocytes, PAF and prostacyclin production were measured under control and thrombin-stimulated conditions. Thrombin stimulation led to some increase in PAF production, but this was not statistically significant (Fig. 5), suggesting that little lysophospholipid is acetylated after production. It should be noted that with the use of this assay system, the amount of radiolabeled PAF produced is a relative measure of the acetylation of lysophospholipid rather than a direct measurement of the mass of biologically active PAF produced. Further metabolism of free arachidonic acid in ventricular myocytes was evidenced by the significant increase in prostacyclin production after thrombin stimulation (Fig. 6). This increase was blocked by inhibiting iPLA₂ with BEL (Fig. 6).

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Platelet-activating factor (PAF) production in thrombin-stimulated (1 U/ml) and unstimulated isolated rabbit ventricular myocytes. Values are means ± SE for independent results from 10 separate animals.

View larger version (22K): [in this window] [in a new window]   Fig. 6.   Production of prostacyclin in isolated rabbit ventricular myocytes after thrombin stimulation (0.05 U/ml). Increase in prostacyclin in thrombin-stimulated myocytes was blocked by pretreatment with BEL (10 µM, 10 min). Prostacyclin content of unstimulated myocytes is represented by dotted line. Values are means ± SE of duplicate measurements for independent results from 4 separate animals. ** P < 0.01 compared with untreated myocytes.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Previous studies (24-26) have demonstrated that thrombin stimulation of rabbit ventricular myocytes results in a selective increase in membrane-associated PLA₂ activity. On the basis of in vitro assay measurements, thrombin-stimulated PLA₂ is able to hydrolyze plasmenylcholine, phosphatidylcholine, and alkylacyl glycerophosphorylcholine substrates in the absence of Ca²⁺; thus thrombin-stimulated PLA₂ does not exhibit a catalytic requirement for Ca²⁺. However, it should be noted that myocytes suspended in a Ca²⁺-free buffer and stimulated with thrombin do not demonstrate increased PLA₂ activity (data not shown), suggesting that although the catalytic activity of thrombin-stimulated PLA₂ is Ca²⁺ independent, activation of PLA₂ in intact myocytes requires the presence of intracellular Ca²⁺, possibly for the function of intracellular regulators of iPLA₂. Thrombin-stimulated iPLA₂ represents an integral membrane protein, because membrane-associated PLA₂ activity cannot be removed by repeated sonication, washing in the presence of EGTA, high- or low-salt concentrations, or submicellar concentrations of detergent (26). The increase in membrane-associated PLA₂ activity appears to be a result of activation of a latent membrane-associated PLA₂, because increases in activity are observed after 30 s of thrombin stimulation, and no change in the amount of membrane-associated iPLA₂ protein is observed by immunoblot analysis in thrombin-stimulated cells (data not shown).

Accompanying the increase in PLA₂ is an increase in free arachidonic acid release from the myocytes and an accumulation of total choline lysophospholipids. Previous studies have not identified the endogenous diradyl phospholipid substrates targeted for hydrolysis by thrombin-stimulated PLA₂. Similarly, previous measurements of thrombin-induced lysophospholipid production did not distinguish between lysophospholipids derived from diacyl phospholipids and lysoplasmalogens derived by PLA₂-catalyzed plasmalogen hydrolysis. In addition, production of other "downstream" bioactive metabolites (e.g., PAF and eicosanoids) in response to thrombin activation of iPLA₂ has not been investigated.

In this study, we have demonstrated that plasmalogen phospholipids serve as the endogenous substrates for thrombin-stimulated PLA₂ activity. Thus, although in vitro measurements of PLA₂ activity demonstrate increased hydrolysis of both plasmalogen and diacyl phospholipids, in vivo hydrolysis of ventricular myocyte phospholipids in response to thrombin stimulation occurs almost exclusively in plasmalogens. This suggests that plasmalogens serve as a highly metabolically active pool of membrane phospholipids that are readily hydrolyzed in response to agonist stimulation and may play a role in second messenger generation during signal transduction. Plasmalogens are the predominant phospholipids found in the sarcolemma and sarcoplasmic reticulum of myocardial cells (13, 14). The physiological consequences of plasmalogens in these subcellular pools remain to be fully established. However, plasmalogens play an important role in regulating transmembrane concentration gradients and myocardial electrical excitability (37). Additionally, plasmalogens have been demonstrated to regulate sodium-calcium exchange in cardiac sarcolemmal preparations (10), suggesting a role for plasmalogens in ion-transport regulation.

Myocardial ischemia is associated with the accelerated selective hydrolysis of plasmenylcholine by phospholipase C (9) and PLA₂ (17, 28), suggesting the activation of multiple pathways for plasmalogen hydrolysis. The accumulation of lysoplasmenylcholine is the direct, specific result of accelerated PLA₂-catalyzed hydrolysis of plasmalogen phospholipids and may be accompanied by other effects on the myocardium. For example, lysoplasmenylcholine has been shown to activate myocardial cAMP-dependent protein kinase, suggesting a role for lysoplasmenylcholine in signal transduction analogous to that of diacylglycerol activation of protein kinase C (35). Additionally, lysoplasmenylcholine has been shown to have direct effects on the function of integral proteins within the membrane. Schonefeld et al. (34) have demonstrated a greater degree of inhibition of kidney cortex Na⁺-K⁺-ATPase with lysoplasmenylcholine than that observed with the structurally similar amphiphilic compounds lysophosphatidylcholine and palmitoylcarnitine. We have shown that lysoplasmenylcholine produces action potential derangements in cardiac myocytes (28) at much lower concentrations than that described previously for either lysophosphatidylcholine or palmitoylcarnitine (29), suggesting a direct interaction of lysoplasmenylcholine with ionic channels in the membrane. The action potential derangements caused by lysoplasmenylcholine occur as a result of the action of this amphiphilic compound on multiple membrane currents (5, 22). The increase in lysoplasmenylcholine observed in thrombin-stimulated myocytes (0.4 nmol/mg protein) is identical to the increase in lysophosphatidylcholine measured in the ischemic myocardium in vivo (0.5 nmol/mg protein) and the increase in vitro (0.4 nmol/mg protein) required to produced electrophysiological alterations (29). Thus lysoplasmenylcholine is an important bioactive metabolite independent of its metabolism to other products.

PAF has also been shown to exert multiple effects on the myocardium, either through a direct action modifying chronotropic and ionotropic activity or indirectly via the release of eicosanoids or cytokines (8). Functional PAF receptors have been found in ventricular myocytes (3); however, the ability of cardiac myocytes to directly produce PAF themselves has not been demonstrated previously. Thrombin stimulation of rabbit ventricular myocytes did not lead to increased PAF production, suggesting that little, if any, of the lysophospholipid produced in our studies was further metabolized by acetylation. Our studies suggest that PAF-mediated effects on the myocardium are a result of response to PAF released from cells other than the cardiac myocytes, such as neutrophils, platelets, or endothelial cells.

The synthesis of eicosanoids in the heart may also have direct inotropic or chronotropic effects and may have been implicated in several pathological conditions in the heart (21). Results from our studies illustrate a significant increase in prostacyclin production in thrombin-stimulated myocytes that is completely abrogated by BEL pretreatment, which blocks both membrane-associated iPLA₂ activity and the resultant arachidonic acid release.

iPLA₂ in mammalian cells has been shown to be involved in both arachidonic acid release and phospholipid remodeling, which appears to be dependent on the cell type and stimulus. For example, in P388D₁ cells, inhibition of iPLA₂ with BEL does not block arachidonic acid release in response to PAF stimulation (2) but does block arachidonic acid release in response to zymosan (1), indicating that iPLA₂ has a different role in the same cell in response to different stimulation. Similarly, zymosan stimulation of mouse peritoneal macrophages stimulates cPLA₂, leading to arachidonic acid release (32), whereas zymosan stimulation of arachidonic acid release in RAW 264.7 cells is sensitive to BEL (which does not inhibit cPLA₂, 16), illustrating that the same stimulus in different cells may activate different PLA₂ isoforms, each of which result in increased arachidonic acid release. The role of iPLA₂ in arachidonic acid release in recent studies suggests that this isoform is responsible for hydrolysis of membrane phospholipids in several different tissues (15, 25, 28, 30, 33, 36). In this study, we have measured PLA₂ activity in the absence of Ca²⁺, which is blocked by BEL pretreatment. BEL also completely blocks membrane phospholipid turnover and the production of metabolites in thrombin-stimulated myocytes, suggesting that iPLA₂ is the principal isoform responsible for thrombin-stimulated phospholipid hydrolysis. A major role for cPLA₂ was not found in ventricular myocytes, because pretreatment with AACOCF₃ or MAFP (both of which inhibit cPLA₂) had little effect on either basal or thrombin-stimulated PLA₂ activity.

In conclusion, we propose that thrombin stimulation of isolated adult rabbit ventricular myocytes results in activation of a membrane-associated iPLA₂ that selectively hydrolyzes plasmalogen phospholipids. The resultant products of this hydrolysis, namely lysoplasmenylcholine, arachidonic acid, and eicosanoids, may contribute directly to the production of electrophysiological abnormalities and arrhythmogenesis in the ischemic myocardium.