The role of calcium-independent phospholipase A2 in cardiolipin remodeling in the spontaneously hypertensive heart failure rat heart.

Cardiolipin (CL) is an essential phospholipid component of the inner mitochondrial membrane. In the mammalian heart, the functional form of CL is tetralinoleoyl CL [(18:2)4CL]. A decrease in (18:2)4CL content, which is believed to negatively impact mitochondrial energetics, occurs in heart failure (HF) and other mitochondrial diseases. Presumably, (18:2)4CL is generated by remodeling nascent CL in a series of deacylation-reacylation cycles; however, our overall understanding of CL remodeling is not yet complete. Herein, we present a novel cell culture method for investigating CL remodeling in myocytes isolated from Spontaneously Hypertensive HF rat hearts. Further, we use this method to examine the role of calcium-independent phospholipase A2 (iPLA2) in CL remodeling in both HF and nonHF cardiomyocytes. Our results show that 18:2 incorporation into (18:2)4CL is: a) performed singly with respect to each fatty acyl moiety, b) attenuated in HF relative to nonHF, and c) partially sensitive to iPLA2 inhibition by bromoenol lactone. These results suggest that CL remodeling occurs in a step-wise manner, that compromised 18:2 incorporation contributes to a reduction in (18:2)4CL in the failing rat heart, and that mitochondrial iPLA2 plays a role in the remodeling of CL's acyl composition.

ing in the Spontaneously Hypertensive HF (SHHF) rat as well as the potential remodeling role of iPLA 2 in this model of cardiac stress. We report that the incorporation of 18:2 into (18:2) 4 CL in SHHF cardiomyocytes: a ) occurs singly over time, b ) is attenuated with the development of HF, and c ) is partially sensitive to inhibition of iPLA 2 .

Materials
All materials used in this study were purchased from the Sigma Chemical Co. with the following exceptions: Type-2 collagenase was purchased from Worthington, racemic bromoenol lactone (BEL) and the iPLA 2 ␥ -specifi c enantiomer R-BEL were purchased from Cayman Chemical Co., and carbon stable isotope linoleic acid ( 13 C 18 -18:2, >98% isotope enrichment, abbreviated in the text as 13 C-18:2) was purchased from Spectra Stable Isotopes (Cambridge Isotope Laboratories).

Animals
The SHHF rat is a model of human HF that is genetically predisposed to death from hypertrophic followed by dilated cardiomyopathy, the etiology of which has been described by McCune et al. ( 37 ). Briefl y, female SHHF rats become hypertensive by approximately 3 months of age and this development progresses to overt hypertension by 5 months. Secondary to this hypertension, myocardial hypertrophy begins around 17 months of age and progresses to dilated cardiomyopathy between 23-25 months ( 38,39 ). Female SHHF rats (colony kept at the University of Colorado by S.A.M.) were designated as nonHF or HF based on age (2-3 months and 21-23 months, respectively), left ventricular (LV) function as assessed by echocardiography, and the absence or presence, respectively, of at least one of the following symptoms: labored breathing, piloerection, or orthopnea. Aged-matched (3 months and у 22 months) female Fisher Brown Norway (Fischer 344 x Brown Norway F1, FBN) rats (Harlan) were used in select experiments as an aging control. The FBN rat was chosen as an aging control, rather than the Fischer 344, Wistar, or Sprague Dawley rat, because the documented incidence of cardiovascular dysfunction and disease is milder and of later onset in FBN rats relative to the other strains (40)(41)(42). All animals were housed in groups of 2-3 on a 12:12 h light:dark cycle with ad libitum access to chow and water. The n values for each experiment are located within fi gure legends. All animal treatment was conducted in conformity with the Public Health Service Policy on Humane Use and Care of Laboratory Animals and in accordance with guidelines set forth on animal care at the University of Colorado, Boulder.

Echocardiographic analysis
Transthoracic echocardiography was performed on all rats 2-5 days prior to euthanasia under inhaled isofl urane anesthesia (5% initial, 2% maintenance) using a 12 MHz pediatric transducer connected to a Hewlett Packard Sonos 5500 Ultrasound system. Short axis M-mode echocardiograms on the LV were obtained for measurement of LV internal diameters at diastole (LVIDd) and systole (LVIDs), fractional shortening (FS), ejection fraction (EF), anterior wall thickness in diastole (AWTd), and posterior wall thickness in diastole (PWTd) as previously described ( 43 ).

Cardiomyocyte isolation
Cardiomyocytes were isolated from whole hearts with modifications to methods previously described ( 44,45 ). Animals re-acid (18:2) ( 2,6,16 ). Tetralinoleoyl CL [(18:2) 4 CL] accounts for 75-80% of total CL content in both rat and human cardiac mitochondria ( 2,16 ). The notion that 18:2 is essential for CL's function in the mammalian heart is based on three independent observations. First, the vast majority of CL is remodeled from its de novo form to (18:2) 4 CL subsequent to its biosynthesis. Second, the high prevalence of (18:2) 4 CL over other CL species is unlikely if one considers the remodeling process to be random with respect to acyl selection, which suggests that the loading of CL with 18:2 is purposeful. Lastly and most importantly, a loss of (18:2) 4 CL, along with an increase in 18:2-defi cient CL species, occurs in a number of cardiac disease states ( 6,18 ). The disease most directly associated with a loss of (18:2) 4 CL is Barth syndrome, caused by an X-linked mutation in the tafazzin gene (19)(20)(21)(22). Levels of 18:2 in cardiac CL also decline in congestive heart failure (HF) ( 16 ), ischemic HF ( 23,24 ), and diabetes ( 25 ). Because (18:2) 4 CL seems to be important for myocardial energy homeostasis, a complete understanding of the CL remodeling process is essential in designing future treatments for patients with HF and other mitochondrial diseases.
The formation of (18:2) 4 CL is dependent on the coupling of CL biosynthesis and remodeling. The biosynthesis of CL occurs within the IMM (26)(27)(28)(29), where nascent CL is formed from the condensation of phosphatidylglycerol (PG) and cytidinediphosphate-diacylglycerol (CDP-DAG) in a reaction catalyzed by CL synthase (CLS) (for review, see 30,31 ). Neither the acyl composition of PG and CDP-DAG nor the acyl specifi city of CLS results in an enrichment of CL with 18:2 de novo; thus, nascent CL must be converted to (18:2) 4 CL through an acyl remodeling cycle. Presumably, the remodeling of CL occurs through a series of deacylation-reacylation reactions, though the details of CL remodeling in vivo remain in question ( 18,30 ). To date, three enzymes have been identifi ed that are capable of adding 18:2 to a monolysoCL (MLCL): tafazzin ( 32 ), MLCL-acyltransferase (MLCL-AT, 33 ), and acylCoAlysocardiolipin acyltransferase (ALCAT-1, 34 ). None of these enzymes, however, possess phospholipase activity. In fact, very little research has examined the role of endogenous phospholipases in CL remodeling.
To the best of our knowledge, there are only two reports examining the role of mitochondrial phospholipases in CL remodeling. Mancuso et al. ( 35 ) created a murine model defi cient in the functional form of calciumindependent phospholipase A 2 (iPLA 2 ) ␥ (iPLA 2 ␥ , also PLA 2 GVIB), and a decrease in (18:2) 4 CL in these animals occurred concomitantly with symptoms of myocardial energetic disequilibrium. More recently, Malhotra et al. ( 36 ) examined the role of iPLA 2 ␤ (also PLA 2 GVIA), reporting that iPLA 2 ␤ is not necessary for CL remodeling, but does increase the MLCL/CL ratio in tafazzin-defi cient Drosophila melanogaster . Because these two reports are the fi rst of their kind, much is still unknown about iPLA 2 in CL remodeling. As such, the purpose of this study was to fi rst develop a method to examine CL remodeling at the level of the isolated rat cardiomyocyte, and thereafter, use this method to investigate potential alterations in CL remodel-charge ratios ( m/z ), and text abbreviations for all phospholipids presented can be found in Table 1 .

PG
To determine whether our method monitors the incorporation of 13 C-18:2 into (18:2) 4 CL at the level of CL remodeling or CL biosynthesis, we examined 13 C-18:2 content in PG following up to 72 h of incubation with 13 C-18:2 or 24 h of incubation with 13 C-18:2 plus BEL or R-BEL. For these experiments, we quantifi ed fi ve different species of PG (listed in Table 1 ) and expressed each as a fraction of their sum, which accounts for the vast majority of PG detected in the mass spectra.

Statistical analysis
For echocardiography data, data corresponding to control CL composition, and data involving the incorporation of 13 C-18:2 into PG, a multivariate ANOVA was used to test for an omnibus F-ratio. Rates of CL remodeling in the presence of BEL or R-BEL were evaluated with a two-factor repeated measures ANOVA. A two-factor ANOVA was used to examine differences in remodeling between young and aged SHHF and FBN cardiomyocytes. In the event of a statistically signifi cant F-ratio, post hoc multiple comparisons were made using Tukey's Honestly Signifi cant Difference or simple comparisons. Where necessary, absolute p -values were adjusted with a Bonferroni correction. In all cases, ␣ = 0.05 was set as the marker for statistical signifi cance.

Cardiac function and cardiomyocyte viability
Cardiac function. Rats were subjected to echocardiography to assess LV function preceding sacrifi ce. SHHF echocardiography data clearly demonstrate signifi cant LV hypertrophy and systolic dysfunction in HF compared with nonHF animals ( Table 2 ), consistent with the late stages of hypertensive heart disease and early HF in this animal model ( 47 ). In contrast, aged FBNs exhibited neither LV thickening nor functional defi cits when compared with young FBNs, and the LV morphological and functional ceived 2,000 units/kg body mass heparin and after 12 min were deeply anesthetized with 35 mg/kg sodium pentabarbitol, both through intraperitoneal injection. Hearts were rapidly excised, immersed in ice-cold saline, and cannulated by the aorta on a modifi ed Langendorff perfusion apparatus. Hearts were perfused in a retrograde manner for 5 min with "buffer B" (described in refs. 44,45 ), after which the perfusate was changed to a digestion buffer identical to the fi rst, except containing 1.30 or 1.50 mg/ml type-2 collagenase (for nonHF and HF hearts, respectively), 1.30 mg/ml hyaluronidase, 100 g/ml dialyzed albumin, and 55 M CaCl 2 . When suffi ciently digested (as determined by increases in coronary fl ow and softness to touch), hearts were cut from the cannula and the right ventricular free wall was removed. Remaining LV and septal tissue was cut into smaller pieces and teased apart with blunted glass pipette tips. Cells were washed once with a buffer identical to buffer B, except containing 100 M CaCl 2 , and twice with medium 199 (37°C, pH 7.4, with 100 units/ml penicillin and 5 g/ml gentamycin). After the fi nal wash, cells were seeded on laminin-coated glass microscope coverslips in Springhorn medium (medium 199 with the addition of 2 mg/ml BSA, 100 nM insulin and (in mM): 2 carnitine, 5 creatine, 5 taurine, 1.3 glutamine, 2.5 sodium pyruvate, 10 2,3butane dione monoxime; pH 7.70 before equilibration with 5% CO 2 ) and incubated at 37°C for 2-3 h.

Cardiomyocyte treatment
For each experimental group, three laminin-coated glass coverslips plated with cardiomyocytes were bathed in Springhorn medium in 100 × 15 mm sterile Petri dishes (fi nal volume 12 ml). The fi rst group served as a control and was incubated with 0.17 mM fatty acid-free BSA and 0.1% DMSO vehicle. The second group of cells was incubated with BSA-bound 13 C-18:2 such that the fi nal concentrations were 1 mM 13 C-18:2 and 0.17 mM BSA (a 6:1 18:2:BSA ratio), with 0.1% DMSO vehicle. The third group of cells was treated in a manner similar to the second group, but was incubated for 30 min with 10 M BEL in 0.1% DMSO prior to the addition of BSA-bound 13 C-18:2 ( 13 C-18:2 + BEL). The fi nal group was treated exactly the same as the third group, except 5 M of the iPLA 2 ␥ -specifi c BEL enantiomer, R-BEL, was used instead of the racemic BEL mixture ( 13 C-18:2 + R-BEL). Data was not shown for a fi fth, "BEL control" group (10 M BEL in 0.1% DMSO, 0.17 mM BSA), because this treatment did not result in any measurable effects. In the event that incubations lasted longer than 24 h, Springhorn medium and all necessary chemicals were replaced every 24 h. Myocytes were photographed preceding and subsequent to each treatment period using a Sony Cybershot DSC-S75 digital camera with a VAD-S70 adaptor ring under an inverted light microscope at 100× magnifi cation.

Cardiomyocyte harvest and lipid extraction
Following treatment, myocytes were scraped from coverslips and centrifuged at 1600 g for 5 min at room temperature. Pelleted myocytes were resuspended in HPLC-grade methanol and stored at Ϫ 20°C until lipid extraction. Lipids were extracted according to Bligh and Dyer ( 46 ) and subject to ESI-MS for phospholipid content analysis.

Phospholipid analysis
Singly-ionized CL and PG species were quantifi ed by ESI-MS as described by Sparagna et al. ( 46 ). Tetramyristoyl CL [(14:0) 4 CL] was included as an internal standard to verify the quality of the spectra. Differences in cell yield between groups were controlled for by expressing each analyte as a fraction of its total respective phospholipid content. The specifi c acyl compositions, mass to present under control conditions (Fig. 3A) appear after 24 h and 48 h incubations with 13 C-18:2 (Figs 3B and 3C, respectively), and this appearance is partially prevented by preincubation with BEL (Fig. 3D). Figure 4 shows the quantitative results from these spectra for nonHF (Fig.  4A) and HF (Fig. 4B) cardiomyocytes. In addition to 13 Clabeled (18:2) 4 CL species, Fig. 4 A and 4B also display the levels of endogenous (18:2) 4 CL and the sum of all labeled and nonlabeled (18:2) 4 CL species throughout the incubation period. With added 13 C-18:2, nonHF SHHF myocytes were able to raise and sustain total (18:2) 4 CL levels to 80.4 ± 1.5% of total CL, whereas levels in myocytes isolated from HF animals peaked at 64.8 ± 5.1%. Furthermore, the rate of 13 C-18:2 incorporation into (18:2) 4 CL is attenuated in HF myocytes, as all 13 C-labeled (18:2) 4 CL species peak at higher values in nonHF SHHF myocytes as compared with HF myocytes.
Cardiomyocyte viability. Isolated myocytes were photographed preceding and subsequent to each treatment period. No large differences in myocyte viability were witnessed between the pre-and post-treatment time points (representative micrographs shown in Fig. 1 ).

Incorporation of 13 C-18:2 into (18:2) 4 CL
In a time-course experiment, we incubated SHHF cardiomyocytes in 13 C-18:2 for up to 72 h, monitoring the incorporation of 13 C-18:2 into (18:2) 4 CL via the formation of 13 C-labeled CL species. The results from these experiments are presented in Figs. 3 and 4 . Figure 3 shows representative ESI mass spectra from nonHF myocytes under varying conditions. As shown, 13 C-labeled (18:2) 4 CL peaks not   elevated in nonHF, but not HF, myocytes after 72 h. In both nonHF and HF myocytes, the addition of neither BEL nor R-BEL to the 13 C-18:2-containing medium resulted in any signifi cant changes (all pair-wise P > 0.05).
Incorporation of 13

C-18:2 into (18:2) 4 CL is partially iPLA 2 -dependent
We examined 13 C-18:2 incorporation into singly-labeled CL [( 13 C-18:2)(18:2) 3 CL] in the presence of the iPLA 2 inhibitor, BEL. Preincubation with 10 M BEL attenuated 13 C-18:2 incorporation in myocytes isolated from both nonHF and HF rat hearts for up to 10 h of incubation ( Fig.  6 A , B). Interestingly, the percent of total 13 C-18:2 incorporation inhibited by BEL after 10 h was signifi cantly less in nonHF versus HF (61.2 ± 2.9% and 79.6 ± 2.3%, respectively, P < 0.01). After 24 h of incubation, 13 C-18:2 incorporation was still sensitive to BEL in nonHF myocytes; however, BEL sensitivity was diminished in myocytes isolated from HF SHHF rats. To further quantify the differential effects of BEL treatment in nonHF and HF, we measured the initial rates of singly-labeled CL formation after 10 h and 24 h of incubation in 13 C-18:2 with and without BEL. In both nonHF and HF, the rate of singly-labeled CL formation was signifi cantly reduced by BEL throughout the fi rst 10 h of incubation; however, this incorporation was sensitive to BEL after 24 h in only nonHF myocytes ( Fig. 6 table inserts). Finally, because data from Mancuso et al. ( 35 ) suggested a role of iPLA 2 ␥ in CL remodeling, we examined 13 C-18:2 incorporation in the presence of the iPLA 2 ␥ -specifi c enantiomer, R-BEL. R-BEL had very similar effects on CL remodeling when compared with BEL in nonHF myocytes, and was capable of signifi cantly preventing 13 C-18:2 incorporation into singly-labeled CL following up to 24 h of incubation with respect to both total incorporation over 24 h and the initial rate of incorporation ( Fig.  7 ). Racemic BEL and R-BEL had the same effects on the initial rates of incorporation and resulted in similar rates of 18:2 incorporation after both 10 h and 24 h of incubation ( P > 0.05). Neither 10 M BEL nor 5 M R-BEL had an effect on cardiomyocyte viability during the 24 h treatment period (micrographs not shown).

DISCUSSION
The maintenance of cardiac (18:2) 4 CL levels appears to be extremely important in mitochondrial energetics; however, the exact mechanism by which CL is remodeled to contain 18:2 remains to be determined. We conducted this study, fi rst, to develop a method for monitoring CL remodeling in the isolated rat cardiomyocyte, and next, to use this method to examine both changes in CL remodeling in the context of HF and the role of iPLA 2 in CL remodeling. We presented data that show CL is remodeled singly with respect to its fatty-acyl moieties; the rate of 18:2 incorporation into CL is depressed in HF; and iPLA 2 is partly involved in the incorporation of 13 C-18:2 into (18:2) 4 CL in SHHF cardiomyocytes.
In this report, we presented a new method for studying the remodeling of CL at the level of the individual cardio-whereas aged FBN and nonHF SHHF myocytes had intermediate values (21.2% and 22.0%, respectively) compared with HF SHHF myocytes (15.9%). Notably, there was a trend for decreased ( 13 C-18:2) 2 (18:2) 2 CL content in aged FBN animals when compared with young FBN animals ( P = 0.07).

Incorporation of 13 C-18:2 into PG
To verify that the measured incorporation of 13 C-18:2 into (18:2) 4 CL was due to CL remodeling and not an artifact of 13 C-18:2 content in the biosynthetic pathway, we examined 13 C-18:2 incorporation into PG in nonHF and HF SHHF myocytes. Levels of the predominant species of endogenous PG, (16:0)(18:1)PG ( 49 ), progressively declined throughout 72 h of incubation with 13 C-18:2 in nonHF ( P < 0.05; Fig. 5 A ), and trended to decline after 48 h and 72 h in HF ( Fig. 5 B). Doubly-labeled PG [( 13 C-18:2) 2 PG] increased in cells from both nonHF and HF animals over time; furthermore, ( 13 C-18:2)(18:2)PG, ( 13 C-18:2)(16:0)PG, and ( 13 C-18:2)(18:1)PG levels were  Presumably, the remodeling of nascent CL to (18:2) 4 CL occurs through a series of four discreet deacylationreacylation cycles, such that MLCL is the only necessary intermediate. In fact, dilysoCL is not readily acylated to CL in isolated mitochondria ( 52 ). As far as we know, however, there is no direct evidence for a step-wise incorporation of 18:2 into CL. The results from our time-course experiment are the fi rst to show that CL is remodeled in this step-wise manner. Rather than the sporadic appearance of 13 Clabeled CL with one, two, three, or four 13 C-18:2 moieties, the incorporation of 13 C-18:2 into (18:2) 4 CL occurs singly over time.
There exists a large descriptive precedent documenting abnormal CL composition in the context of disease ( 6,16,20,53 ); however, aside from Barth syndrome, no research has resulted in a mechanism for this decline. In this report, we have shown that 18:2 incorporation into CL is attenuated in myocytes isolated from failing rat hearts. Total (18:2) 4 CL levels peaked at approximately 65% of total CL in HF myocytes, which was much lower than the corresponding value (80%) in nonfailing myocytes. These observations show that the failing myocardium has an attenuated ability to traffi c and/or incorporate 18:2 into CL. Interestingly, the incorporation of 18:2 into PG was also lower in HF myocytes, indicating that acyl-chain remodeling abnormalities are not limited to CL in the failing rat heart.
In a recent two-part publication, Schlame and colleagues ( 54,55 ) provided evidence that the acyl composition of CL depends more on the composition of the local lipid environment than the acyl specifi city of CL remodeling myocyte. Although we argue that our method monitors the incorporation of 13 C-18:2 at the level of CL remodeling rather than CL biosynthesis, we reported 13 C-containing PG species in our cultures following up to 72 h exposure to 13 C-18:2. Because it is CL's precursor, an isotopic enrichment of PG suggests that our method monitors not CL remodeling, but the formation of isotopic CL from isotopic PG. We believe this data is misleading and that our method does monitor CL remodeling for the following reasons: fi rst, the only species of PG containing 13 C-18:2 after 24 h was doubly-labeled PG, and this phospholipid could not condense with CDP-DAG to form the major isotopic CL species at 24 h, ( 13 C-18:2)(18:2) 3 CL. We believe the increase of doubly-labeled PG, which occurs concomitantly with a loss of endogenous (16:0)(18:1)PG, represents a remodeling of PG by mass-action. The accumulation of ( 13 C-18:2) 2 PG suggests that it may not be a substrate for CLS. Although human CLS expressed in Saccharomyces cerevisiae can use (18:2) 2 PG to synthesize CL ( 50 ), rat cardiomyocyte CLS exhibits different substrate specifi city than human CLS ( 51 ). Thus, it is possible that rat CLS cannot use ( 13 C-18:2) 2 PG in the condensation reaction. Second, the 13 C-labeled PG species that could give rise to singly-or triply-labeled CL did not increase until 72 h, whereas these CL species were detected before 72 h. Finally, the incorporation of 13 C-18:2 into CL was sensitive to inhibition of iPLA 2 , whereas the incorporation of 13 C-18:2 into PG was not. Therefore, the incorporation of 13 C-18:2 into (18:2) 4 CL in this method appears to occur largely through remodeling CL and not simply by increasing 13 C-18:2 content in CL de novo. CL with 18:2. In our previous publication, we also reported a 5-fold increase in MLCL-AT activity in HF ( 56 ), which may represent a compensatory response to the reduction in tafazzin content. Regardless of this, HF myocytes were still unable to properly remodel CL; therefore, we postulate that in the absence of tafazzin, the bioavailability of 18:2-CoA for MLCL-AT acyl transfer may be regulated by an additional, currently unknown mechanism. In addition to alterations in CL acyl composition during disease states, data also exist demonstrating declines in 18:2 content in CL with age ( 57 ). To investigate a potential aging effect on CL remodeling, we measured 18:2 incorporation into (18:2) 4 CL in a nonpathological model of aging, the FBN rat. Both young and aged FBN myocytes incorporated 18:2 into CL more readily than HF SHHF myocytes. Although these data suggest that attenuated CL remodeling is associated only with HF, we also noted a trend for lower 18:2 incorporation with age in FBN myocytes. Overall, our data suggest that both aging and the development of HF may impact CL remodeling, although the relative reduction due to aging alone is only half that due to the development of HF (14.7% and 27.7% enzymes (e.g., tafazzin). We have previously reported a 97% reduction in tafazzin transcript in the failing SHHF rat heart ( 56 ) and have shown here that, even in the presence of ample substrate, HF cardiomyocytes are unable to raise (18:2) 4 CL levels to those of nonHF myocytes or values reported in healthy rat and human cardiac mitochondria ( 6,16 ). Thus, it appears that the decline in tafazzin transcript may explain the inability of HF myocytes to load  rat hearts, which is consistent with data put forth by Mancuso et al. ( 35 ). Interestingly, the percent of ( 13 C-18:2)(18:2) 3 CL formation that is inhibited by BEL after 10 h of incubation is signifi cantly greater in HF versus nonHF myocytes. These results indicate a potential increase in the quantity or activity of iPLA 2 in the failing myocardium. Theoretically, increased phospholipase activity would be cytoprotective by preventing an accumulation of lipid peroxidation end products; however, prior research on the role of iPLA 2 in models of cellular stress have yielded confl icting results. Williams and Gottlieb ( 58 ) reported that inhibition of iPLA 2 during ischemia reduced mitochondrial phospholipid loss and was cardioprotective, whereas Seleznev et al. ( 59 ) reported that the presence of iPLA 2 was cytoprotective during apoptotic induction by staurosporine. In our model of HF, increased iPLA 2 activity may be benefi cial to the mitochondrial membrane in the absence of other abnormalities; however, when coupled with a lack of lysophospholipid reacylation (e.g., reduction in tafazzin) increased iPLA 2 activity may result in the net degradation of CL or other mitochondrial phospholipids. Pretreatment of cardiomyocytes with BEL was more potent in HF myocytes for the fi rst 10 h, but after 24 h of incubation, BEL-sensitive CL remodeling disappears in HF, but not in nonHF. Polyunsaturated fatty acids and their derivatives are known ligands for nuclear receptors ( 60-62 ) and the apparent loss of iPLA 2 -dependent CL remodeling may be caused by upregulation of peroxisome proliferatoractivated receptor binding and alterations in cellular levels of lipid-metabolizing enzymes.
Because Mancuso et al.'s ablation of the ␥ -isoform of iPLA 2 resulted in CL abnormalities, we also treated myocytes with R-BEL in the presence of 13 C-18:2. Specifi c inhibition of iPLA 2 ␥ resulted in similar effects on 18:2 incorporation in nonHF myocytes with respect to total isotope incorporation and the initial rate of 18:2 incorporation, consistent with the notion that iPLA 2 ␥ is the calcium-independent phospholipase isozyme involved in remodeling CL. Although iPLA 2 ␥ seems to be involved in CL remodeling, its specifi c role in the process is still unknown. Namely, does iPLA 2 ␥ directly hydrolyze CL to MLCL for acylation, or is it involved in the remodeling of other glycerophospholipids that act as 18:2 donors for CL remodeling? We have shown that iPLA 2 ␥ does not remodel PG; however, its potential role in remodeling 18:2 donors such as phosphatidylcholine or phosphatidylethanolamine ( 32 ) remains to be determined.
The use of BEL as an inhibitor of iPLA 2 is somewhat controversial. Although BEL's selective nature for iPLA 2 over cytosolic PLA 2 or secretory PLA 2 is well accepted (63)(64)(65), previous investigators have warned against the use of BEL as a specifi c inhibitor of iPLA 2 . These reports suggest that BEL inhibits the magnesium-dependent, cytosolic isoform of a lipid phosphate phosphatase, PAP-1, thereby perturbing cellular lipid homeostasis (by inhibiting the formation of DAG from phosphatidate) and promoting apoptosis in prolonged cell culture ( 66,67 ). Indeed, we have unpublished observations that concentrations of BEL at or exceeding 30 M are toxic to SHHF cardiomyocytes reductions in aged FBN and HF SHHF myocytes, respectively).
The experiments in which we inhibited iPLA 2 in the presence of 13 C-18:2 yielded a number of novel results. First, our data suggest CL remodeling is partially iPLA 2dependent for up to 10 h in both nonHF and HF SHHF  after 10 h, which is consistent with observations published by these researchers. However, investigation by Gross et al. ( 68 ) showed that BEL neither diminishes whole cell lipid phosphate phosphatase activity at concentrations up to 100 M, nor the activities of either PAP-1 or its membranebound relative, PAP-2, at concentrations up to 200 M in purifi ed subcellular fractions. In our study, we used a concentration of BEL (10 M) below that which was previously reported to promote apoptosis or attenuate PAP-1 activity (25 M) ( 66,67 ). The role of PAP-1 as a phosphomonoesterase is unlikely to infl uence the remodeling of CL's acyl chains; furthermore, the concentration of BEL used herein was not toxic to myocytes after 24 h of incubation. Though we cannot entirely discount the possibility that BEL is affecting the activity of PAP-1 in our cell culture, we do not believe it seriously confounds the interpretation of our results.
In closing, we have used a novel cell culture method to generate data suggesting a necessary but partial role of iPLA 2 ␥ in CL remodeling in the SHHF rat heart. Further, 18:2 incorporation into (18:2) 4 CL is decreased largely in HF and may decrease slightly during nonpathological aging, CL is remodeled singly with respect to its acyl moieties, and iPLA 2 -dependent CL remodeling accounts for a greater percentage of total CL remodeling in HF versus nonHF after 10 h of observation. To the best of our knowledge, the current body of literature suggests that nascent CL may be remodeled to (18:2) 4 CL by a number of interacting deacylation-reacylation enzyme pairs, and that they all proceed through the intermediate MLCL. Future research examining the specifi c role of iPLA 2 ␥ and the path of 18:2 through the cell are necessary if we are to fully understand the remodeling of this unique phospholipid.