Acylglycerol Recycling from Triacylglycerol to Phospholipid, Not Lipase Activity, Is Defective in Neutral Lipid Storage Disease Fibroblasts*

Neutral lipid storage disease (NLSD) is an autosomal recessive disorder in which excess triacylglycerol (TG) accumulates in most cells. Although it has been hypoth- esized that the TG accumulation is caused by a functional defect in cytosolic lipase activity, we were able to expose TG hydrolysis in NLSD cells by using triacsin C, an inhibitor of acyl-CoA synthetase that blocks the re- incorporation of hydrolyzed fatty acids into glycerolipids. Our data suggest that TG lipolysis in NLSD cells is masked by rapid TG resynthesis, occurring because released acylglycerols cannot be used for phospholipid synthesis. In uptake studies, triacsin C blocked the incorporation of [ 3 H]glycerol into glycerolipids, incorporation of [ 14 C]oleate into TG, but not incorporation of [ 14 C]oleate into phospholipid. Thus, the drug inhibited both de novo synthesis of glycerolipids via the glycerol- 3-phosphate pathway and the synthesis of TG from diacylglycerol. The drug did not appear to block reacyla- tion of lysophospholipids. Triacsin C caused a loss of about 60% of the TG mass from both NLSD and oleate- loaded control cells. Rates of TG lipolysis were similar in NLSD cells and oleate-loaded control cells labeled with [6-(7-nitro-2,1,3-benzoxadiazol-4-yl)-amino]hexanoicacidorlabeledwith[ 14 C]oleate or [

Neutral lipid storage disease (NLSD) is an autosomal recessive disorder in which excess triacylglycerol (TG) accumulates in most cells. Although it has been hypothesized that the TG accumulation is caused by a functional defect in cytosolic lipase activity, we were able to expose TG hydrolysis in NLSD cells by using triacsin C, an inhibitor of acyl-CoA synthetase that blocks the reincorporation of hydrolyzed fatty acids into glycerolipids. Our data suggest that TG lipolysis in NLSD cells is masked by rapid TG resynthesis, occurring because released acylglycerols cannot be used for phospholipid synthesis. In uptake studies, triacsin C blocked the incorporation of [ 3

H]glycerol into glycerolipids, incorporation of [ 14 C]oleate into TG, but not incorporation of [ 14 C]oleate into phospholipid. Thus, the drug inhibited both de novo synthesis of glycerolipids via the glycerol-3-phosphate pathway and the synthesis of TG from diacylglycerol. The drug did not appear to block reacylation of lysophospholipids. Triacsin C caused a loss of about 60% of the TG mass from both NLSD and oleateloaded control cells. Rates of TG lipolysis were similar in NLSD cells and oleate-loaded control cells labeled with [6-(7-nitro-2,1,3-benzoxadiazol-4-yl)-amino]hexanoic acid or labeled with [ 14 C]oleate or [ 3 H]glycerol and chased in the presence of triacsin C. During a 96-h chase, [ 14 C]oleate reincorporation into the different phospholipid species increased only in control cells.
Similar results were observed when NLSD, and control cells were chased after labeling with [ 3 H]glycerol. These data strongly suggest that normal human fibroblasts mobilize stored TG for phospholipid synthesis and that recycling to PC occurs via a TG-derived mono-or diacylglycerol intermediate. Normal recycling to phosphatidylethanolamine may primarily involve TG-derived acyl groups rather than an acylglycerol precursor. NLSD cells appear to have a block in this recycling pathway with the result that both hydrolyzed fatty acids and the acylglycerol backbone are re-esterified to form TG. Because the NLSD phenotype includes ichthyosis, fatty liver, myopathy, cardiomyopathy, and mental retardation, the recycling pathway appears to be critical for the normal function of skin, liver, muscle, heart, and the central nervous system.
Most cells are able to synthesize triacylglycerol (TG) 1 and to store it in cytosolic droplets, but the function and fate of stored TG in non-adipose tissue has not been extensively investigated. It has been suggested that diacylglycerol (DAG) derived from stored TG may be recycled as a precursor for phospholipid synthesis (1) or, in hepatocytes, for very low density lipoprotein TG synthesis (2). The identification of a human genetic disorder in which cellular TG accumulation is associated with a variety of clinical problems suggests that the metabolism of cytosolic TG is critical for normal cell function.
Neutral lipid storage disease (NLSD) is an autosomal recessive disorder in which most cells accumulate intracellular droplets of TG. The clinical phenotype variably includes ichthyosis, fatty liver, mental retardation, myopathy, ataxia, neurosensory hearing loss, cataracts, and cardiomyopathy (3,4). In NLSD, the uptake, transport, and ␤-oxidation of fatty acids (5), lipase and carboxyesterase activities (6 -8), and several enzymes of glycerolipid synthesis (8) are all normal. Nevertheless, the TG content of NLSD lymphocytes (6), macrophages (9), and fibroblasts (5) in culture is 5-, 2-, and 20-fold higher, respectively, than in normal cells. Radiolabeled TG does not deplete even when cells are placed in lipid-deficient medium (4,5,10). Unlike the lipase/cholesterol esterase deficiency of Wolman's disease, the defect in NLSD is extra-lysosomal (10,11), and lipolysis of short chain TG may be normal (12). Salvayre's group has hypothesized that the NLSD cells might have a functional block in cytosolic lipase activity such that even though the lipase is normal in in vitro studies, it does not function in situ (11,12).
In order to understand the normal role of stored TG, we examined recycling of TG's fatty acid and glycerol moieties to phospholipid in NLSD fibroblasts and in normal fibroblasts that had been loaded with oleate to approximate the TG content of NLSD cells. Using the acyl-CoA synthetase inhibitor, triacsin C, we were able to expose the extent of TG hydrolysis by preventing the activation and reutilization of hydrolyzed fatty acid (13,14). Our studies show that NLSD cells are not defective in lipase function but, instead, appear to have a defect that involves abnormal recycling of TG-derived mono-or diacylglycerols to specific phospholipids.  Serdary. C6-NBD, C6-NBD-PA, C6-NBD-PC, and C6-NBD-PE were from Avanti Polar Lipids. Triacsin C  was from Biomol Research Lab, Inc. Tissue culture supplies and fetal  bovine serum were from Life Technologies, Inc. Cell Culture-Normal human skin fibroblasts were obtained from the American Type Tissue Culture Collection (cell line CCD). NLSD fibroblasts were obtained with parental informed consent from a child with lamellar ichthyosis, mental retardation, fatty liver with fibrosis and cholestasis, and vacuolated neutrophils. 2 Normal fibroblasts and NLSD fibroblasts were cultured at 37°C in a humidified atmosphere of 5% CO 2 in minimum essential medium with Earle's salts plus 1% non-essential amino acids (E-MEM) and 10% fetal bovine serum (FBS) that had been heat-inactivated for 60 min at 56°C. The medium was changed every 2-3 days. For each experiment total cellular DNA content was measured fluorometrically (15).

Materials-Silica
Cellular Triacylglycerol Mass Quantification-Normal and NLSD fibroblasts were grown to confluence in 100-mm dishes. In order to obtain a high cellular TG content, similar to that present in NLSD cells, some control cells were incubated for 24 h with 1.0 mM sodium oleate in 1% BSA. This treatment produced abundant cytoplasmic lipid droplets similar in appearance to those present in NLSD cells. Control and NLSD cells were washed three times with 3 ml of 0.1% BSA in PBS. Then 8 ml of fresh medium (E-MEM, 10% FBS, supplemented with 1% BSA) was added in the presence of either 5 M triacsin C in Me 2 SO (0.1%) or in Me 2 SO alone. After 96 h, cell monolayers were washed once with warm PBS, and the cells were trypsinized. DNA was measured in aliquots from each cellular suspension. The remaining cells were pelleted by centrifugation, resuspended in water, and probe-sonicated (three pulses of 15 s each). Cell lipids were extracted (16), and total TG content was determined using a commercial kit (GPO-Trinder, Sigma Co.).
Cell Radiolabeling with Oleic Acid-Fibroblasts were seeded in 60-mm dishes and grown to near confluence. The cultures were incubated with 0.5 Ci of [ 14 C]oleic acid in 3 ml of 10% FBS, E-MEM. Sodium oleate (0.1 or 1.0 mM) was included in the labeling medium to promote incorporation of label into TG. Sodium oleate was dissolved in H 2 O at 65°C and added to dry [ 14 C]oleate. Then 1% BSA (final concentration) in E-MEM was added. After 24 h, the medium was removed, and residual radiolabel was removed by washing the cells three times with 0.1% BSA in PBS at 37°C. Cells were then chased in 2 ml of 10% FBS, E-MEM plus 0.1% BSA at 37°C in the absence or the presence of 5 M triacsin C dissolved in Me 2 SO. The medium was not replaced during the chase. The concentration of Me 2 SO in the culture medium never exceeded 0.2% (v/v). Triacsin treatment did not affect cell viability. At the end of each incubation, the medium was removed, and the cells were washed once with 0.1% BSA in 37°C PBS. Cells were scraped in two additions of 1 ml of methanol and 0.5 ml of H 2 O. 1 ml of CHCl 3 was added, and total cell lipids were extracted (16) and concentrated using a SpeedVac concentrator (Savant, Hicksville, NY).
Cell Radiolabeling with Glycerol-Near confluent monolayers of normal and NLSD fibroblasts in 60-mm dishes were incubated at 37°C with 2 Ci/ml [ 3 H]glycerol (specific activity, 1.0 Ci/mol) in 10% FBS, E-MEM, supplemented with 0.1 or 1.0 mM sodium oleate in 1% BSA. After 20 h, the labeling medium was removed, and cells were washed three times with 0.1% BSA in warm PBS to remove any residual label. Control and NLSD fibroblasts were then chased for 24 -96 h in 2 ml E-MEM, 10% FBS, supplemented with 0.1% BSA in the presence or the absence of 5 M triacsin C in Me 2 SO (final Me 2 SO concentration was 0.1%, v/v). Medium was not replaced during the chase period. At the end of each incubation, cells were washed with 1 ml of 0.1% BSA in PBS, and lipids were extracted as described above.
Lipid Analysis-Aliquots of the lipid extracts were spotted on 0.25-mm silica gel G plates. Neutral lipids were resolved in heptane: isopropylether:glacial acetic acid (60:40:4, v/v/v) with authentic standards. Phospholipids were separated in a double one-directional solvent system. First, the plate was chromatographed in CHCl 3 :CH 3 OH: C6-NBD-Phosphatidic Acid Incorporation into Cellular Lipids-Normal and NLSD fibroblasts were grown to near confluence in 100-mm dishes. Control cells were first incubated with 1.0 mM sodium oleate in 1% BSA for 24 h in order to increase TG content. Then, normal and NLSD fibroblasts were incubated for another 24 h with 5 M C6-NBD-PA (in 0.1% Me 2 SO) and 1.0 mM sodium oleate in 1% BSA. At the end of this incubation, cells were washed three times with 0.1% BSA in warm PBS to eliminate residual NBD-PA. Then cells were chased in E-MEM, 10% FBS for 24 -48 h. At the end of each incubation, fibroblasts were washed once with 0.1% BSA in PBS, and lipids were extracted as described previously. C6-NBD-labeled lipids were separated in a double one-directional solvent system with authentic standards. The C6-NBD-DAG standard was produced by phospholipase C treatment of C6-NBD-PC. The C6-NBD-TG standard was synthesized from C6-NBD-DAG (18). After chromatography with CHCl 3 :CH 3 OH: 30% NH 4 OH (65:35:5, v/v) to 10 cm, plates were air dried for 30 min and then chromatographed to the top of the plate in CHCl 3 :CH 3 OH: CH 3 COOH (100:2.5:5, v/v).

Triacylglycerol Mass Is Mobilized Similarly from Control and NLSD Cells-Previous studies have shown that after TG
in normal fibroblasts is labeled with 14 C-labeled fatty acid, most of the radiolabeled TG disappears during a 24-h chase. In contrast, radiolabeled TG in NLSD cells is not depleted (4,5,10). Because lipase activities are normal in NLSD cells, it has been suggested that the lipase may require a missing cofactor, be mislocalized, or be otherwise unable to reach the droplet TG (11,12). However, because the amount of endogenous TG in normal fibroblasts is very low and that of NLSD cells is high, we suspected that even normal cells would give the appearance of a stable TG pool if radiolabeled TG were diluted into a large unlabeled TG pool similar to that present in NLSD cells.
Confluent NLSD cell contained 17-fold more TG/g DNA than did control fibroblasts (Table I). When control fibroblasts were incubated with 1.0 mM sodium oleate, their cellular TG content increased 15-fold. The oleate-loaded control cells depleted their TG content slowly; after a 96-h chase, the total TG content in control cells decreased by only 15%. NLSD cells maintain their high TG content under these conditions. In order to block the possible re-esterification of any hydrolyzed fatty acid, we incubated control and NLSD cells with triacsin C, a potent inhibitor of acyl-CoA synthetase (13,14). When cells were incubated for 96 h in the presence of triacsin C, the content of TG in both oleate-loaded control and NLSD fibroblasts each decreased by approximately 50% (Table I). Thus, when fatty acid re-esterification was prevented, the two cell lines showed no apparent difference in their ability to hydrolyze TG. Under these conditions, it appeared that 35 and 50% of any fatty acid hydrolyzed from TG was being re-esterified back to TG in oleate-loaded control and NLSD cells, respectively. Triacsin Exposes TG Hydrolysis in Control and NLSD Cells-Phospholipid synthesis and deacylation/reacylation cycles may be tightly controlled such that any excess fatty acid is channeled to TG synthesis (1). In order to examine the fate of hydrolyzed TG, we studied the turnover of [ 14 C]oleate-labeled glycerolipids in control cells whose TG content had been increased to the level normally present in NLSD cells by incubation with 1.0 mM [ 14 C]oleate. After 24 h, 90% of the incorporated label in control cells was present in the TG fraction. In NLSD fibroblasts incubated with only 0.1 mM [ 14 C]oleate, 75% of the incorporated labeled fatty acid was present as TG (Fig.  1). The amount of radiolabel incorporated into TG by control cells was three times higher than in NLSD cells, consistent with a greater than 15-fold increase in TG mass in the presence of 1.0 mM oleate (see Table I). In contrast, the amount of [ 14 C]oleic acid incorporated into phospholipids was similar in both cell lines. These data are consistent with the view that excess fatty acid is used for TG synthesis (1).
After a 48-h chase, [ 14 C]TG in the oleate-loaded control cells decreased by 60%, whereas no loss was observed in NLSD cells (Fig. 1, A and B). However, when triacsin C was added during the chase to block reutilization of hydrolyzed fatty acid, labeled TG decreased 50% in NLSD cells and 70% in control cells by 48 h. This experiment confirms that NLSD cells contain a functional TG lipase capable of hydrolyzing TG in lipid droplets but that lipase activity is masked by rapid resynthesis of TG. Without triacsin C, oleate-loaded control cells lose labeled TG slowly, suggesting that both cell lines reduce their TG content at similar rates and that previous reports showing rapid loss of labeled fatty acid from control cells was likely due to their low TG mass and the consequent difference in TG-specific activity within the control and NLSD cells (5, 6 -8).
Triacsin Exposes Fluorescent TG Loss in NLSD and Lipidloaded Control Cells-Because the triacsin C might alter TG hydrolysis by a mechanism unrelated to the ability of the drug to inhibit acyl-CoA synthetase, we tested the metabolism of TG synthesized from the fluorescent precursor, C6-NBD-PA. When incubated with cultured cells, C6-NBD-PA is cleaved to C6-NBD-diacylglycerol, which enters the cells and is metabolized to C6-NBD-labeled glycerolipids (19). When a fluorescent glycerolipid is hydrolyzed, the released C6-NBD is not reincorporated into glycerolipid (19). After a 24-h incubation in medium with 1.0 mM oleate, NLSD cells and oleate-loaded control cells incorporated similar amounts of the fluorescent label into TG and PC (Fig. 2). During a 48-h chase, both cell lines lost the fluorescent labels rapidly. PC was undetectable after 24 h, and the rate of decrease from TG was equally rapid in both cell lines, further supporting the conclusion that the NLSD cells are not deficient in their ability to hydrolyze TG.
Triacsin Alters the Turnover of [ 14 C]Oleate-labeled Glycerolipids-Both cell lines initially incorporated similar amounts of [ 14 C]oleate into phospholipid. During the chase, however, the phospholipid fraction in control cells increased 66% (Fig. 1A), whereas no change was observed in phospholipid labeling in NLSD cells (Fig. 1B). The addition of triacsin C did not alter the incorporation or metabolism of total [ 14 C]oleate-labeled phospholipids, but in NLSD cells, the acyl-CoA synthetase inhibitor caused the amount of labeled phospholipid to decrease 40% during the first 48 h. These results suggest that the control cells are less dependent on reacylation to maintain label in phospholipid.
In order to determine which phospholipid species were altered during the chase, phospholipids were separated by TLC and quantified. 73% of the label was incorporated into PC (Fig.  3). At 48 h in the absence of triacsin, each phospholipid species in control cells increased 1.4-(PC), 2-(PI-PS), 3-(PE), or 5.7-fold (sphingomyelin), whereas in NLSD cells, labeling of these phospholipids either decreased (PC and PI-PS) or increased minimally (PE and sphingomyelin) (Fig. 3). In contrast, when triacsin was present for 48 h, control cells showed a 20% increase in [ 14 C]oleate incorporation into PC at 48 h, but NLSD cells showed a 20% decrease (Fig. 3A). Increased label in PC could result if triacsin increased the amount of cellular fatty acid or diacylglycerol, thereby activating CTP:phosphocholine cytidylyltransferase (20). Differences in the activity of diacylglycerol cholinephosphotransferase were not observed in the two cell lines. Activities in control cells and NLSD cells were 0.314 and 0.306 nmol/min/mg protein, respectively (averages of two independent preparations that each varied by less than 10%). During the same chase period, triacsin C caused the amount of label in PE and sphingomyelin to decrease approximately 40% in both control and NLSD cells at 48 h but had no effect on the amount of oleic acid incorporated into PI-PS.
The increased labeling of PC and the lack of change in labeled PI-PS in both the presence and the absence of triacsin C suggests that normally PC and PI-PS acquire new acylchains both through reacylation and through transfer of labeled acyl-groups that remain esterified to a glycerol backbone, either as a mono-or diacylglycerol. In control cells, triacsin actually increased incorporation of [ 14 C]oleate into PC during the chase period, suggesting that the transferred oleate was still attached to a glycerol backbone. For PE, however, hydrolyzed fatty acid must be a major component of the chase labeling because this was blocked 40% during the first 48 h of the chase. Triacsin C also appeared to block the acyl-group addition to sphingosine to form ceramide, the precursor of sphingomyelin (Fig. 3D). In NLSD cells, on the other hand, little increase in any phospholipid was observed, and in the presence of triacsin C, labeling decreased greatly in every phospholipid except PI-PS. These data suggest that in NLSD cells, even in the absence of triacsin C, neither a fatty acid nor a mono-or diacylglycerol derived from labeled TG is available for phospholipid synthesis or remodeling.
Triacsin Inhibits de Novo Glycerolipid Synthesis but Not Reacylation-In order to determine the effect of triacsin C on glycerolipid synthesis, normal and NLSD cells were incubated with the acyl-CoA inhibitor plus radiolabeled glycerolipid precursors. The inhibitor blocked the synthesis of phospholipid and TG from [ 3 H]glycerol 80 and 99%, respectively, in control and NLSD fibroblasts (Fig. 4A), indicating that acylation of glycerol-3-phosphate, lysophosphatidic acid, and DAG were severely impaired. Incorporation of [ 14 C]oleate into TG was blocked 95%, consistent with impaired acylation in the de novo pathway. On the other hand, the incorporation of [ 14 C]oleate into phospholipid in the presence of triacsin remained unchanged (Fig. 4B), suggesting that the ability of the cells to reacylate lysophospholipids remained intact. Thus, triacsin C has differential effects on TG synthesis, de novo phospholipid synthesis, and reacylation of lysophospholipids. These data also suggest either that transfer of labeled fatty acid from TG to phospholipid either does not require activation to an acyl-CoA or that individual acyl-CoA synthetases exist that differ in both their sensitivity to triacsin C and in their production of acyl-CoAs that are present in separate, non-mixing pools that vary in their metabolic fates. Compartmentalization of acyl-CoA synthetases and their differential inactivation by triacsin has been described in yeast (21,22).

NLSD Cells May Contain Separate Functional Diacylglycerol Pools for TG and Phospholipid
Synthesis-In order to determine whether TG hydrolysis and fatty acid reutilization allows the glycerol backbone of TG to be used for phospholipid synthesis, we labeled cells for 19 h with [ 3 H]glycerol (2 M). We added 0.1 mM sodium oleate to promote moderate TG synthesis (1). This was followed by a chase in the presence or the absence of triacsin C. Control cells incorporated more label into phospholipid than TG; the reverse was true for NLSD cells (Fig. 5). The 3 H label most certainly remains as glycerol because the 80% block in phospholipid synthesis and 100% block in TG synthesis (Fig. 4A)   As demonstrated earlier, this difference is due to dilution of the [ 3 H]TG into a large unlabeled pool in the NLSD cells. All the [ 3 H]glycerol label lost from cellular lipids was recovered as free glycerol in the medium (data not shown). The rate of loss of [ 3 H]glycerol from phospholipid was similar in both cell lines. The glycerol uptake studies (Fig. 4), strongly suggest that the defect in NLSD cells lies in impaired phospholipid synthesis from a TG-derived mono-or diacylglycerol precursor.
When triacsin C was present to block re-esterification, 30% of the labeled TG in NLSD cells was lost. The increase in phospholipid label observed in the presence of triacsin C in control cells (Fig. 5A) suggests that some MAG or DAG that would have normally been used for TG resynthesis by the controls was diverted to phospholipid synthesis. Lack of a similar increase in phospholipid labeling in the NLSD cells (Fig.  5B) suggests that the cells cannot use released MAG or DAG for phospholipid synthesis even though reuse of DAG for TG synthesis has been blocked.
Analysis of the [ 3 H]glycerol-labeled phospholipids from this experiment identified the movement of a DAG or MAG precursor more specifically (Fig. 6). Phosphatidylcholine was the major labeled phospholipid in both cell lines. When triacsin C was absent, 55 and 65% of the labeled phosphatidylcholine was lost from control and NLSD cells, respectively, by 48 h. Triacsin C prevented much of this loss from control cells during the first 48 h but had little effect on the loss from NLSD cells, suggesting that DAG resulting from TG hydrolysis might be used for PC biosynthesis in control cells but not by NLSD cells. Surprisingly, PE, which is believed to be synthesized from the same pool of DAG in the endoplasmic reticulum as PC, showed a completely different pattern of labeling. In both cell lines about 33% of the [ 3 H]glycerol label was lost from PE during the chase. Further, PI-PS labeling changed very little whether or not triacsin C was present. These differences imply the existence of different functional precursor pools for each of the phospholipid species. Others have inferred compartmentalization of several phospholipid pathways because of differences in both [ 32 P]/[ 3 H] labeling ratios (23) and channeling of PC metabolites (24).
In order to determine whether oleate-loaded cells would behave differently in their metabolism of 3 H-labeled glycerolipids, control cells were incubated with 1.0 mM sodium oleate for 24 h, and then both control and NLSD cells were labeled with [ 3 H]glycerol and 0.1 mM oleate. Both cell lines primarily incorporated [ 3 H]glycerol into TG (Ͼ78% of total glycerolipid) (Fig.   7). In contrast to the loss of 82% of labeled TG in 24 h from control cells that had not been oleate-loaded (Fig. 5A), oleateloaded control cells lost only 52% of the initial dpm. (Fig. 7A). Even when triacsin C was present, very little increased loss was seen. With the NLSD cells, however, no label was lost from TG unless triacsin was present to block released fatty acid reacylation of a (presumably) partially hydrolyzed [ 3 H]glycerol backbone. In the oleate-loaded control cells with triacsin C present during the chase period, [ 3 H]glycerol label increased in phospholipid, again suggesting that diacyl[ 3 H]glycerol or monoacyl[ 3 H]glycerol was being used as a precursor for phospholipid biosynthesis. No such triacsin-induced increase was observed in the NLSD cells.
When the phospholipids were analyzed, (Fig. 8) the rate of [ 3 H]glycerol label lost from each phospholipid species in both cell lines was similar to each other and to the rates of loss from control cells that had not been oleate-loaded (see Fig. 6). Triacsin C, however, had different effects in the lipid-loaded cells. It did not block loss of labeled glycerol from PC, suggesting that the [ 3 H]glycerol was now diluted into a large unlabeled pool of TG that becomes a precursor for PC synthesis. Additionally, the slight protection against loss from PI-PS was similar in both cell lines and in the non-oleate-loaded cells (Fig. 6C). The labeling pattern was markedly different, however, in PE where triacsin C prevented the loss of label only in the control cell line (Compare Figs. 6B and 8B). This experiment suggests that in oleate-loaded control cells, some of the DAG released from TG forms a pool from which PE, but not PC, can be synthesized. 3 In the NLSD cells, which always contain a large amount of TG, the phospholipid species showed identical changes in both experiments (compare Figs. 5 and 6 with Figs. 7 and 8) and appear unable to use this putative DAG pool. DISCUSSION NLSD is an autosomal recessive disease characterized by the accumulation of TG in non-membrane bounded cytosolic droplets in a variety of cell types. Vacuolated neutrophils are present in all reported patients, but the presence of ichthyosis, fatty liver, myopathy, cardiomyopathy, mental retardation, neurosensory hearing loss, and cataracts is variable. The TG content of control human skin fibroblasts is normally extremely low; nearly confluent control fibroblasts contained only 0.06% as much TG as NLSD cells. Others have reported that NLSD fibroblasts contain 20-fold more TG than in control cells and that even when lipid is removed from culture medium, NLSD cells do not substantially decrease their TG content (5,7). Because cellular TG content is high and TG turnover is severely impaired, it was hypothesized that NLSD is caused by a deficiency of a cytosolic lipase that is unable to function appropriately, either because it is mislocalized or because an essential activator or colipase is missing (11,12). Our studies now clearly show that lipase activity in NLSD cells functions normally and that the defect instead lies in phospholipid synthesis from a TG-derived diacylglycerol or monoacylglycerol intermediate.
These studies also strongly suggest that de novo and recycling pools of diacylglycerol may not mix, that cellular acyl-CoA synthetases differ in their sensitivity to triacsin inhibition, and that acyl-CoAs may be channeled within separate metabolic pathways.
In order to increase cell TG content, we supplemented normal human fibroblasts with 1.0 mM sodium oleate. Cultured cells incubated with a large amount of fatty acid increase the synthesis and storage of TG proportional to the amount of offered fatty acid, whereas net incorporation into phospholipids and neutral lipids other than TG does not change (1). The resulting 15-fold increase in TG in oleate-loaded control cells made them similar to NLSD cells in both TG content and histological appearance. When oleate-loaded control cells were put into fresh medium to mobilize the lipid pool, TG decreased at a rate of approximately 0.5 g/g of DNA/day. Because the TG content in normal human fibroblasts was 0.9 g/g of DNA (Table I), the entire TG storage pool would turn over in less than 48 h, consistent with previous studies showing rapid depletion of radiolabeled TG from control fibroblasts (5,7,10).
Because the acyl-CoA synthetase inhibitor triacsin C prevents hydrolyzed fatty acids from being activated for re-esterification, the drug can expose the true amount of TG hydrolysis that occurs in human fibroblasts. When control and NLSD fibroblasts were treated with triacsin C, both lost the same absolute mass of TG and hydrolyzed their intracellular TG at similar rates. Similarly, in cells labeled with [ 14 C]oleate or [ 3 H]glycerol and chased in the presence of triacsin C, the loss of label from TG was similar in NLSD cells and in oleate-loaded control cells. Taken as a whole, these studies show that although NLSD cells contain a TG lipase activity that functions normally, lipase function is masked because TG is rapidly resynthesized. This resynthesis appears to occur because of a block in TG recycling to phospholipid. If fibroblasts and other non-adipose cells mobilize TG primarily in order to channel a mono-or diacylglycerol intermediate into phospholipid, the difference between the rate of TG loss from control and NLSD cells may approximate the normal turnover of TG to form phospholipid.
The function of stored TG within cells other than adipocytes has not been established. Although several studies have indi-cated that mono-or diacylglycerol derivatives of TG are used for phospholipid synthesis, how this might occur is unclear. The terminal steps in the synthesis of both TG and the phospholipids have all been localized to the endoplasmic reticulum; thus, unless such activities have an additional location associated with the lipid droplet itself, a mechanism would be required to transport mono-or diacylglycerol intermediates to the endoplasmic reticulum. If monoacylglycerols were the major intermediates used for phospholipid synthesis, once at the endoplasmic reticulum, they would require phosphorylation to form lysophosphatidic acid and esterification to form phosphatidic acid (25) (Fig. 9). Phosphatidic acid is the precursor both for CDP-DAG and the anionic phospholipids PI and PG and for DAG and its phospholipid products, PC, PE, and PS. If diacylglycerols were the major intermediates transported from the lipid droplet to the endoplasmic reticulum, PC and PE could be synthesized directly, or the DAG could be phosphorylated by diacylglycerol kinase to form phosphatidic acid. In either case, one might expect concordance of PC and PE synthesis. Lack of concordance in our studies suggests that the synthesis of PC and PE from DAG is regulated independently.
We detected major differences in phospholipid synthesis between control and NLSD fibroblasts. Although the total amount of [ 14 C]oleate initially incorporated into phospholipid was similar in control and NLSD cells (Fig. 2), during the chase, radiolabel increased as much as 5-fold in phospholipid species in the control cells, whereas there was no change in the NLSD cells. This progressive increase in labeled phospholipids during the chase period suggests that acyl groups are actively recycled from TG to phospholipid. Cook and Spence (1) suggest that intact DAG produced from TAG forms the backbone for new phospholipid molecules. Triacsin C seems to decrease this turnover in PE and sphingomyelin, perhaps by inhibiting acyl-CoA formation specific for remodeling these phospholipids. To explain the increase in labeled acyl-groups transferred to PC when triacsin C is present, one must assume that either acylated DAG is transferred from TG to PC by condensation of the TG-derived diacylglycerol with CDP-choline (26) or that triacsin's block in fatty acid activation is incomplete and that virtually all of the acyl-CoA synthesized is used to selectively reacylate lysoPC but not lysoPE. Because triacsin C blocks incorporation of [ 3 H]glycerol into phospholipid (Fig. 4), selective regulation seems unlikely, and de novo synthesis of DAG also seems to be ruled out. Instead, these results suggest that DAG released from stored TG can be used to form PC and PI-PS, but that PE and sphingomyelin depend more on the reutilization of released acyl groups. Partial hydrolysis of TG would produce DAG and fatty acids, both activators of CTP: phosphocholine cytidylyltransferase, the rate-limiting step in CDP-choline synthesis.
Although the uptake and incorporation of labeled glycerol is insignificant in neuroblastoma cells (1) and in Swiss 3T3 cells (25), in the presence of 0.1 mM oleate, fibroblasts were able to take up glycerol and incorporate enough of it into phospholipid and neutral lipid for the study of the metabolism of the labeled products. Normal and NLSD cells incorporated virtually identical amounts of [ 3 H]glycerol into total glycerolipid. However, the loss of labeled PC was rapid compared with that of the other phospholipids, suggesting a more rapid turnover of the glycerol backbone of PC. In contrast to the changes observed in incorporation of labeled oleate into phospholipid, the rate of glycerol loss was similar in control and NLSD cells.
NLSD cells did not modify the pattern of phospholipid acylation during a 4-day chase (PC decreased slightly during the chase period, Fig. 2), and triacsin C treatment decreased not only radiolabeled PE and sphingomyelin but also PC. These observations suggest that in NLSD cells, the acylated glycerol backbone of TG cannot be incorporated into phospholipid but is instead normally reacylated and converted back into TG. In addition, no accumulation of radiolabeled DAG and/or free fatty acids is observed in NLSD cells (8,11). If the cytosolic TG lipase functions normally and its fatty acid, MAG, and DAG products cannot be metabolized into phospholipid, phospholipid in NLSD cells must originate primarily via de novo synthesis. When Williams et al. (8) pulsed NLSD fibroblasts with [ 14 C]oleic acid, oleate was rapidly incorporated into PE and PC, suggesting that the cellular DAG pool destined for phospholipid synthesis might be extremely small in NLSD cells. The small DAG pool could be caused by failure to release DAG from a TG storage pool.
What might be the source of the NLSD clinical phenotype? The total amount of phospholipids and the distribution of individual phospholipid species in NLSD cells from one family was similar to that of control fibroblasts (5). In 24-h labeling experiments using [ 14 C]octanoic acid or [ 3 H]oleic acid, no differences were observed in NLSD or control fibroblasts in the percentages of label incorporated into the total phospholipid fraction or into PC, PE, and sphingomyelin (12). If no differences in the relative amounts of phospholipid species exist in NLSD cells, there might be abnormalities in the fatty acid composition of phospholipids that are precursors of lipid second messengers. Although the fatty acid composition of total lipid in NLSD fibroblasts was identical to that of control fibroblasts, individual phospholipid species were not analyzed (5). Further, fatty acid composition may well be similar in cells grown for many generations under identical culture conditions. The phenotypic abnormalities of ichthyosis, fatty liver, myopathy, and mental retardation in NLSD make it clear, however, that the recycling pathway may be critical for the normal function of skin, liver, muscle, and the central nervous system.