The turnover of cytoplasmic triacylglycerols in human fibroblasts involves two separate acyl chain length-dependent degradation pathways.

Cultured fibroblasts from patients affected with the genetic metabolic disorder named neutral lipid storage disease (NLSD) exhibit a dramatic accumulation of cytoplasmic triacylglycerols (Radom, J., Salvayre, R., Nègre, A., Maret, A., and Douste-Blazy, L.(1987) Eur. J. Biochem. 164, 703-708). We compared here the metabolism of radiolabeled short-, medium- and long-chain fatty acids in these cells. Short/medium-chain fatty acids (C4-C10) were incorporated into polar lipids (60-80%) and triacylglycerols (20-40%) at a lower rate (5-10 times lower) than long-chain fatty acids. Pulse-chase experiments allowed to evaluate the degradation rate of cytoplasmic triacylglycerols in normal and NLSD fibroblasts and to discriminate between two catabolic pathways of cytoplasmic triacylglycerols. Short/medium-chain (C4-C10) triacylglycerols were degraded at a normal rate in NLSD fibroblasts, whereas long-chain (C12 and longer) triacylglycerols remained undegraded. These data are confirmed by mass analysis. The use of diethylparanitrophenyl phosphate (E600) and parachloromercuribenzoate (PCMB) inhibitors allows to discriminate between the two triacylglycerol degradation pathways. E600 inhibited selectively the in situ degradation of short/medium-chain triacylglycerols without inhibition of the degradation of long-chain triacylglycerols, whereas PCMB inhibited selectively the in situ hydrolysis of long-chain triacylglycerols without affecting the degradation of long-chain triacylglycerols. This was correlated with the in vitro properties of cellular triacylglycerol-hydrolyzing enzymes characterized by their susbtrate specificity and their susceptibility to inhibitors; the neutral lipase specific to long-chain triacylglycerols is inhibited by PCMB, but not by E600, in contrast to short/medium-chain lipase, which is inhibited by E600 but not by PCMB. The data of in vitro and in situ experiments suggest the existence in fibroblasts of two separate acyl chain length-dependent pathways involved in the degradation of cytoplasmic triacylglycerols, one mediated by a neutral long-chain lipase and another one mediated by a short/medium-chain lipase.

In human fibroblasts, triacylglycerols are degraded in two separate and independent subcellular (lysosomal and cytoplasmic non-lysosomal) compartments (1). In the lysosomal compartment, triacylglycerols and cholesteryl esters of low density lipoprotein and very low density lipoprotein taken up by cells are degraded by the acid lysosomal lipase (2,3), which is genetically deficient in Wolman disease (4). In the cytoplasmic compartment of fibroblasts, triacylglycerols are degraded by a lipase system (yet poorly characterized) different from the other known cellular (hormone-sensitive lipase) or secretory lipases (1,5) and different from the cholesteryl ester degradation pathway (6). As shown by using radiolabeled oleic acid or pyrene-containing fluorescent fatty acids, this degradative pathway of cytoplasmic triacylglycerols is deficient in the neutral lipid storage disease (NLSD) 1 (1,5,(7)(8)(9)(10), a rare inherited metabolic disease generally characterized by the association of muscular weakness, ichthyosis, and multisystemic triacylglycerol storage (11,12). We have recently reported that the pool of cytoplasmic triacylglycerols accumulated in fibroblasts from neutral lipid storage disease is constituted by triacylglycerols endogenously biosynthesized (7)(8)(9)(10) and by triacylglycerols taken up from high density lipoproteins (13). An increased biosynthesis of phosphatidylethanolamine has also been reported (14).
Comparative studies of the uptake and metabolic utilization of short-chain and long-chain fluorescent pyrene fatty acids in cultured lymphoblasts (15) and in fibroblasts (16) have shown that short-chain (pyrene-butanoic) acid was incorporated into phospholipids but not in triacylglycerols, whereas long-chain (pyrene-decanoic and pyrene-dodecanoic) fatty acids were incorporated into phospholipids and triacylglycerols. In lymphoblasts from neutral lipid storage disease pulsed with shortchain pyrene fatty acid, we detected no significant accumulation of fluorescent triacylglycerols (15). These results led us to speculate that replacing natural long-chain fatty acids by short-chain fatty acids could be of interest to slow down the triacylglycerol accumulation in neutral lipid storage disease cells. However, the conclusions obtained with fluorescent pyrene fatty acids cannot be directly transposed to natural fatty acids because natural and pyrene fatty acids exhibited several known metabolic differences (5,8,17). The apparent lack of incorporation of short-chain fluorescent fatty acid (pyrene-butanoic) in triacylglycerols and of accumulation of pyrene-butanoic containing triacylglycerols in NLSD cells could be due either to a lack of biosynthesis of pyrene-butanoic-containing triacylglycerols (despite the biosynthesis of pyrene-butanoic-containing glycerophospholipids) or to a rapid degradation of pyrene-butanoic-containing triacylglycerols by a cytoplasmic pathway independent of the neutral lipase system (deficient in intact NLSD cells). This latter hypothesis is supported by the existence in fibroblast homogenates of five different enzymes able to hydrolyze in vitro long-or short-chain fluorescent triacylglycerols. Both groups of enzymes referred to as long-chain lipases or short/medium-chain lipases, respectively, can be discriminated by their enzymatic properties (heat stability, effect of inhibitors) (1,18), with short/medium-chain lipases exhibiting properties of (nonspecific) carboxylesterases (19).
This prompted us to investigate the metabolism of short-and long-chain fatty acids in cultured fibroblasts from controls and from neutral lipid storage disease to examine the metabolic fate of short-chain fatty acids and their potential influence on the triacylglycerol accumulation in neutral lipid storage disease.
The data reported here showed that, in comparison to longchain fatty acids, short/medium-chain fatty acids 1) are incorporated into cellular lipids at a much lower rate, 2) induce a lower accumulation of triacylglycerols in NLSD cells, and 3) contained in triacylglycerols are hydrolyzed at a normal rate through a catabolic pathway involving a short/medium-chain lipase activity not deficient in situ in NLSD cells (in contrast to the defect of the long-chain lipase activity). C]decanoic acid (10 mCi/mmol), and unlabeled fatty acids (butanoic, hexanoic, octanoic, decanoic, dodecanoic, palmitic, and oleic acids) were from Sigma; [ 14 C]oleic acid (47 mCi/ mmol) and LM-1 autoradiographic emulsion for light microscopy was from Amersham (Les Ulis, France); Silica Gel G and RP18 thin layer chromatography plates were from Merck (Darmstadt, Germany); RPMI 1640 with Glutamax®, fetal calf serum, streptomycin, and penicillin were from Seromed (Strasbourg, France); Ultroser G was from IBF (Villeneuve-la-Garenne, France); Instafluor was from Packard (Warrenville, IL); Aquasafe 300 Plus was from Zinsser Analytic (Maidenhead, United Kingdom); and the other reagents were from Merck or Prolabo (Paris).

Chemicals-[
Cell Cultures-Skin fibroblasts were obtained from normal individuals (no1, no2), from two patients affected with NLSD (Bo. and Dem.) and with Wolman disease (GM1606). NLSD fibroblasts (N1 and N2) were kindly provided by Drs. J. M. Mussini and S. Billaudel (Nantes, France) and by Dr. B. Winchester (London). GM1606 cells were purchased from the NIGMS, National Institutes of Health, Human Genetics Mutant Cell Repository (Camden, NJ.). Other fibroblasts from normal subjects (no1, no2) were from our laboratory. Lymphoblastoid cell lines were obtained by Epstein-Barr virus (B95/8), and transformation of blood B lymphocytes from normal individuals (no3 and no4) or from a patient affected with NLSD (Bo) was as previously indicated (5). Fibroblasts and lymphoblasts were grown at 37°C in 5% CO 2 , 95% air in RPMI 1640 medium with penicillin (100 units/ml) and streptomycin (100 g/ml) and supplemented with 10% fetal calf serum or 2% Ultroser G or HY (lipoprotein-free serum substitutes for fibroblasts and lymphoblasts, respectively) under the previously used conditions (14,15). Cell cultures were always used in the same growth state since the level of fatty acid incorporation into cellular lipids may be influenced by the growth rate of cells (18); fibroblasts were used at confluency, and lymphoblasts were used in exponential growth phase.
Cell viability was assessed by 3-(4,5-dimethyl thiazol-2-yl)-2,5diphenyl tetrazolium bromide test (20) or by trypan blue test and by morphological examination as previously used (21), which shows that the number of necrotic and apoptotic cells did not exceed 8 -10% during the metabolic experiments.
Pulse-Chase Experiments with Fatty Acids-The culture conditions defined to induce a large cellular biosynthesis and accumulation of radiolabeled triacylglycerols were derived from data previously reported (7, 8) and from preliminary experiments. Before studies were initiated, fibroblasts and lymphoblasts were grown in a lipoprotein-free medium (RPMI 1640 containing 2% Ultroser G) for the period of time indicated in the figure legends.
Cells were pulsed with fatty acids, used at the concentrations indicated in the text, generally 30 nmol/ml and 10 6 dpm, solubilized in Me 2 SO (0.5% final concentration), and preincubated with the culture medium (for 30 min at 37°C) before addition to the cell culture. At the end of the pulse period (12 or 24 h, as indicated in the text), the fatty acids were removed by washing the cells twice with phosphate-buffered saline containing 10 nmol/ml bovine serum albumin (essentially fatty acid-free) and twice with phosphate-buffered saline. One batch of cells was harvested at the end of the pulse (time 0 of the chase); the other batches were grown in a fresh medium containing 2% Ultroser G (but no additional fatty acids) and harvested at the indicated time of this chase period. Under the experimental conditions used here, fatty acids and Me 2 SO had no adverse effect on cell viability.
Lipid Extraction and Analysis-At the indicated time, cells were washed twice with phosphate-buffered saline supplemented with 10 nmol/ml bovine (essentially fatty acid-free) serum albumin and twice with phosphate-buffered saline. Lymphoblasts were pelleted by centrifugation (2,000 ϫ g for 10 min), and fibroblasts were harvested by scraping with a rubber policeman. Cells were homogenized in 1 ml of distilled water by sonication (2 cycles of 15 s, Soniprep 150). 1 aliquot (50 l) was used for protein determination, and another (50 l) was used for counting the total cellular radioactivity. The remaining aliquot was used for lipid analyses by two procedures. Cellular lipids were extracted by the Folch procedure (22) and separated by TLC on Silica Gel G plates, using petroleum ether/diethyl ether/acetic acid (80/20/1 (v/v/v)) for the separation of the neutral lipids and chloroform/methanol/ water (100/42/6 (v/v/v)) for the separation of the phospholipids. Lipid spots were visualized by iodine vapors, and the radiolabeled lipids were determined directly on the TLC plate by using a TLC-radiochromatoscanner Berthold under the previously used conditions (9, 10). Alternatively (when using low levels of radiolabeled lipids), triacylglycerols (the major radiolabeled neutral hydrophobic lipids) were separated from polar lipids by a solvent partition system derived from the Dole's procedure (23) as previously used (13). Briefly, the chloroformic phase of the Folch extract was evaporated under nitrogen, and the 1.25 ml of alkaline Dole's mixture (isopropyl alcohol/heptane/1 M, pH 10.5, NaOHglycine buffer, 40/10/1), 0.75 ml of heptane, and 1 ml of water were added to the lipid residue. After mixing and centrifuging, the phases were separated and backwashed once with the fresh phases (the upper phase with 2 ml of fresh lower phase and the lower phase with 1 ml of fresh upper phase). Under these conditions, the recovery of triacylglycerols in the heptane phase was better than 98%, and the contamination of this heptane phase by phospholipids was lower than 0.4%; reversely, the contamination of the aqueous phase by triacylglycerols was negligible (lower than 0.1%). These data were confirmed by the thin layer chromatography analysis (and quantification by TLC-radiochromatoscanner Berthold) of radiolabeled lipids contained in each phase (data not shown). The radioactivity was determined by liquid scintillation counting in Picofluor® using a Packard ␤ counter (Tricarb 4530).
Mass analysis of cellular lipids was performed by gas liquid chromatography. After adding tripentadecanoin (50 g used as standard), cellular lipids were extracted according to the Folch procedure. After drying of the chloroformic phase, lipids were partitioned in the Dole's biphasic solvent system under the above indicated conditions. After evaporation of the heptane phase, neutral lipids were solubilized in 50 l of hexane, and triacylglycerols were determined by gas liquid chromatography under the previously used conditions (24) (Carlo-Erba GC-8000; 6 M CP-Sil-5CB, 0.32-mm diameter capillary column; oven temperature programmed from 220 to 350°C, 5°C per min, flame ionization detector 370°C, carrier gas N 2 , 35 pK a ). Phospholipids (of the chloroformic phase of the Folch extract) were evaluated by their phosphorus content determined according to the method of Chen et al. (25). Alternatively, gross triaglycerol mass was estimated by a method derived from Jefferson et al. (26) (briefly, after TLC separation of neutral lipids on Silica Gel G plates as above indicated (petroleum ether/diethyl ether/acetic acid, 80/20/1 (v/v/v) as developing system) and staining by dipping in 1 N sulfuric acid in methanol, followed by air drying for 15 min and heating at 180°C for 15 min). Lipid component were quantified with a Biocom Image Station (Biocom-France) using a standard of tripentadecanoin treated under the same conditions on the same TLC plate (standard was linear between 5 and 50 g).
In Vitro Enzymatic Assays and Preparation of [ 14 C]Decanoic-containing Triacylglycerols-Radiolabeled [ 14 C]decanoic-containing triacylglycerols were biosynthesized in cultured fibroblasts or lymphoblasts pulsed for 24 h with [ 14 C]decanoic acid (30 nmol/ml, 70,000 dpm/nmol) in the presence of 1 mol/liter E600 (used to block the degradation of short-chain triacylglycerols, see Fig. 6). Then, cells were washed twice with phosphate-buffered saline and homogenized in water by sonication. Neutral lipids were extracted by partition in the Dole's solvent system (21) and [ 14 C]decanoic-containing triacylglycerols were purified by preparative thin layer chromatography on silica gel G (solvent system petroleum ether/diethyl ether/acetic acid, 80/20/1). The specific radioactivity of the [ 14 C]decanoic-containing triacylglycerols (70,000 dpm/nmol) was calculated on the basis of that of the [ 14 C]decanoic acid incorporated into the culture medium (assuming that the non-labeled and radiolabeled decanoic acids were incorporated into triacylglycerols at the same rate).
Enzymatic Assays-Enzyme solutions were prepared by homogenizing fibroblasts or lymphoblasts in distilled water and by sonication (3 cycles of 10 s, using MSE sonicator, Soniprep 150). The standard enzymatic assays of the enzymes hydrolyzing [ 14 C]decanoic-containing triacylglycerols (referred to as carboxylesterases) contained 5 nmol of radiolabeled [ 14 C]decanoic-containing triacylglycerols and 50 nmol of egg phosphatidylcholines dispersed in the 0.2 M citrate/phosphate buffer, pH 7.2, and the enzyme preparation (100 g of protein) in a final volume of 200 l. The standard assay for determining lipase activity contained 10 nmol/assay and 10 5 dpm [ 3 H]triolein (10,000 dpm/nmol), 0.1% Triton X-100, 0.2 M citrate/phosphate buffer, pH 7.2, and the enzyme solution (100 g of protein) in a final volume of 200 l. When varying the pH of the assay, we used citrate/phosphate buffers (from pH 3.5 to 7.0) and Tris-HCl (from pH 7.2 to 9.0). At the end of the incubation time (2 h at 37°C, under the standard conditions), the liberated fatty acids were extracted according to the procedure of Belfrage and Vaughan (27), and the radioactivity extracted in the aqueous phase was determined by liquid scintillation counting (in Aquasafe 300-Plus, using a Packard ␤ counter, Tricarb 4530). Enzyme activities of the homogenates were generally expressed as nmol of fatty acid liberated per hour and per milligram of cell protein. Protein concentrations were determined using the method of Lowry et al. (28).
Cell Microautoradiography-Fibroblasts, grown on microscopy cover glasses, were pulsed for 12 h with [ 14 C]decanoic or [ 14 C]oleic acid (30 nmol/ml and 2.5 Ci/ml). After washing the cells under the previously described conditions, one batch was immediately used for microautoradiography, and another batch was chased for an additional 48-h period in RPMI containing 2% Ultroser G before use for microautoradiography. Cells were fixed for 20 min at 4°C in cacodylate buffer 0.1 M, pH 8.0, containing 2% glutaraldehyde, then washed twice with distilled water; cover glasses were then immersed in LM-1 autoradiographic emulsion for light microscopy, under the procedure indicated by the manufacturer. After 1-3 days (at 4°C in the dark), microautoradiographies were developed and examined by light microscopy (Leica Diaplan).

The Accumulation of Triacylglycerols in NLSD Fibroblasts Is Largely Dependent on the Presence of Lipids in the Culture
Medium-As shown in Fig. 1, triacylglycerol levels are higher in NLSD than in normal cells, in agreement with previously reported results (7)(8)(9)(10). But these data also show that the rate of accumulation of triacylglycerols in NLSD cells is largely dependent on the intake of extracellular lipids. NLSD cells grown in culture medium supplemented with 10% fetal calf serum exhibit a large accumulation of triacylglycerols, whereas the triacylglycerol accumulation was relatively lower in NLSD cells grown in a serum-free medium for 15 days (when triacylglycerol levels are related to the cell proteins) (Fig. 1). This led us to hypothesize that the rate of triacylglycerol accumulation in NLSD cells may be slowed down by modifying the lipid composition of the culture medium. Moreover, as we have previously reported that short-chain pyrene fatty acids did not induce any storage of fluorescent triacylglycerols in NLSD cells (because of the lack of incorporation of short-chain fluorescent fatty acid analogs into triacylglycerols) (15), we studied the metabolism of natural (non-fluorescent) fatty acids with various acyl-chain lengths and compared their influence on the triacylglycerol storage in NLSD cells.
Incorporation of Radiolabeled Fatty Acids with Various Acyl Chain Lengths into Cellular Lipids-When fibroblasts were grown for 12 h in a medium containing a fixed concentrations of radiolabeled fatty acids (30 nmol/ml, 33,000 dpm/nmol) with various chain lengths, the level of the radiolabeled fatty acids incorporated into cellular lipids at the end of this pulse period was largely dependent on the acyl chain length (Fig. 2). Shortand medium-chain fatty acids used here (from butanoic to dodecanoic acid) were incorporated into cellular lipids at a lower rate (5-10 times lower) than the long-chain fatty acids (palmitic and oleic acids). Under the conditions used here (pulse for 12 h with 30 nmol/ml fatty acid), incorporation of radiolabeled fatty acids into triacylglycerols of normal cells ranged between 20 and 40% (of total radiolabeled lipids) (Fig.  3). A large part (60 -80%) of the cell-associated fatty acids was incorporated into phospholipids (Fig. 3). Medium-chain and long-chain fatty acids were incorporated in the main classes of phospholipids, but [ 3 H]oleic acid incorporated in phosphatidylethanolamine was higher than that of [ 14 C]octanoic acid (Table I). Under the used conditions, we observed no significant difference between the phospholipid classes of NLSD and controls. At the end of the pulse, the radioactivity was localized in both the plasma membrane and the cytoplasm (Fig. 4, A, C, E, and G).
It is noteworthy that, under the used experimental conditions of the pulse (12 or 24 h), elongation of C10 incorporated into cellular lipids was negligible. Reversely, we did not detect any appreciable amount of short derivatives of [ 3 H]oleic acid incorporated into cellular lipids. Similar data were observed with lipids extracted at the end of the chase (data not shown). These data allow to conclude that the chain lengths of the main part of fatty acids incorporated into cellular lipids remain unaltered during the time of the experiments reported here.
Study of the Degradation of Endogenously Biosynthesized Triacylglycerols Containing Fatty Acids with Various Chain Lengths (Figs. 3 and 4)-Pulse-chase experiments clearly showed that the short-and medium-chain radiolabeled fatty acids (from C4 to C10) incorporated into triacylglycerols were degraded at a similar rate in NLSD and in control fibroblasts (Fig. 3, A-D). In contrast, triacylglycerols containing radiolabeled C12 and longer fatty acids (C16 and C18:1) were degraded only in normal cells but not in NLSD fibroblasts (Fig. 3,  E-H). The data, summarized in Fig. 3, D and H, show the clear-cut discrimination between the group of fatty acids that accumulated in triacylglycerols of NLSD (C12 and longer fatty acids) and the group of short/medium-chain fatty acids that did These conclusions were also supported by morphological (microautoradiographic) studies of the cellular radioactivity during pulse-chase experiments (Fig. 4). At the end of the pulse period, the level of radioactivity incorporated into the cells was higher with [ 14 C]oleic acid than with [ 14 C]decanoic acid, and the radioactivity was localized in the membranes and in the cytoplasmic compartment (Fig. 4, A, C, E, and G). These morphological features are in agreement with the metabolic studies (Fig. 3), which showed the incorporation of radiolabeled fatty acids into phospholipids and triacylglycerols. At the end of the chase, when cells were labeled with [ 14 C]oleic acid, the cytoplasmic radioactivity was lost in normal cells (Fig. 4F) but was still apparent in NLSD cells (Fig. 4H). This was quite consistent with the degradation of oleyl-containing triacylglycerols in normal but not in NLSD cells (Fig. 3G). In contrast, in cells labeled with [ 14 C]decanoic acid, the cytoplasmic radioactivity disappeared almost completely at the end of the chase in normal cells and in NLSD cells as well, the radioactivity being persistent only in cell membranes (Fig. 4, B and D). This was consistent with the normal degradation of [ 14 C]decanoic acidcontaining triacylglycerols in NLSD cells (Fig. 3D).
As the accumulation of long-chain triacylglycerols in NLSD cells is due to a defect of the degradation pathway, it was suggested that, in the cytoplasmic compartment, short/medium-chain triacylglycerols could be degraded by an enzymatic system specific to short/medium-chain triacylglycerols and different from that hydrolyzing long-chain triacylglycerols (neutral lipase).
Mass Evaluation of Medium-chain and Long-chain Triacylglycerol Metabolism (Fig. 5)-To confirm by mass quantitation of medium-and long-chain triacylglycerols the results obtained with radiolabeled fatty acids, we used immortalized lymphoblasts from normal subjects and from a patient affected with NLSD. We used lymphoblasts for mass quantitations because 1) immortalized lymphoblasts are rapidly growing cells (doubling time, 24 -36 h) and grow in suspension, thus permitting to obtain relatively easily large amounts of cell material (necessary for mass quantitation and analysis of lipids), and 2) NLSD lymphoblasts exhibit the same block of triacylglycerol degradation as fibroblasts (5).
Lymphoblasts grown in the presence of 30 nmol/ml C10 or C18:1 (for 24 and 12 h, respectively) showed that the levels of triacylglycerols, at the end of the pulse, were much lower in cells grown in the presence of C10 than in cells grown in the presence of C18:1 (Fig. 5). In normal cells, the levels of triacylglycerols were considerably reduced at the end of the 48-h chase (Fig. 5, A and B). In NLSD cells grown in the presence of oleic acid, the triacylglycerol level was almost unchanged (Fig.  5B), whereas it was reduced by about 50% in cells fed with C10 (Fig. 5A). Analysis of the triacylglycerol species in NLSD cells fed with C10 showed that triacylglycerols with a carbon number lower than 45 (triacylglycerols containing C10 or short/ medium-chain fatty acid) were degraded during the chase, whereas triacylglycerols with a carbon number higher than 48 (long-chain triacylglycerols) were not degraded (Fig. 5A). In   FIG. 2. Incorporation of radiolabeled short-, medium-, and long-chain fatty acids into cellular lipids of cultured fibroblasts from controls and patients with NLSD and Wolman disease at the end of the pulse period. Cells were incubated for 12 h in RPMI 1640 medium supplemented with 2% Ultroser G and fatty acids (radiolabeled fatty acids, 10 6 dpm/ml, were mixed with non-labeled fatty acid, 30 nmol/ml, i.e. 33,000 dpm/nmol). At the end of the pulse, fibroblasts were washed and harvested. Lipids were extracted by the procedure of Folch, and the radioactivity of the chloroformic phase was determined by liquid scintillation counting. The level of radiolabeled fatty acids incorporated into cellular lipids was calculated assuming that the radiolabeled and non-labeled fatty acids have the same metabolic fate (and on the basis of 33,000 dpm/nmol). Mean Ϯ S.E. of four experiments.

FIG. 3. Degradation of triacylglycerols and polar lipids (insets) endogenously biosynthesized from radiolabeled short-, medium-, and long-chain fatty acids incorporated in the culture medium of fibroblasts from normal subjects (filled squares and circles) and from patients with NLSD (empty squares and circles) and Wolman disease (triangles).
Cells were incubated at 37°C for 12 h (pulse period) in RPMI 1640 supplemented with 2% Ultroser G, and radiolabeled fatty acids (10 6 dpm/ml and 30 nmol/ml non-labeled fatty acid) are as described under "Materials and Methods." At the end of the pulse period, cells were washed and incubated for an additional 24 or 48 h (chase period) in RPMI 1640 supplemented with 2% Ultroser G. At the indicated times, cells were washed and harvested, and the lipids were extracted and analyzed as described under "Materials and Methods." Mean Ϯ S.E. of three experiments. D and H, comparison of the levels of undegraded radiolabeled triacylglycerols in the various cell types at the end of the pulse (whole bars) and at the end of the 48-h chase period (black sections) with various fatty acids. cells fed with C18:1, the major part of triacylglycerols have a carbon number higher than 48 and were not degraded in NLSD cells (Fig. 5B). These data confirm the results observed with radiolabeled fatty acids, supporting the idea that only longchain containing triacylglycerols accumulate in NLSD cells, whereas medium-chain containing triacylglycerols do not accumulate in these cells. (Figs. 6 and 7)-The above reported data showed that the in situ degradation of C10-containing triacylglycerols is not defective in NLSD cells in contrast to the degradation of long-chain triacylglycerols, which is genetically deficient. This led us to formulate the hypothesis that these two types of triacylglycerols are probably degraded through the two different degradation pathways. This suggests the existence of two cytoplasmic enzymes, a short/medium-chain lipase and a long-chain lipase, respectively. This hypothesis was supported by two types of experiments: 1) by blocking selectively in situ each degradation pathway by irreversible inhibitors and 2) by evaluating in vitro the activity of each enzyme in homogenates of cells pretreated by these irreversible inhibitors. These inhibitors have been selected, in preliminary experiments, on the basis of their selective inhibitory effect in vitro on candidate enzymes, namely lipases hydrolyzing long-chain fluorescent triacylglycerols and lipases hydrolyzing short-chain fluorescent triacylglycerols (this latter enzyme exhibiting properties of nonspecific carboxylesterases) (1,18). The organophosphorous compound E600 is an irreversible inhibitor of various esterases in vitro (29,30), among them the short/medium-chain lipase, but it did not inhibit the long-chain lipase (1, 18). Reversely, the sulfhydryl-reactive compound p-chloromercuribenzoate

composition of fibroblasts and lymphoblasts from NLSD patients and from normal subjects grown in the presence of radiolabeled [ 14 C]octanoic or [ 3 H]oleic acids
Cells were incubated at 37°C for 24 h (pulse period) in RPMI 1640 supplemented with 2% Ultroser G and radiolabeled fatty acids (10 6 dpm/ml and 30 nmol/ml non-labeled fatty acid) as described under "Materials and Methods." At the end of the pulse period, cells were washed and harvested, and the lipids were extracted and analyzed by TLC as described under "Materials and Methods." Mean Ϯ S.E. of three experiments. PS/PI, phosphatidylserine/phosphatidylinositol; Sm, sphingomyelin; PC, phosphatidylcholine; PE, phosphatidylethanolamine.   (lower panel, B, D, F, H). Magnification, 800ϫ.

FIG. 5. Mass analysis of cellular triacylglycerols in pulse-chase experiments with a fixed concentration of decanoic (A) or oleic acid (B) in lymphoblasts from NLSD patient or normal subject.
Cells were grown in a serum-free RPMI medium for 15 days, under the standard conditions of Fig. 1. Then, cells were pulsed with 30 nmol/ml decanoic acid for 24 h or oleic acid for 12 h. After washing, one batch was harvested for analysis (p), and another batch was grown in lipidfree medium (i.e. serum-free and fatty acid-free medium) for a 24-h chase period and washed and harvested for analysis (ch). Lipids were extracted and analyzed either by gas liquid chromatography or by thin layer chromatography as described under "Materials and Methods." Hatched and black bars represent triacylglycerols with carbon number Ͻ45 and Ͼ48, respectively (note that no significant level of triacylglycerols with the carbon number Ͻ45 was detected in cells fed with oleic acid (panel B)). Mean Ϯ S.E. of three experiments. In insets are shown the triacylglycerol TLC spots of the relative cells treated as indicated above.
(PCMB) is an irreversible inhibitor of the neutral long-chain lipase, but it did not inhibit the short/medium-chain lipase in vitro (1). Preliminary experiments allowed to show that optimal (i.e. effective in inhibiting the degradation pathways) con-centrations of these inhibitors can be used without any adverse effect on the incorporation of radiolabeled fatty acid into cellular lipids (Table II) and without any cytotoxic effect to cultured cells (data not shown).
When intact normal cells were incubated in the presence of 1 mol/liter E600, the in situ hydrolysis of short/medium-chain triacylglycerols was completely blocked (Fig. 6A), whereas that of long-chain triacylglycerols was not affected (Fig. 6B). In contrast, when intact normal cells were incubated in the presence of 10 mol/liter PCMB, the in situ hydrolysis of long-chain triacylglycerols was severely impaired (Fig. 6C), whereas that of medium-chain triacylglycerols was not altered (Fig. 6D). The separate inhibition of the degradation pathways of short/medium-and long-chain triacylglycerols by these two inhibitors supports the hypothesis of two separate degradation pathways.
The activities of the two candidate enzymes were evaluated in homogenates of cells treated with the irreversible inhibitors under conditions of Fig. 6, using in vitro assays containing either [ 14 C]decanoic-containing triacylglycerols (prepared from cells loaded either with radiolabeled [ 14 C]decanoic acid or [ 3 H]triolein). The enzyme hydrolyzing C10-containing triacylglycerols was inhibited in cells treated by E600, whereas the enzyme hydrolyzing triolein was not inhibited in the same cells (Fig. 8, C and D). Reversely, in cells treated with PCMB, the enzyme hydrolyzing triolein was inhibited, whereas the enzyme hydrolyzing C10-containing triacylglycerols retained its activity (Fig. 8, C and D).
These results support the hypothesis that cytoplasmic C10containing triacylglycerols are degraded in situ by an E600sensitive short-chain lipase (not deficient in NLSD cells), which is clearly discriminated from the neutral lipase activity (specific to long-chain triacylglycerols) involved in the liberation of long-chain fatty acids of cytoplasmic triacylglycerols and deficient in NLSD cells. DISCUSSION The data previously reported (8 -10, 31, 32) and those reported here clearly show that the cytoplasmic triacylglycerol accumulation in cultured fibroblasts is largely dependent on the extracellular lipids, since it is supplied by (at least) two separate pathways, the first one being the intracellular pathway of triacylglycerols endogenously biosynthesized from the cytoplasmic pool of fatty acids (8 -10) and the second one (probably minor) resulting from the cellular uptake of triacylglycerols contained in high density lipoproteins (13). Both sources of triacylglycerols have been shown to participate to the storage of cytoplasmic triacylglycerols in NLSD (8 -10, 13).  7. In vitro determination of activities of enzymes degrading radiolabeled decanoic acid-or oleic acid-containing triacylglycerols (A-D and E, respectively) in fibroblast homogenates (prepared in distilled water as described under "Materials and Methods"). A, linearity of the enzyme activity versus time (standard assay conditions using 100 g of protein of the homogenate per assay); B, linearity of the enzyme activity versus enzyme concentration (expressed as g of homogenate protein/assay; incubation time, 2 h); C, effect of increasing concentrations of the substrate (100 g of protein/ assay; incubation time, 2 h); D and E, hydrolytic activity when varying pH from 3.5 to 9.0 using decanoic-or oleic-containing triacylglycerols as substrates (D and E, respectively). Mean Ϯ S.E. of three experiments. In the experiments reported here, the difference between the rates of incorporation of fatty acids with various acyl chain length into cellular lipids (under the used conditions, longchain fatty acids were incorporated around 10 times faster than short-or medium-chain fatty acids) cannot be attributed to any cytotoxic artefactual effect (as assessed by the viability of 3-(4,5-dimethyl thiazol-2-yl)-2,5-diphenyl tetrazolium bromide or trypan blue tests). The selectivity of incorporation of long-chain natural fatty acids into cellular lipids is consistent with the results observed with fluorescent fatty acid analogs (15,33,34) and with natural fatty acids (35). As the ratio of triacylglycerols/polar lipids is nearly constant whatever the acyl chain length, it is suggested that the limiting step of short-chain fatty acid incorporation is probably an early metabolic step, such as transmembrane transport or biosynthetic pathway (for instance, acyl-CoA synthases are known to be acyl chain length-dependent) (36).
Natural short-chain fatty acids were really incorporated into triacylglycerols, in contrast to short-chain fluorescent fatty acids (pyrene-butanoic acid), which were incorporated into phospholipids but not in triacylglycerols (15). Therefore, the mechanism of the lack of accumulation of short/medium-chain triacylglycerols in NLSD cells is different with fluorescent fatty acids (lack of biosynthesis) and with natural fatty acids (lack of degradation block).
On the basis of genetic and inhibitor-induced metabolic blocks, it was suggested that cytoplasmic triacylglycerols are degraded through two separate catabolic pathways, one being specific to long-chain triacylglycerols (long-chain specific lipase activity) and the other one specific to short-chain triacylglycerols (short-chain specific lipase or carboxylesterase activity).
The genetic defect of the in situ triacylglycerol degradation in NLSD cells is specific to triacylglycerols containing dodecanoic and longer chain fatty acids, thus suggesting an in situ defect of the long-chain lipase activity. As the activity of the long-chain neutral lipase is apparently not deficient in vitro (10,15), it may be speculated that the mutation affects either a site involved in the lipase activity in situ (for instance, site for routing or enzyme activation) or a co-factor necessary to the lipase activity in situ by analogy with the co-lipase for the pancreatic lipase (37) or with the activator proteins (saposins) for several lysosomal enzymes (38, 39). The normal degradation of short/medium-chain triacylglycerols (up to 10 carbons) in NLSD cells suggests that these lipids are degraded through a catabolic pathway different from and unable to compensate the in situ deficient activity of the long-chain lipase.
The hypothesis of the existence of two separate hydrolytic pathways is also supported by experiments with irreversible inhibitors specific to each degradation pathway and able to inhibit the enzyme activities in situ and in vitro as well. E600 inhibited concomitantly the degradation of short/mediumchain triacylglycerols in situ and the short-chain lipase activity in vitro but not the long-chain lipase in vitro activity nor the in situ degradation of long-chain triacylglycerols. Reversely, the sulfhydryl reagent PCMB inhibited the in situ degradation of long-chain triacylglycerols and the long-chain lipase activity in vitro, but not the in situ degradation of short/medium-chain triacylglycerols and the short/medium-chain lipase activity. Moreover, the inhibition of the enzymes (activities determined in vitro by using their respective substrates) persisting after lysis of cells treated with irreversible inhibitors, E600 and PCMB (this study), and the in vitro studies of enzymatic properties previously reported (1, 40 -42) are consistent with the idea that the short-chain lipase is different from the long-chain lipase. The enzymatic properties (specificity to long-chain triacylglycerols, heat stability, susceptibility to PCMB but not to E600) of the neutral long-chain lipase (1,5,18,42) are similar to those of the microsomal neutral lipase (43)(44)(45), which has not been cloned to date, to our knowledge, and is different from the hormone-sensitive lipase (no activation of triacylglycerol degradation by dibutyryl-cAMP) (data not shown) and other known lipases by its enzymatic properties in vitro (1,5). The neutral short/medium-chain lipases exhibit some enzymatic properties (specificity to various short-chain lipophilic esters, heat lability, susceptibility to organophosphorous compound E600, but relative resistance to PCMB) similar to those of carboxylesterases (1, 40 -42), some of them having been cloned in liver or other tissues (45)(46)(47)(48). As the enzymes are not identified at the molecular level, the alternative hypothesis that there is one lipase with differential substrate specificity cannot be excluded.
Finally, from the data observed in intact cells, i.e. degradation at a normal rate of short-chain triacylglycerols in NLSD (in contrast to the accumulation of long-chain triacylglycerols), it is suggested to use diets enriched with short/medium-chain lipids (and poor in long-chain lipids) to tentatively slow down the long-chain triacylglycerol accumulation in NLSD patients.