Triglyceride alters lysosomal cholesterol ester metabolism in cholesteryl ester-laden macrophage foam cells.

In late-stage atherosclerosis, much of the cholesterol in macrophage foam cells resides within enlarged lysosomes. Similarly, human macrophages incubated in vitro with modified LDLs contain significant amounts of lysosomal free cholesterol and cholesteryl ester (CE), which disrupts lysosomal function similar to macrophages in atherosclerotic lesions. The lysosomal cholesterol cannot be removed, even in the presence of strong efflux promoters. Thus, efflux of sterol is prevented. In the artery wall, foam cells interact with triglyceride-rich particles (TRPs) in addition to modified LDLs. Little is known about how TRP metabolism affects macrophage cholesterol. Therefore, we explored the effect of TRP on intracellular CE metabolism. Triglyceride (TG), delivered to lysosomes in TRP, reduced CE accumulation by 50%. Increased TG levels within the cell, particularly within lysosomes, correlated with reductions in CE content. The volume of cholesterol-engorged lysosomes decreased after TRP treatment, indicating cholesterol was cleared. Lysosomal TG also reduced the cholesterol-induced inhibition of lysosomal acidification allowing lysosomes to remain active. Enhanced degradation and clearance of CE may be explained by movement of cholesterol out of the lysosome to sites where it is effluxed. Thus, our results show that introduction of TG into CE-laden foam cells influences CE metabolism and, potentially, atherogenesis.-Ullery-Ricewick, J. C., B. E. Cox, E. E. Griffin, and W. G. Jerome. Triglyceride alters lysosomal cholesterol ester metabolism in cholesteryl ester-laden macrophage foam cells.

was obtained from Atlanta Biologicals (Norcross, GA), and RPMI, L -glutamine, Eagle's vitamins, streptomycin, and penicillin were purchased from Mediatech (Herndon, VA). All tissue culture plasticware was purchased from Corning (Corning, NY). All other chemical reagents and chemical solvents were obtained from VWR (West Chester, PA).

Lipoprotein isolation and aggregation
Human LDL was isolated from plasma collected from fasted, normocholesterolemic human volunteers who had provided informed consent. Collection of blood followed procedures approved by the Human Subjects Institutional Review Board. LDL (1.006 g/ml < d < 1.063 g/ml) and VLDL (d < 1.006 g/ml) were isolated by sequential ultracentrifugation ( 16 ). LDL and VLDL were dialyzed against 0.9% NaCl containing EDTA (0.3 mmol/l) for 72 h, fi lter-sterilized through a Millipore fi lter (0.45 µm), and stored under nitrogen at 4°C. Isolated LDL was aggregated by vortex (1 min) followed by sonication with a Branson sonifi er (10 min, 50% duty cycle) on ice to break up large aggregates. The resultant aggregates were passed through a 0.45 µm fi lter to produce small (approximately 30-75 nm) aggregates that induce maximal uptake and lysosomal delivery. Aggregation of LDL and size were confi rmed by negative staining with 2% phosphotungstic acid. Measurement of thiobarbituric acid-reactive substances and conjugated diene levels confi rmed the absence of oxidation after the aggregation procedure ( 38,39 ).

Preparation of lipid dispersions
CE-rich or TG-rich lipid dispersions (DISP) were prepared under sterile conditions as described by Mahlberg et al. ( 26 ). Briefl y, phosphatidylcholine (1 mg), and phosphatidylserine (0.1 mg) and either cholesteryl oleate (30 mg, anisotropic) or triolein (10.35 mg, isotropic) were combined in a sterile 50 ml Corex glass tube and dried under nitrogen. RPMI medium (17 ml) supplemented with HEPES (12.5 mM) was then added, and the suspension was heated in an 80°C water bath for 20 min to melt the dried lipids. The solution was sonicated for 20 min using a Branson sonifi er (50% duty cycle). Thiobarbituric acid-reactive substance levels showed no oxidation after sonication of DISP. Dispersion size was confi rmed by negative staining electron microscopy with 2% phosphotungstic acid (see supplementary Figure I).
Cell culture of THP-1 macrophages THP-1 macrophages were plated onto 35 mm wells or coverslips at a density of 1.5 × 10 6 cells and incubated for 3-4 days at 37°C in RPMI containing 10% FBS and 50 ng/ml phorbol ester (TPA) to allow for differentiation into macrophages. Culture media for all incubations was supplemented with HEPES (20 mmol/l), Eagle's vitamins, L -glutamine (200 mmol/l), streptomycin (100 µg/ml), penicillin (100 IU/ml), and ␤ -mercaptoethanol (0.008 µl/ml). TPA was included in the incubation medium throughout the duration of the experiments. Macrophages were incubated with medium containing 1% FA-free BSA for 24 h before cholesterol loading to minimize excess TG in cells prior to lipid loading.

Lipid loading and analysis
To measure lipid loading, macrophages were incubated for 0-6 days at 37°C in culture medium containing 1% FBS with or without lipid particles, including aggregated LDL (aggLDL), VLDL, CE-rich dispersions, or TG-rich dispersions. In some experiments, only one type particle was employed, while in other experiments, both CE-rich particles (aggLDL or CE dispersions) and TG-rich particles (VLDL or TG dispersions) were used either lesion and could have an impact on foam cell cholesterol metabolism (17)(18)(19)(20)(21). Surprisingly, though, the infl uence of TRP on metabolism of CEs has not been extensively studied. We do know that lesion macrophages, primary cultures of human monocyte-derived macrophages, and ex vivo human foam cell macrophages contain triglyceride (TG) (22)(23)(24)(25). Moreover, TGs are more metabolically active than CE and represent a dynamic lipid pool with the potential to infl uence cellular cholesterol ester metabolism. Thus, defi ning the interaction between TG and intracellular cholesterol pools is critical for a full understanding of atherogenesis.
A key observation linking TRP to CE metabolism is that macrophages hydrolyze CE more effi ciently when it is introduced into lysosomes as a mixed CE and TG particle, compared with CE-containing particles alone ( 26 ). Additionally, it has been shown that TG can alter the physical state of CE by keeping it more fl uid and accessible ( 26 ). This is consistent with studies of cytoplasmic CE droplet metabolism, which demonstrate increased activity of lipolytic enzymes in the presence of mixed CE and TG droplets compared with CE alone ( 27,28 ). Furthermore, FAs generated by lysosomal or extralysosomal hydrolysis of TG are known ligands for, and can upregulate, cholesterol homeostatic genes, including liver X receptors and peroxisome proliferator-activated receptors (29)(30)(31)(32)(33)(34)(35)(36)(37). Therefore, it is clear that increased levels of TG in cells have the potential to affect macrophage cholesterol ester metabolism.
In this study, we investigate the potential for TG to reestablish lysosomal CE hydrolysis and to enhance the mobilization of the resulting FC out of lysosomes. Since CE cannot be cleared from lysosomes, lysosomal CE hydrolysis is the mandatory fi rst step in cellular metabolism of lipoprotein-derived cholesterol and cellular cholesterol use or effl ux. In this report, we show that treatment of macrophages with TRP before, during, or after cholesterol accumulation reduced both lysosomal FC and CE stores and promoted the eventual effl ux of sterol from the cells. The alterations in lysosomal CE metabolism occurred, at least in part, through TG's ability to maintain normal lysosomal activity. Thus, we conclude that modulation of lysosomal CE metabolism, through alterations of cellular TG levels, has profound infl uences on the ability of foam cells to clear cholesterol. TRP can fl ux through the atherosclerotic lesion, and our studies indicate that uptake of these particles by macrophage foam cells can infl uence the ability of foam cells to metabolize the extensive lysosomal CE stores found in late-stage lesions.

Materials
THP-1 human monocytes/macrophages were purchased from ATCC (Manassas, VA). BSA, EDTA, cholesteryl oleate, trioleate, and cholesteryl methyl ether were purchased from Sigma-Aldrich (St. Louis, MO). Phosphatidylcholine and phosphatidylserine were purchased from Avanti Polar Lipids (Alabaster, AL). FBS previously published, vesicles with a pH < 4.8 were classifi ed as active, while those with a pH > 4.8 were considered inactive vesicles ( 45 ). A pH of 4.8 was chosen because the lysosomal acid lipase should have no activity above a pH of 4.8, and this value is above the Pk a of LysoSensor such that there is a signifi cant blue shift in the fl uorescence of the probe. The vesicles in at least 20 cells per condition from three separate experiments were counted.
Negative stain electron microscopy was used to determine the ultrastructural characteristics of the isolated lysosomes and to analyze changes in average lysosome diameter with the various treatments. Lysosomal isolates were absorbed to Formvar-caoted grids for 30 s and then negatively stained for 20 s with 2% phosphotungstic acid, pH 7.0. Digital images were collected using a FEI CM-12 electron microscope operated at 80 keV accelerating voltage and equipped with an AMT cooled CCD camera. To determine the average lysosome diameter, a grid of points was superimposed over the images to select, in an unbiased fashion, the lysosomes for quantifi cation. The diameter of each selected lysosome was computed as the distance between the two most distant points on the lysosome periphery. At least 100 diameters were computed for each condition.
Electron microscopy was also used to assess the distribution of lipid in foam cells between lysosomes and cytoplasmic droplets. Lysosomes and related organelles were identifi ed by the presence of acid phosphatase using a modifi cation of the Gomori lead precipitation method ( 46,47 ). ␤ -Glycerol phosphate was used as a substrate, and the reaction control was incubated in identical medium not containing the enzymatic substrate. After incubation, cells were postfi xed in 1% osmium tetroxide, dehydrated, and embedded in epoxy resin. Before further staining, the sections were viewed to verify the enzymatic reaction and then counterstained with uranyl acetate.

Lysosomal isolation and modifi cation
THP-1 macrophages were treated with various lipid particles, as described above, and then isolated as described previously ( 45 ). Briefl y, the cells were rinsed in cold STE buffer (0.25 M sucrose, 0.01 M Tris-HCl, 1 mM EDTA, and 0.1% ethanol) and scraped into 1 ml/dish of STE buffer containing protease inhibitors (Sigma-Aldrich). The cell suspension was placed in a cell disruption chamber (Kontes) and disrupted using three passes of 20 min each at 150 p.s.i. This method consistently resulted in disruption of >95% of cells but left lysosomes intact. The suspension was centrifuged at 1,500 rpm to separate the postnuclear supernatant from the nuclear pellet. The postnuclear supernatant density was raised to 1.15 g/ml through the addition of sucrose and then applied to a sucrose density gradient ranging from 1.28-1.00 g/ml. The gradient was centrifuged at 19,400 rpm for 4 h at 4°C to separate lysosomal fractions based on their buoyant density. Morphological analysis of the fractions by negative stain electron microscopy (as described above) revealed the presence of a reasonably pure lysosomal population with no apparent structural features of other organelles, including Golgi, rough endoplasmic reticulum, and mitochondria. The purity of the lysosomal populations was assessed further by Western blotting for markers of cellular organelles that might share common morphology to lipid-engorged lysosomes, including early endosomes (EEA1) and lipid droplet associated proteins (perilipin A). Fractions were negative for EEA1 and perilipin and were positive for the lysosomal marker [lysosomal-associated membrane protein (LAMP-1); see supplementary Figure II]. Lysosomal recovery was verifi ed using Western blotting of both the lysosome isolate and the nuclear pellet for LAMP-1 in comparison with whole-cell extracts. LAMP-1 was not detected in the nuclear pellet. The isolated lysosomes were resuspended in buffer containing simultaneously or sequentially. In sequential treatment during pulse-chase experiments, the pulse media was removed after 3 days of treatment, and cells were washed briefl y with 1% FBS prior to addition of the chase media for an additional 3 days. Concentrations and specifi cs of incubation order are described for each experiment. The lipid loading medium was changed every 3-4 days to fresh medium containing the cholesterol or TG loading vehicle.
Loading with aggLDL or VLDL was done using standard culture techniques. In contrast, experiments using DISP were conducted using an inverted culture technique that has been previously described ( 26 ). This method maximizes contact of cells with dispersions and the subsequent internalization of the particles ( 26 ). For the inverted technique, cells were fi rst plated on glass coverslips on the bottom of 35 mm dishes. After adherence and differentiation, the coverslips were inverted and placed on sterile rings. Loading medium was then added so that the coverslip was submerged and the fl oating DISP came into contact with the adherent cells.
For quantitation, cellular lipids were extracted by incubation in 2 ml isopropanol containing 5-10 µg of cholesteryl methyl ether as an internal standard. Lysosomal lipids were extracted from isolated lysosomes using the method of Bligh and Dyer ( 40 ) with 5 µg of cholesteryl methyl ether as an internal standard. The cholesterol content of the lipid extracts were quantifi ed by gasliquid chromatography according to the procedure of Ishikawa et al. ( 41 ) as modifi ed by Klansek et al. ( 42 ). TG content was quantifi ed using a GPO Trinder assay kit from Raichem, according to the manufacturer's instructions. Cellular proteins and proteins in isolated lysosomes were solublized in 1 N NaOH overnight, and protein content was measured using the method of Lowry et al. ( 43 ). Cellular lipid values are reported as micrograms of cholesterol or TG normalized to milligrams of cell protein, while isolated lysosomal lipid values are reported as micrograms of cholesterol or TG normalized to milligrams of lysosomal protein.
Cell viability during loading was assessed by counts of cell number and by protein levels. Experiments were performed in triplicate to assess experimental variability. The mean value for the three measures was used as the value for that experiment for subsequent statistical analysis of multiple experiments.

Microscopy
After lipid loading, microscopy was used to examine the subcellular localization of the accumulated lipids as well as to analyze changes in lysosomal environment. LysoSensor Yellow/Blue DND-160 staining (Molecular Probes, Eugene, OR) was used to determine changes in lysosomal pH ( 44,45 ). This dye fl uoresces yellow in an acidic environment, but the fl uorescence wavelength shifts toward blue as the environment becomes more alkaline. For staining, cells were washed two times in PBS, and the dye was added to cells at a concentration of 5 µM in medium containing 1% FBS. All images were collected within 10 min after the placement of dye on the cells to avoid artifacts produced by the alkaline properties of the dye. As a positive control, macrophages in which active lysosomes were increased by incubation of macrophage with polystyrene beads rather than lipoproteins were stained with LysoSensor Yellow/Blue DND-160. Images were collected using a Zeiss Axioplan Imaging E fl uorescence/brightfi eld microscope (Zeiss, Oberkochen, Germany) equipped with a Photometrics Coolsnap HQ digital camera with a cooled CCD chip (Roper Scientifi c, Tucson, AZ). Image analysis was conducted using MetaMorph imaging software (Universal Imaging, Downingtown, PA). To quantify changes in the number of active lysosomes, a grid of points was superimposed over each image. This provided an unbiased selection of vesicles to evaluate. As This correlated with a signifi cant increase in cellular TG ( Fig. 1B ). Signifi cant TG accumulation was not seen when cells were incubated with aggLDL alone. Thus, the presence of TG, delivered to the cell as a component of VLDL, reduced cholesterol accumulation, specifi cally CE, from aggLDL in THP-1 macrophages.
The simplest explanation for our results would be that TG-rich and CE-rich lipoproteins compete for uptake. In order to determine the extent to which our observations were the result of competition for uptake, THP-1 cells were treated with I 125 labeled aggLDL (50 g aggLDL protein/ml) in the presence or absence of increasing concentrations of VLDL (0, 10, or 50 g VLDL protein/ml) for 150 mM KCl to generate high K + levels inside the lysosome, which provided a membrane potential during the stimulation of the v-ATPase with ATP. Aliquots were obtained for cholesterol, TG, and protein analyses as described above.

Vacuolar-type ATPase activity
Measurement of lysosomal v-ATPase activity was carried out using a modifi cation of a procedure described previously ( 45,48 ). Briefl y, isolated lysosomes were placed in a cuvette containing activation buffer and 6.7 µM acridine orange. After achieving a steady spectrophotometric baseline, v-ATPases were primed with MgCl 2 . After approximately 60 s to allow the baseline to be reestablished, v-ATPases were activated by the addition of ATP (1.4 µM fi nal concentration) and valinomycin (to promote the movement of K + from inside to out for membrane potential generation). v-ATPase-driven pumping of hydrogen ions into the lysosome lumen, as measured by the quenching of acridine orange fl uorescence when excited at 495 nm and recorded at 530 nm, was determined using an SLM Aminco 8100 dual-wavelength spectrophotometer. As a control, lysosomes were activated in the absence of the requisite membrane potential by including valinomycin in the medium.

Statistics
For most analyses, the experiments were repeated at least three separate times. For replicate experiments, a mean, standard deviation, and standard error of the mean were determined. The value used for each separate experiment was the mean value determined from triplicate measures. Two group comparisons used Student's t -test. For multiple comparisons, following an ANOVA, group comparisons were performed with the multiple comparison honestly signifi cant differences method of Tukey ( 49 ). The criterion for signifi cance was set at P < 0.05 for a type I error.

RESULTS
TRPs could potentially infl uence the foam cell metabolism of cholesterol. This is particularly true of the CEs derived from the uptake of CE-rich particles. These CE-containing particles must fi rst be processed within lysosomes. We have previously shown that this CE hydrolysis is inhibited in heavily cholesterol-laden foam cells due primarily to the accumulation of FC within the lysosome lumen and lysosome membrane ( 45,50 ). The lysosomally sequestered free and esterifi ed cholesterol is trapped and cannot be removed from the lysosome even under conditions that promote the removal of nonlysosomal cholesterol from membranes and intracellular droplets ( 16 ). To determine if TRPs affect the lysosomal metabolism of CE, cellular lipid levels were measured in THP-1 macrophages treated with VLDL and aggLDL at the same time. As controls, macrophages were incubated with aggLDL or VLDL alone. Consistent with what we have shown previously, incubation of THP-1 human macrophages with 100 g aggLDL protein/ml produced a dramatic accumulation of both FC and CE ( Fig. 1A ). As in previous studies ( 50 ), thin section electron microscopy of acid phosphatase-stained samples showed that >75% of the lipid volume was within lipid-engorged lysosomes. In contrast, coincubation of THP-1 with both aggLDL and VLDL resulted in a significant ( P < 0.05) reduction the accumulation of CE ( Fig. 1A ).  1. Accumulation of lipids in THP-1 macrophages incubated with aggLDL and/or VLDL. THP-1 macrophages were treated for 6 days at 37°C in RPMI containing 1% FBS and TPA (50 ng/ml) alone or with 100 µg protein/ml of aggLDL and/or VLDL. The cells were harvested, and the cellular lipid levels were determined as described in Materials and Methods. A: Incubation of THP-1 with aggLDL produced a dramatic increase in total cellular cholesterol seen primarily as a signifi cant increase ( P < 0.05) in CE (lightgray portion of bar). Although on average the FC (dark-gray portion of bar) also increased to almost double the control levels, this difference was not statistically signifi cant. Incubation with both aggLDL and VLDL reduced the cellular CE accumulation compared with that seen with aggLDL alone. B: Incubation of cells with VLDL produced a signifi cant ( P < 0.05) increase in cellular TG levels compared with control or aggLDL-treated cells both when used alone or in combination with aggLDL. Values are the mean ± SEM for three experiments. Within each panel, bars with the same letter indicate that means were not statistically different. All other comparisons were signifi cantly different ( P < 0.05).
particles, including oxidized LDL, aggLDL, and CE-DISP is, in large part, the result of inhibition of lysosomal CE hydrolysis in response to an initial accumulation of excess FC within the lysosome ( 45,50,51 ). Lysosomal CE hydrolysis is a key rate-limiting step in the cellular clearance of exogenously derived sterol. Thus, the TG-induced enhancement of cellular cholesterol ester metabolism and loss of sterol from the cell suggest that TG alters lysosomal 48 h (see supplementary Figure III). Results show no difference in particle uptake over a range of VLDL concentrations, indicating that VLDL treatment does not reduce uptake of aggLDL. This suggests that the two particles do not compete for uptake and indicates a TRP-specifi c mechanism by which lysosomal CE metabolism and cellular clearance is enhanced.
Lipoproteins are complex aggregates of lipids and proteins, both of which can infl uence the uptake and cellular metabolism of internalized material. To rule out an effect of lipoprotein-derived protein on the process, protein-free lipid dispersions of phospholipid and CE (CE-DISP, 60 g CE/ml) were substituted for aggLDL, and dispersions of phospholipid and TG (TG-DISP, 50 g TG/ml) replaced VLDL. TG-DISP signifi cantly reduced ( P < 0.05) the CE-DISP-induced accumulation of CE ( Fig. 2A ). As with lipoproteins, this correlated with a signifi cant increase ( P < 0.05) in cellular TG. There was no signifi cant difference in cellular TG levels between incubation with TG-DISP alone and coincubation with TG-DISP and CE-DISP, indicating that uptake of CE-DISP and TG-DISP was not signifi cantly affected by coincubation. Thus, it appears that TG is required for the reduction of cellular cholesterol and is the primary mediator of the effects on cellular sterol metabolism. An even more dramatic reduction in cellular cholesterol concentration was accomplished when cells were fi rst loaded with cholesterol from CE-DISP (60 g CE/ml) and then chased, after removing CE-DISP from the media, with media containing TG-DISP (50 g CE/ml). The loading with CE-DISP more than doubled the cellular cholesterol content, and all of the accumulation was as CE. The chase with TG-DISP produced a 60% reduction in cellular total sterol with all of the reduction occurring as loss of CE.
To further confi rm that competition for uptake was not the explanation for reduced cellular cholesterol and to more completely defi ne the contribution of intracellular TG concentration on cellular lipid levels, we performed pulse-chase experiments with a constant concentration of aggLDL (50 g aggLDL protein/ml) and varying concentrations of VLDL (5, 20, and 50 g VLDL protein/ml). TG treatment reduced cholesterol content in aggLDL-treated THP-1 macrophages in a concentration-dependent manner ( Fig. 3A ). Importantly, the TG-effect was observed when the cells were incubated with VLDL either following ( Fig. 3A ) or prior to ( Fig. 3B ) aggLDL incubation. Moreover, the effect on cellular cholesterol was roughly proportional to cellular TG accumulation ( Fig. 3C, D ). Therefore, TG reduces cellular cholesterol levels in a manner that is dependent on the cellular concentration of TG. Furthermore, TG can prevent subsequent cellular cholesterol accumulation and can mobilize preexisting stores of cellular CE, including the large volume of sterol that accumulates in cellular lysosomes ( 50 ). This is in stark contrast to direct stimulation of extralysosomal cholesterol mobilization, which does not affect the cholesterol trapped within lysosomes ( 16,51 ).
We recently demonstrated that lysosomal accumulation of CE occurring in macrophages incubated with CE-rich lysosomes, but digestion of the particles is inhibited ( 45 ). In contrast to treatment with cholesterol-rich vehicles, cellular TG enrichment results in lysosomes that are much smaller and have relatively homogenous luminal contents, indicative of active lipid particle digestion ( Fig. 4C, D ). Additionally, coincubation of aggLDL and VLDL results in lysosomes that are similar in size and morphology to control lysosomes ( Fig. 4A, D ). The quantitative differences in lysosomal diameter are shown in Fig. 4E . These results are consistent with our hypothesis that cellular TG enrichment enhances lysosomal activity and clearance of lipid from the lysosome.
Changes in the size and morphology of isolated lysosomes suggest enhanced CE hydrolysis and subsequent clearance of lysosomal sterol upon TG enrichment. To confi rm this, we examined the effect of TG specifi cally on lysosomal sterol content. Treatment of cells with TRPs alone or in combination with cholesterol-rich molecules induced an increase in lysosomal TG levels, while treatment with cholesterol-rich molecules alone resulted in increased lysosomal total cholesterol ( Fig. 5 ). However, coincubation of cells with both TRPs and cholesterol-rich cholesterol ester hydrolysis. To study this further, we isolated lysosomes from cholesterol normal cells and cells under our various treatment conditions. Electron microscopy of negatively stained isolates from various cellular subfractions indicates that the lysosomal fraction contained primarily membrane-limited vesicles having the appearance of lysosomes. Furthermore, these fractions were positive for LAMP-1, a marker for lysosomes/late endosomes ( 45,50 ), and lysosomal acid lipase (LAL), the acid CE hydrolase (see supplementary Figure II). Additionally, these fractions were negative for markers of early endosomes (EEA-1) and cytoplasmic lipid droplets (perilipin A) (see supplementary Figure II). Although the lysosomal frac tions from each treatment group contained membranelimited vesicles, differences in lysosome size and morphology were observed in lysosomes isolated from TG-enriched macrophages compared with those treated with cholesterolenriched particles. The presence of small LDL-sized particles within the lysosomes isolated from cells incubated with aggLDL ( Fig. 4B ) is consistent with our previous observations that, under conditions of cholesterol enrichment, CE-rich particles continue to be delivered to the

Fig. 3. Incubation of THP-1 with VLDL decreases cellular FC and CE in a TG dose-dependent manner. A:
In cells preloaded with sterol by incubation with 50 g aggLDL protein/ml, a chase with TG signifi cantly ( P < 0.05) reduced total cholesterol accumulation over a range of VLDL concentrations (5, 20, and 50 g protein/ml) in a dose-dependent manner. B: Preincubation of cells with VLDL also produced a signifi cant ( P < 0.05) and dose-dependent reduction in total cellular cholesterol accumulation from subsequent incubation with 50 g aggLDL protein/ml. C, D: Incubation of THP-1 with increasing concentrations of VLDL produced a stepwise increase in cellular TG levels. Values are the mean ± SEM for three experiments. Within each panel, bars with the same letter indicate that means were not statistically different. All other comparisons were signifi cantly different ( P < 0.05).
only functional at an acidic pH ( 52 ). Thus, the observed changes in lysosomal sterol levels are likely the result of effects on LAL. We do not observe changes in the cellular or lysosomal expression of LAL under the various lipid loading conditions (see supplementary Figure IV). Therefore, it is likely that the effects of TG involve changes in the environment in which the enzyme must function. Previously, we have shown that accumulation of excess FC in molecules resulted in a 33% reduction in lysosomal cholesterol ( Fig. 5 ). As with the analysis on a whole-cell basis, the reduction was almost exclusively in CE. This suggests that the reduction in cellular cholesterol begins with TGinduced changes in lysosomal cholesterol ester hydrolysis.
Conversion of CE to FC is a critical fi rst step in the removal of cholesterol from lysosomes. The hydrolysis of CE is mediated by the enzyme LAL, an acid hydrolase that is Fig. 4. THP-1 macrophages treated with TRPs for 6 days had reduced lysosome diameter compared with macrophages that were only cholesterol enriched. A-D: Negative stain electron micrograph of lysosomes isolated from control cells (A), aggLDL-treated cells (100 g aggLDL protein/ml; B), VLDL-treated cells (100 g VLDL protein/ml; C), or cells coincubated with aggLDL and VLDL (100 g aggLDL protein/ml and 100 g VLDL protein/ml; D). E: The average lysosomal diameter determined from three separate experiments. Lysosomes from control cells were generally small and had a homogenous appearing lumen (A and E). In contrast, lysosomes from aggLDL-treated cells had signifi cantly larger ( P < 0.05) lysosomes that had a variety of appearances but often contained a heterogenous mixture of apparently undigested material (arrows in B) within their lumen (B and E). Lysosomes from VLDL-treated cells were small with homogenous lumens similar to those isolated from control cells (C and E). Signifi cantly, when THP-1 were incubated with VLDL in combination with aggLDL, their lysosomes remained small and failed to develop the large, heterogenous appearance of lysosomes isolated from cells incubated with aggLDL alone. Within each panel, bars with the same letter indicate that means were not statistically different. All other comparisons were signifi cantly different ( P < 0.05). Magnifi cation for A, C, and D = 40,000×; magnifi cation for B = 25,000×; bar = 500 nm. possesses little activity above pH 4.5 ( 52 ). Consistent with our previous studies, the majority of vesicles in untreated macrophages exhibited an active pH ( Fig. 6A, F ). Additionally, the majority of vesicles in cholesterol-enriched macrophages were inactive ( Fig. 6B, F ). However, examination of the vesicle pH in TG-enriched macrophage foam cells revealed differences in pH upon treatment with CErich compared with TG-rich lipoproteins. In contrast to cholesterol-rich foam cells ( Fig. 6B ), macrophages enriched with TG via treatment with VLDL did not display an alteration in the number of active vesicles ( Fig. 6C, F ). In fact, the presence of TG prevented cholesterol-induced lysosome neutralization when TG enrichment occurred, concurrent with aggLDL accumulation ( Fig. 6D, F ). Moreover, when cells were loaded with sterol via aggLDL, subsequent treatment with VLDL reestablished an active pH to the previously inhibited vesicles ( Fig. 6E, F ). To confi rm that most of the vesicles we analyzed were lysosomes, in separate experiments we used LAMP-1 staining to investigate changes in the number of LAMP-1-positive vesicles under our various treatment conditions. Greater than 75% of vesicles in THP-1 foam cells were LAMP-1 positive, and the number of LAMP-1 positive vesicles did not significantly change ( P < 0.05) with treatment. Thus, the shift in percentage of active lysosomes was not the result of changes in lysosome number but rather represents the ability of lysosomes to maintain an active pH. Control cells incubated with polystyrene beads did not show a change in pH, indicating that time duration was not the cause of changes in pH measurement; this was consistent with previous observations ( 45 ). Electron microscopy analysis of acid phosphatase-stained cells also indicates that most of the lipid accumulation was in lysosomes and the number of lysosomes was not appreciably altered with our various treatments.
We have previously shown that cholesterol-induced inhibition of lysosomal acidifi cation occurs primarily through the inhibition of the vacuolar ATPase, an integral membrane protein responsible for pumping H + ions into the lysosomal lumen ( 45 ). Loss of v-ATPase function occurred concurrent with the accumulation of unesterifi ed cholesterol (FC) in the lysosome and, specifi cally, within the lysosomal membrane. Our LysoSensor Yellow/Blue data suggest that TRPs accentuate CE hydrolysis by restoring lysosome pH. We hypothesized that the reduction in pH was due to TG-induced restoration of lysosomal v-ATPase activity. To confi rm this, we examined the activity of the lysosomal v-ATPase in isolated lysosomes from cells treated with TG, cholesterol, or both. Quenching of acridine orange fl uorescence was our measure of the v-ATPase proton pump activity ( 45,48 ). Consistent with our published results ( 45 ), treatment with aggLDL alone (50 g aggLDL/ml, 6 days) resulted in defi cient v-ATPase activity ( Fig. 7 ). However, lysosomes isolated from cells treated with both aggLDL and VLDL (50 g lipoprotein/ml each, 6 days) exhibited a rapid quenching of the acridine orange, comparable to control lysosomes ( Fig. 7 ), indicating that the v-ATPases remained active. This suggests that TRP can help lysosomes maintain an active pH by suppressing foam cell lysosomes leads to an inhibition of CE hydrolysis as a result of failure of the lysosomes to maintain the correct acid pH. Thus, the most probably candidate for the TG reversal of the inhibition of CE hydrolysis would be the reestablishment of the acid pH to the lysosome.
We measured the pH of vesicles within cells under our various treatments using Lysosensor Yellow/Blue DND-160, a pH-sensitive dye ( 45 ). LysoSensor Yellow/Blue DND-160 fl uoresces yellow in an active lysosomal environment and has a signifi cant blue shift as pH approaches neutrality ( 44,53 ). We classifi ed intracellular vesicles as active if the pH was <4.8 and inactive if the pH was >4.8. This pH is at the upper limit of the accepted normal lysosomal pH and is well above the narrow active pH range of human LAL, which displays peak activity at pH 3.8-4.0 and Fig. 5. Accumulation of lipids in lysosomes isolated from THP-1 macrophages incubated with aggLDL and/or VLDL. THP-1 macrophages were treated for 6 days at 37°C in RPMI containing 1% FBS and TPA (50 ng/ml) alone or with 100 µg protein/ml of aggLDL and/or VLDL. The cells were harvested following 6 day lipid accumulation and lysosomes were isolated. The lysosomal lipid levels were determined as described in Materials and Methods. A: Incubation of cells with aggLDL produced a signifi cant increase ( P < 0.05) in the total cellular cholesterol that was seen primarily as increase in CE. In contrast, incubation of cells with VLDL did not produce an increase in lysosomal sterol. Importantly, coincubation of THP-1 with both VLDL and aggLDL limited signifi cantly reduced the amount of CE found in lysosomes. B: Incubation of THP-1 with aggLDL and VLDL together signifi cantly increased ( P < 0.05) the amount of TG accumulating within lysosomes. Values are the mean ± SEM for three experiments.
Our data and that from previous studies suggest the size and metabolic activity of intracellular TG pools can be an important component of macrophage lipid metabolism. Thus, it is important to defi ne the interaction between TG and intracellular cholesterol pools to fully understand the fl ux of lipids within foam cells within atherosclerotic lesions. However, despite some provocative reports ( 28,54 ), the role of TRP in foam cell cholesterol metabolism has not been extensively studied. In this report, we demonstrate that cellular TG greatly infl uences lysosomal cholesterol homeostasis and signifi cantly impacts atherosclerotic macrophage foam cell CE accumulation. Previous research on the interaction of TG-rich and cholesterol-rich lipoproteins has indicated a complex interaction between lipids that can have multiple effects within the macrophage. These potential interactions include the possibility that TG or its metabolites can alter the physical properties of the mixed lipid particles to enhance metabolism, can affect the affi nity and activity of the enzymes responsible for lipolysis, and can regulate cholesterol metabolic metabolism cholesterol's normal ability to suppress the v-ATPases proton pumping.

DISCUSSION
Much of our understanding of atherosclerotic foam cells focuses on the study of cholesterol ester metabolism and accumulation, and most tissue culture models limit themselves to the study of the metabolism of cholesterol-rich particles. However, extracellular areas of atherosclerotic lesions contain a complex milieu with multiple lipid species, each of which could have a distinct infl uence on foam cell biology. In this study, the infl uence of TRPs on macrophage cholesterol ester metabolism was examined. TG was introduced through incubation of cells with VLDL or artifi cial TG-rich dispersions. Treatment with TRPs led to a reduction in total cellular cholesterol, but more importantly, in lysosomal CE macrophage foam cells. This suggests that TG (or a metabolite of TG) is the mediating agent inducing the removal of sequestered sterol from foam cell lysosomes. Fig. 6. LysoSensor Yellow/Blue DND-160 staining of macrophages shows that enrichment of THP-1 macrophages with TG maintains and/or restores active lysosome pH. A: After 6 days in culture, untreated lysosomes maintained an active pH (pH < 4.8), as indicated by a yellow fl uorescence pattern. B: After 6 days of incubation with 100 g aggLDL protein/ml, most lysosomes displayed a predominantly blue fl uorescence, indicating an increase in pH to levels above 4.8. At this pH, CE hydrolysis is inhibited. C: Macrophages treated with 100 g VLDL protein/ml maintained a large population of active lysosomes (pH < 4.8). D: Similarly, when cells were simultaneously incubated with 100 g aggLDL protein/ml and 100 g VLDL protein/ml, the lysosomes maintained activity, as evidenced by predominantly yellow fl uorescence. E: When macrophages were incubated with 100 g VLDL protein/ml after the cells were fi rst incubated with 100 g aggLDL protein/ml, the VLDL treatment was able to restore the pH of the lysosomes back to normal levels. F: Quantifi cation of the percentage of active (pH < 4.8) vesicles following the various treatment conditions. Magnifi cation of A-E = 500×. HDL does not reduce lysosomal cholesterol stores ( 16,56 ). Therefore, TG-containing particles appear to be unique in their ability to promote lysosomal cholesterol clearance.
Our studies also suggest that the effect of TG on lysosomal CE metabolism and cholesterol clearance from the lysosome is more complex than just alterations in enzymatic hydrolysis, as evidenced by the infl uence of TG on lysosomal v-ATPase activity. The lysosomal v-ATPases are critical lysosome integral membrane proteins responsible for pumping H + into the lysosomal lumen to maintain the acidic pH necessary for the function of lysosomal lipolytic enzymes. It is not clear whether TG directly or indirectly affects the pumps. We have previously shown that the activity of v-ATPases is sensitive to the FC content of the lysosome membrane ( 45 ). This is consistent with similar studies that have shown that the macrophage endoplasmic reticulum calcium pump, sarcoplasmic-endoplasmic reticulum calcium ATPase-2b, is also sensitive to changes in membrane fl uidity induced by cholesterol content ( 57 ). Therefore, it is possible that the TG effect is, at least in part, mediated through lysosome membrane changes that enhance proton pumping, lysosomal acidifi cation, and lipolytic enzyme activity. It is also possible that TG, either directly or through its metabolites, directly infl uences the removal of cholesterol from lysosomes and into cholesterol effl ux pathways. In this regard, while the eggression of cholesterol out of lysosomes is not a well-defi ned process, it is known that it primarily involves the intercalation of FC into lysosomal membrane and its subsequent removal via formation of FC-enriched vesicles that traffi c to and fuse with other cellular membranes ( 58 ). Thus, TGmediated effects on lysosomal membrane properties have the potential to infl uence this aspect of lysosomal cholesterol clearance. It remains to be seen how exactly lysosomal TG infl uences lysosomal cholesterol clearance and to what extent membrane alterations are required. However, our study does indicate that the presence of lysosomal TG enhances the metabolism and removal of cholesterol both from the lysosome and, ultimately, from the cell. Further study is required to defi ne the precise mechanism or mechanisms that drive the enhanced clearance. However, our studies to date show that TG can restore lysosome function in foam cells by reestablishing the ability of lysosomes to maintain an acid pH. This leads to increased CE hydrolysis as the initial step in enhanced sterol clearance both from lysosomes and from the cell.
In addition to the potential modulation of the activity of lipolytic enzymes, the metabolic byproducts of TG could affect lysosomal and cellular lipid metabolism. Macrophages have previously been shown to metabolize TG from TRPs to glycerol and FAs through surface hydrolysis and by internalization of TRPs and lysosomal hydrolysis ( 59 ). Thus, TG and its metabolic products would be found in lysosomes. The TG pool is metabolized more effi ciently than CE ( 60 ). This suggests other ways that TG metabolites might impact lysosomal function. For instance, free FAs generated from hydrolysis of TG are known to be infl uential in cholesterol homeostasis. FAs are key signaling and traffi cking (26)(27)(28). For instance, cellular TG can infl uence intracellular increase of the rate and effi ciency of CE hydrolysis in macrophage lysosomes when the CE resides in a mixed lipid pool containing TG with lysosomes ( 26 ). This results, at least in part, from the ability of TG to alter the physical state of CE, keeping it more fl uid ( 26 ). This physical state effect is not limited to lysosomal hydrolysis, as association of TG with cytoplasmic CE droplets has previously been shown to make CE more susceptible to hydrolysis by neutral cholesterol ester hydrolase in the cytoplasm ( 27,28 ). This is important because the mobilization of cytoplasmic and lysosomal CE stores is a key mechanism for cellular cholesterol clearance ( 28 ).
Here, we show that increases in cellular TG drastically reduce cellular and, specifi cally, lysosomal cholesterol levels by maintaining the activity of foam cell lysosomes. Thus, our results are in line with previous studies demonstrating that TG plays a signifi cant role in altering the metabolism of cholesterol in macrophage foam cells. However, our studies are the fi rst to reveal a reduction in lysosomal cholesterol levels in response to increased cellular and lysosomal TG content. This is not an unimportant fi nding. In advanced atherosclerotic lesions, it is estimated that >70% of the lipid in a macrophage foam cell is sequestered in lipid-enriched lysosomes ( 6,55 ). Moreover, our previous studies indicate that the cholesterol within lysosomes is trapped and is extremely resistant to treatments that mobilize and effl ux other intracellular cholesterol stores ( 8,16 ). We and others have shown that enhancement of effl ux promoters, including increases in ATP-binding cassette A1 or the extracellular concentration of free apoproteins or Fig. 7. Activation of v-ATPases in isolated lysosomes following TG and/or cholesterol enrichment. Untreated lysosomes exhibited activation of v-ATPase and the pumping of hydrogen ions into the lysosomal lumen when stimulated by the addition of ATP and valinomycin (time 0), as indicated by the decrease in the relative fl uorescence intensity of acridine orange. Lysosomes from macrophages that had been treated with 100 g aggLDL protein/ml exhibited a lack of v-ATPase activity as evidenced by no reduction in the relative fl uorescence intensity. In contrast, lysosomes from cells treated with 100 g VLDL protein/ml, either alone or in combination with 100 g aggLDL protein/ml, exhibited rapid quenching of acridine orange fl uorescence, indicating active v-ATPases. Data are a representative of example chosen from multiple separate experiments. molecules that greatly affect the expression of critical genes controlling cellular cholesterol mobilization (29)(30)(31)(32)(33)(34)(35)(36)(37). FAs can act at the level of nuclear receptors to affect the transcription of a number of genes important in cholesterol homeostasis (29)(30)(31)(32)(33)(34)(35)(36)(37). In particular, the individual or cooperative upregulation of peroxisome proliferatoractivated receptor and liver X receptor expression by FA has been shown to regulate the expression of a number of cholesterol homeostatic genes, including the ATP-binding cassette gene family members A1 and G1, which enhance cholesterol movement and effl ux ( 37 ). Free FAs might also elicit an effect within the lysosome. Additionally, it is possible that the free FAs generated from the lipolysis of TG within the lysosome may infl uence lysosomal membrane properties either by direct intercalation into the membrane or by infl uencing which acyl chains are present on the lysosomal membrane phospholipids. This modification in the FA composition of the lysosomal membrane might improve membrane fl uidity and enhance the activity of lysosomal integral membrane proteins, including the lysosomal v-ATPase ( 61-64 ). Previously, we have shown that increased membrane cholesterol, which decreases membrane fl uidity, can inhibit lysosomal proton pumping ( 45 ). Combined, these data suggest several explanations for how TG uptake into lysosomes might infl uence macrophage foam cell lysosomal cholesterol ester metabolism. It will take additional experimentation to determine which of these possibilities may actually play a role in the TGinduced enhanced lysosomal CE hydrolysis and clearance.
There is an apparent contradiction between published epidemiologic studies suggesting that hypertriglyceridemia (HTG) may increase atherosclerosis, and our current cellular studies, which indicate a role for TG in cholesterol clearance. Two points are worth noting in this regard. First, the epidemiologic evidence is controversial and does not determine whether HTG has a direct or indirect effect on coronary disease ( 59 ). A better understanding of how HTG affects cells in the artery wall is required to resolve this aspect of the paradox. An important component of that effort is defi ning the role of TRP as a modulator of cholesterol ester metabolism and foam cell biology. Second, it remains to be determined if the increased lysosomal cholesterol clearance induced by TG has a positive or negative impact on macrophage foam cell biology and, ultimately, lesion development. Although in most settings removal of foam cell cholesterol is thought to have a positive impact, the massive removal of cholesterol from previously engorged lysosomes may generate high levels of cellular FC that overwhelm the normal homeostatic mechanisms. In this regard, it is known that high FC levels within certain cellular pools are harmful to macrophages ( 65 ). Thus, cellular health is regulated not only by the levels, but also the cellular location, of cholesterol. As such, FC is essential for proper cellular growth and membrane stability, but excess cellular FC is cytotoxic ( 66,67 ). The sequestration of cholesterol within the lysosomal compartment may be a protective measure to save the cell from the toxic effects of accumulated FC. Further studies are required to sort out this conundrum.