Critical role of neutral cholesteryl ester hydrolase 1 in cholesteryl ester hydrolysis in murine macrophages.

Hydrolysis of intracellular cholesteryl ester (CE) is the rate-limiting step in the efflux of cholesterol from macrophage foam cells. In mouse peritoneal macrophages (MPMs), this process is thought to involve several enzymes: hormone-sensitive lipase (Lipe), carboxylesterase 3 (Ces3), neutral CE hydrolase 1 (Nceh1). However, there is some disagreement over the relative contributions of these enzymes. To solve this problem, we first compared the abilities of several compounds to inhibit the hydrolysis of CE in cells overexpressing Lipe, Ces3, or Nceh1. Cells overexpressing Ces3 had negligible neutral CE hydrolase activity. We next examined the effects of these inhibitors on the hydrolysis of CE and subsequent cholesterol trafficking in MPMs. CE accumulation was increased by a selective inhibitor of Nceh1, paraoxon, and two nonselective inhibitors of Nceh1, (+)-AS115 and (−)-AS115, but not by two Lipe-selective inhibitors, orlistat and 76-0079. Paraoxon inhibited cholesterol efflux to apoA-I or HDL, while 76-0079 did not. These results suggest that Nceh1 plays a dominant role over Lipe in the hydrolysis of CE and subsequent cholesterol efflux in MPMs.

excessive amounts of modifi ed lipoproteins generated during prolonged retention in the arterial walls ( 1 ). Hydrolysis of intracellular CE is the rate-limiting step in the cholesterol effl ux from macrophage foam cells ( 2 ). As the hydrolysis of CE takes place at neutral pH, the enzymes catalyzing it have been collectively called neutral CE hydrolases (NCEHs). Because this step is rate-limiting, particularly in macrophage foam cells ( 3,4 ), it is important to clarify the mechanisms that mediate the hydrolysis of CE in foam cells.
To date, at least three enzymes have been proposed to serve as NCEHs in macrophages. One is hormone-sensitive lipase (LIPE ) ( 5 ). Another is CE hydrolase ( 6 ), which is identical to human liver carboxylesterase 1 (CES1 ) ( 7 ). It is also identical to macrophage serine esterase 1 ( 8 ), also known as a human ortholog of triacylglycerol hydrolase ( 9 ). A third such enzyme is neutral cholesterol ester hydrolase 1 (NCEH1) ( 10 ), which is also known as KIAA1363 or arylacetamide deacetylase-like 1 ( 11 ).
Contradictory results, however, have been reported with regard to the relative contribution of each enzyme to the hydrolysis of CE in macrophages. Lipe is expressed in mouse peritoneal macrophages (MPMs), and its overexpression inhibits the accumulation of CE in macrophages derived from a human acute monocyte leukemia cell line, THP-1 ( 12 ). The reported contributions of Lipe to the hydrolysis of CE in MPMs have varied from negligible ( 13,14 ) to intermediate ( 15 ) to substantial ( 16 ). A mouse ortholog of CES1, carboxylesterase 3 (Ces3), was barely detectable in MPMs ( 10 ) and had negligible NCEH activity ( 17,18 ). In contrast, we found that Nceh1 was robustly Abstract Hydrolysis of intracellular cholesteryl ester (CE) is the rate-limiting step in the effl ux of cholesterol from macrophage foam cells. In mouse peritoneal macrophages (MPMs), this process is thought to involve several enzymes: hormone-sensitive lipase (Lipe), carboxylesterase 3 (Ces3), neutral CE hydrolase 1 (Nceh1). However, there is some disagreement over the relative contributions of these enzymes. To solve this problem, we fi rst compared the abilities of several compounds to inhibit the hydrolysis of CE in cells overexpressing Lipe, Ces3, or Nceh1. Cells overexpressing Ces3 had negligible neutral CE hydrolase activity. We next examined the effects of these inhibitors on the hydrolysis of CE and subsequent cholesterol traffi cking in MPMs. CE accumulation was increased by a selective inhibitor of Nceh1, paraoxon, and two nonselective inhibitors of Nceh1, (+)-AS115 and ( ؊ )-AS115, but not by two Lipe-selective inhibitors, orlistat and 76-0079. Paraoxon inhibited cholesterol effl ux to apoA-I or HDL, while 76-0079 did not. These results suggest that Nceh1 plays a dominant role over Lipe in the hydrolysis of CE and subsequent cholesterol effl ux in MPMs.

Cells
HEK293A cells were cultured in DMEM containing 10% (v/v) FBS and antibiotics. MPMs were obtained 3 days after a 2 ml intraperitoneal injection of 5% (w/v) thioglycollate broth. MPMs were plated on 48-or 96-well plates and cultured in DMEM containing 10% (v/v) FBS and antibiotics for 2 h. Thereafter, cells were washed with PBS, and if not stated otherwise, the adherent macrophages were maintained in DMEM supplemented with 10% (v/v) FBS and antibiotics.

Western blot analyses
Cells were sonicated in buffer A (50 mM Tris-HCl, 250 mM sucrose, 1 mM EDTA, 2 g/ml leupeptin, pH 7.0). Each lysate was separated on a 10% SDS-PAGE and transferred to a nitrocellulose membrane. For detection of the proteins, the membranes were incubated with each anti-murine Nceh1 ( 10 ), anti-murine Ces3 ( 17 ), anti-murine Lipe ( 10 ), or anti-murine GAPDH at a dilution of 1:500-4,000. Specifi cally bound immunoglobulins were detected in a second reaction with a horseradish peroxidase-labeled IgG conjugate and visualized by ECL detection (GE Healthcare) with Image Quant LAS 4000 Mini (GE Healthcare).

Enzymatic assays
Whole cell lysates were prepared from transfected HEK293A and used for the enzymatic assays. PNPB hydrolase activity was determined as described previously ( 10 ). NCEH activity was determined as described by Hajjar et al. ( 28 ), using a reaction mixture containing 6.14 M cholesterol [

Intracellular neutral lipids stained with Oil Red O
After CE turnover assay, MPMs were washed with PBS, fi xed with 4% paraformaldehyde for 30 min, and then stained with Oil Red O and hematoxylin eosin for microscopic analysis (IX70, Olympus).

Cholesterol effl ux assay
Cholesterol effl ux was determined as described previously ( 15,29 ). Briefl y, MPMs (1 × 10 6 cells/well) were loaded with [1,2,6, H(N)]cholesteryl oleate by incubating the cells with 50 g/ml acLDL. After 24 h, cholesterol effl ux was initiated by the addition of 100 g/ml HDL or 25 g/ml apoA-1 in the presence of K-604 and continued for 24 h. An aliquot of the medium was removed and centrifuged at 15,000 g for 2 min to remove cellular debris, and the radioactivity in the supernatant was measured with a liquid scintillation counter. The cells were lysed in 0.05% SDS buffer, and the radioactivity in an aliquot of the cell lysate was measured. The percent effl ux was calculated as (media dpm)/(cell + media dpm) × 100.
To determine which enzyme is more relevant, we used a pharmacological approach, which can be more advantageous, because genetic modifi cation might confound the results by potentially leading to not only unpredictable developmental changes but also compensatory regulation of other genes.
We selected six inhibitors, four of which have been reported to have inhibitory activity toward either Nceh1 or Lipe. Cravatt and his colleagues have previously reported that phosphatase activity of KIAA1363 (NCEH1) was inhibited by paraoxon ( 11 ) or AS115 ( 20,21 ). We confi rmed the inhibitory activity of AS115 on NCEH activity of NCHE1 ( 18 ), and 76-0079 was originally developed as a selective inhibitor of Lipe ( 18,22 ). Benzil inhibits CES1 ( 23 ), and orlistat inhibits pancreatic lipase ( 24 ).

Preparation of lipoproteins
After an overnight fast, blood was collected from normolipidemic volunteers to isolate plasma. LDL (d 1.019-1.063 g/ml) and HDL (d 1.063-1.21 g/ml) were isolated from the plasma by sequential density ultracentrifugation ( 26 ). LDL was acetylated by repeatedly adding acetic anhydride ( 27 ).

Mice
All mice [C57BL/6J (WT), Nceh1 knockout ( Nceh1 ( 15 ), and Lipe knockout ( Lipe ( 14,15 ) mice] were maintained in a temperature-controlled (25°C) facility with a 12 h light/dark cycle and given free access to food and water. Mice were maintained and cared for according to the regulations of the Animal Care Committees of Jichi Medical University. All animals used in these studies were male. enzymes caused substantial increases in PNPB hydrolase activity (Ad-Nceh1, 28.1-fold; Ad-Ces3, 26.5-fold; Ad-Lipe, 15.3-fold) (supplementary Fig. IB). NCEH activity was increased 24.9-fold by overexpression of Lipe and was increased 4.4-fold by overexpression of Nceh1, but it was not increased by overexpression of Ces3 (supplementary Fig. IC). Therefore, we used only Ad-Nceh1 and Ad-Lipe for further studies.

Selectivity of compounds against NCEH enzymes
We compared the inhibitory effects of each compound on NCEH enzymatic activities, which were expressed by overexpression of Nceh1 or Lipe in cell lysates. The IC 50 values and selectivity index (SI) values are summarized in Table 1

CE turnover in MPMs
To examine whether inhibition of the hydrolysis of CE by each compound affected cholesterol traffi cking, we labeled MPMs with oleic acid and measured the amounts of cholesteryl oleate after exposure to each compound. Treatment with (+)-AS115 or ( Ϫ )-AS115, which are nonselective inhibitors, and with paraoxon, which is an Nceh1-selective inhibitor, increased CE in MPMs from WT mice ( Fig. 1A-C ). Similar increase in CE was observed in MPMs from Lipe Ϫ / Ϫ mice. However, these CEincreasing effects were not observed in MPMs from Nceh1 Ϫ / Ϫ mice. These results indicate that Nceh1 is critical for hydrolyzing CE and removal of cholesterol from the cell. On the other hand, treatment with orlistat or 76-0079, which are Lipe-selective inhibitors, did not signifi cantly increase the CE contents in MPMs from any of three types of mice ( Fig. 1D, E ). Moreover, treatment with benzil, a Ces3-selective inhibitor, did not significantly increase the CE content either ( Fig. 1F ). These results indicate that neither Lipe nor Ces3 is signifi cantly involved in the process.
To rule out the possibility that the compounds affect cholesterol traffi cking via their cytotoxicity, we measured Chemical (Ann Arbor, MI). Assay was performed following the manufacturer's protocol. Briefl y, MPMs (5 × 10 4 cell/well) were incubated in DMEM containing 5 mg/ml BSA with each compound for 24 h. Four hours after the addition of Dye solution, Solubilization/Stop solution was added to the medium for measurement of absorbance using a spectrometer (E Max, Molecular Devices).

Quantitative real-time PCR
Total RNA was prepared from MPMs using TRIzol. Relative amounts of mRNA were calculated using a standard curve or the comparative cycle threshold method with the StepOnePlus Real-Time PCR instrument (Applied Biosystems) according to the manufacturer's protocol. Mouse ␤ -actin ( Actb ) mRNA was used as the invariant control. Primer sequences were as follows: Nceh1 forward,

Statistical analyses
Results are presented as the mean ± SD. Statistical differences between groups were analyzed by one-way ANOVA and the Dunnett's multiple comparisons test. All calculations were performed with Graph Pad Prism version 6.0 for Macintosh (MDF).

NCEH activity in the cells infected with Ad-Nceh1, Ad-Ces3, and Ad-Lipe
To confi rm the ability of the overexpressed enzymes to hydrolyze CE, we infected HEK293A cells with recombinant adenoviruses to overexpress Nceh1, Ces3, or Lipe. Whole cell lysates were subjected to Western blot analyses and measurements of enzymatic activities (supplementary   MTT activities in the cells treated with the compounds. These compounds did not show cytotoxicity against MPMs in an MTT assay (supplementary Fig. II). To examine whether the compounds affect the expression of each enzyme, we performed RT-PCR analysis of Nceh1 and Lipe . Paraoxon, orlistat, and 76-0079 did not affect the expression of Nceh1 and Lipe (supplementary Fig.  III). On the other hand, (+)-AS115 decreased the expression of Lipe , and ( Ϫ )-AS115 decreased the expression of Nceh1 and Lipe at 25.6 M. However, the effects were not signifi cant at the lower concentrations (0.3 and 2.6 M). Therefore, (+)-AS115 and ( Ϫ )-AS115 inhibited both the hydrolase activity and the expression of Nceh1 and Lipe (supplementary Fig. III). Conceivably, the inhibition of the expression of Nceh1 and Lipe did not mediate the CE-increasing effects of AS115s at the lower concentrations.

Lipid droplet accumulation in MPMs
After loading the cells with CE by incubation with acLDL, intracellular neutral lipid droplets were stained with Oil Red O. Treatment with (+)-AS115, ( Ϫ )-AS115, or paraoxon increased lipid droplet accumulation compared with the control ( Fig. 2A-D ). On the other hand, neither orlistat, 76-0079, nor benzil caused signifi cant lipid droplet accumulation ( Fig. 2E-G ). These results suggest that selective inhibition of Nceh1, but not Lipe or Ces3, increased CE accumulation in MPMs.

Cholesterol effl ux in MPMs
To directly investigate whether the inhibition of CE hydrolysis decreases the release of free cholesterol from the cell, we measured cholesterol effl ux in MPMs treated with paraoxon, an Nceh1-selective inhibitor, and 76-0079, a Lipe-selective inhibitor in the presence of K-604 to inhibit de novo esterifi cation of cholesterol ( Fig. 3 ). When HDL or apoA-1 was used as a cholesterol acceptor, only treatment with paraoxon decreased cholesterol effl ux from MPMs of WT mice ( Fig. 3A, B ). Similar decrease was observed in MPMs from Lipe Ϫ / Ϫ mice. However, these effects of paraoxon were not observed in MPMs from Nceh1 Ϫ / Ϫ mice ( Fig. 3A, B ). To examine whether the changes in cholesterol effl ux are associated with changes in the expression of Abca1 and Abcg1 , we measured the expressions of these genes by RT-PCR. While TO-901317, a liver X receptor agonist, increased the expressions of Abca1 and Abcg1 , neither paraoxon nor 76-0079 affected them signifi cantly (supplementary Fig. IVA, B). K-604 (19.7 M), an ACAT1 inhibitor, did not inhibit Nceh1 and Lipe in NCEH assay (supplementary Fig. V). These results indicate that Nceh1 is primarily involved in CE hydrolysis and that it is the rate-limiting step in the cholesterol effl ux from MPMs.

DISCUSSION
Based on their selectivities on Nceh1 or Lipe, the six inhibitors were classifi ed into four groups: (1) nonselective inhibitors [(+)-AS115 and ( Ϫ )-AS115], (2) Nceh1-selective inhibitor (paraoxon), (3) Lipe-selective inhibitors (orlistat and 76-0079), and (4) inhibitor of PNPB-hydrolyzing activity of Ces3 (Benzil). Treatment with paraoxon, an Nceh1-selective inhibitor, increased CE accumulation, as shown by the accumulation of neutral lipid droplets, in MPMs ( Figs. 1C, 2D ). Moreover, paraoxon decreased cholesterol effl ux from MPMs without changing the expression of Abca1 or Abcg1 (supplementary Fig. IVA, B). Similar effects were observed in MPMs from Lipe Ϫ / Ϫ mice. In contrast, these effects were not detectable in MPMs from Nceh1 Ϫ / Ϫ mice. These results indicate that Nceh1 substantially contributes to the NCEH and subsequent cholesterol effl ux in MPMs, which is in good agreement with our previous reports ( 15,18 ). On the other hand, orlistat and 76-0079 (Lipe-selective inhibitors) did not signifi cantly increase the CE contents ( Fig. 1D, E ) or lipid droplets in MPMs ( Fig. 2E, F ). It is possible that orlistat is not transported to the intracellular sites where Lipe is localized. In contrast, 76-0079 has been widely used for cell-based experiments showing its effi cacy ( 18,22 ). Thus, we conclude that Lipe does not contribute to the NCEH activity in MPMs. These observations are very consistent with the results that AS115s (nonselective inhibitors) increased CE accumulation as much as paraoxon did. Based on these results, we can conclude that Nceh1 plays a dominant role over Lipe in the hydrolysis of CE in MPMs.
In contrast to these current and previous observations that Nceh1 signifi cantly contributes to CE hydrolysis in MPMs, Buchebner et al. ( 16 ) proposed that Lipe, but not Nceh1, is essential for the hydrolysis of CE in MPMs. It is unclear why they reached opposite conclusions even though they used a similar strategy: use of MPMs obtained from genetically modifi ed mice. We assume that the completeness of the defi ciency of Nceh1/KIAA1363 might be different between the mice used in the two labs because of the use of different targeting vectors. Furthermore, sensitivity of the assay to measure NCEH activity might have complicated the results.
The negligible contribution of Lipe to the hydrolysis of CE in MPMs is very consistent with the results of two earlier studies on MPMs from Lipe Ϫ / Ϫ mice ( 13,14 ). Buchebner et al., however, reported that NCEH activity was almost abrogated in MPMs from Lipe Ϫ / Ϫ mice. We do not know the reason for this discrepancy. Because Lipe defi ciency leads to developmental changes in several tissues such as testis and adipose tissue ( 14 ), it is possible that Lipe deficiency also causes developmental changes in the macrophage lineage in an age-dependent manner. If Lipe defi ciency somehow decreases the NCEH activity of Nceh1 in macrophages under certain conditions, this may explain the contradiction stated previously.
The compounds used in the present study were not strictly specifi c. Paraoxon and/or AS115, the Nceh1-inhibiting compounds, also inhibit PNPB-hydrolyzing activity of Ces3 ( Table 1 ). Furthermore, it is well known that organophosphorus toxicants, to which paraoxon belongs, target at least 50 serine hydrolases and receptors including acetylcholinesterase, butyrylcholinesterase, chymotrypsin, arylformamidase, and fatty acid amide hydrolase ( 30 ). However, overexpression of Ces3 did not show a signifi cant NCEH activity (supplementary Fig. IC) ( 17 ). Moreover, benzil, a Ces3-selective inhibitor, did not inhibit the hydrolysis of CE in MPMs ( Figs. 1F, 2G ). Therefore, it is unlikely that paraoxon or AS115 increased CE accumulation by specifi cally inhibiting Ces3. Recently, Marcel and his colleagues ( 31,32 ) proposed a novel and intriguing pathway for CE hydrolysis: autophagy. In one of the key experiments, they used chloroquine to disrupt the lysosomal pathway. Treatment with chloroquine increased cellular CE as much as treatment with paraoxon did. These results were interpreted as evidence of the involvement of autophagy in the hydrolysis of CE. However, a high concentration of chloroquine can be cytotoxic. Indeed, we found that incubation of MPMs with 30 and 100 µM chloroquine for 24 h decreased MTT activity by 60% and 90%, respectively (unpublished observations). The resulting dying cells might be taken up by neighboring macrophages by efferocytosis, where cellular CE is directly targeted to lysosomes via fusion with phagosomes. This pathway involves lysosomes, but certainly not autophagy. Another caveat concerning the use of chloroquine is its potential effect on ataxia telangiectasia mutated (ATM). Schneider et al. ( 33 ) reported that treatment with low-dose chloroquine attenuated atherosclerosis in apoE knockout mice by suppressing c-Jun N-terminal kinase activity, which suppresses LPL activity via activating ATM. In a pioneering paper addressing the lysosomal pathway for CE hydrolysis, Avart and his colleagues ( 34 ) showed that chloroquine inhibited the hydrolysis of CE only when CE is in anisotropic inclusions. Further studies are needed to correctly interpret the antiatherosclerotic effect of chloroquine.
In conclusion, we show pharmacological evidence that Nceh1 has a critical role in the hydrolysis of CE in MPMs. These fi ndings should provide the basis for understanding the pathophysiology of atherosclerosis and can be exploited to develop new therapeutic approaches.