Identification of a novel intracellular cholesteryl ester hydrolase (carboxylesterase 3) in human macrophages: compensatory increase in its expression after carboxylesterase 1 silencing

Abstract

Cholesteryl ester (CE) hydrolysis is the rate-limiting step in the removal of free cholesterol (FC) from macrophage foam cells, and several enzymes have been identified as intracellular CE hydrolases in human macrophages. We have previously reported the antiatherogenic role of a carboxylesterase [carboxylesterase 1 (CES1)], and the objective of the present study was to determine the contribution of CES1 to total CE hydrolytic activity in human macrophages. Two approaches, namely, immune depletion and short hairpin (sh)RNA-mediated knockdown, were used. Immuneprecipitation by a CES1-specific antibody resulted in a 70–80% decrease in enzyme activity, indicating that CES1 is responsible for >70% of the total CE hydrolytic activity. THP1-shRNA cells were generated by stably transfecting human THP1 cells with four different CES1-specific shRNA vectors. Despite a significant (>90%) reduction in CES1 expression both at the mRNA and protein levels, CES1 knockdown neither decreased intracellular CE hydrolysis nor decreased FC efflux. Examination of the underlying mechanisms for the observed lack of effects of CES1 knockdown revealed a compensatory increase in the expression of a novel CES, CES3, which is only expressed at <30% of the level of CES1 in human macrophages. Transient overexpression of CES3 led to an increase in CE hydrolytic activity, mobilization of intracellular lipid droplets, and a reduction in cellular CE content, establishing CES3 as a bona fide CE hydrolase. This study provides the first evidence of functional compensation whereby increased expression of CES3 restores intracellular CE hydrolytic activity and FC efflux in CES1-deficient cells. Furthermore, these data support the concept that intracellular CE hydrolysis is a multienzyme process.

hydrolysis of intracellular cholesteryl esters (CEs) in artery wall-associated macrophage foam cells is the obligatory first step in the mobilization of CE and extracellular acceptor-mediated removal of the resulting free cholesterol (FC). In association with HDL, FC is transported to the liver for the final elimination from the body as biliary cholesterol or bile acids. The significance of this process of reverse cholesterol transport in the regulation of atherosclerotic plaque formation is well established, and, as the rate-limiting first step of reverse cholesterol transport, the importance of intracellular CE hydrolysis cannot be overemphasized. While low intracellular CE hydrolysis is linked to susceptibility to atherosclerosis (15, 23), an increase in hydrolysis by overexpression of a human CE hydrolase [CEH; carboxylesterase 1 (CES1)] leads to an increase in cellular CE mobilization (5), FC efflux (26), and attenuation of diet-induced atherosclerosis (25).

Several enzymes have been identified as intracellular CEHs, and whether any one of them should be considered as the major CEH is currently controversial, mainly due to the incomplete loss of activity after single gene deficiency. Gene ablation of the first candidate CEH, namely, hormone-sensitive lipase (HSL; gene: Lipe), was initially reported to have no effect on CE hydrolysis in macrophages (16), and it was concluded that HSL does not play a role in intracellular CE hydrolysis (2). Although Buchebner et al. (1) recently demonstrated a significant decrease in CE hydrolytic activity in macrophages from HSL-deficient macrophages, the observed lack of cellular CE turnover led the authors to conclude that an in vitro assay of this enzyme may not truly represent the situation in vivo. Despite the complete lack of mRNA and protein expression of the recently identified CEH KIAA1363 [also known as neutral CEH 1 (NCEH1)], macrophages from homozygous KIAA1363^−/− mice showed only a 40–50% reduction in total CE hydrolytic activity (no change in activity associated with the cytosolic fraction and a ∼30% decrease in endoplasmic reticulum-associated activity) and a corresponding 40% reduction in FC efflux (21). The recent characterization of a different line of KIAA1363 knockout mice generated earlier showed no change in CE hydrolysis in macrophages (1). Triglyceride lipase (TGH; gene: Ces3) is the murine ortholog of human CES1, and effect of its deficiency on macrophage CE hydrolysis has not been examined as yet. However, ablation of TGH expression resulted in only a 30% reduction in enzyme activity in the liver (22). Collectively, these studies demonstrate that single gene deficiency is insufficient to completely abolish intracellular CE hydrolysis and allude to the contribution of multiple enzymes to total CE hydrolysis. However, the mechanisms underlying the observed discordance between expression levels and activity were neither examined nor addressed in any of these studies.

We have previously reported on the cloning and characterization of human macrophage CEH (6) and established its role in intracellular CE mobilization (5), FC efflux (26), and attenuation of diet-induced atherosclerosis (5). Li and Hui (14) demonstrated the expression of carboxyl ester lipase in THP1 and primary human macrophages and attributed a limited role for this enzyme in intracellular CE metabolism due to its secretory nature. The expression of HSL in human monocyte/macrophages remains controversial (18, 2). Igarashi et al. (10) recently demonstrated the presence of KIAA1363 in human macrophages. However, the relative contribution of each of these enzymes to total CE hydrolysis has not been evaluated. In the present study, we used two approaches, namely, immune depletion as well as short hairpin (sh)RNA-mediated knockdown in human THP1 cells, to determine the contribution of CES1 to total CE hydrolysis in human macrophages. Our results demonstrate that CES1 accounts for >70% of the total CE hydrolytic activity; <30% activity remained after CES1 immune depletion. However, despite the significant shRNA-mediated knockdown of CES1, there was no significant decrease in cellular CE hydrolysis and FC efflux, and the underlying mechanism was a compensatory increase in the expression of CES3. The cloning and characterization of a novel variant of CES3 expressed in macrophages are described.

EXPERIMENTAL PROCEDURES

Immunoprecipitation.

CES1 antibody (AF4920, R&D) was cross-linked to Dynabeads protein A (Invitrogen) according to the manufacturer's protocol. THP1 cells were sonicated using a Branson sonifier fitted with a microtip, as previously described (4), and cell lysates were incubated with Dynabeads-antibody complex overnight at 4°C. Dynabeads-antibody-antigen complexes were removed by centrifugation, and the postimmunoprecipitation (post-IP) supernatant was analyzed for the presence of unprecipitated residual CES1 protein. The ratio of antigen (THP1 cell lysates) and CES1 antibody conjugated to Dynabeads was optimized such that after IP no immunoreactive protein was detected in the post-IP supernatant by Western blot analysis (data not shown). Optimized conditions (5 μg antibody and 250 μg THP1 lysate protein) were used to immunoprecipitate CES1, and the post-IP supernatant as well as cell lysates before IP (pre-IP) were used to determine CE hydrolytic activity.

Enzyme assays.

CE as well as triacylglycerol hydrolase activities were determined using cholesteryl [1-¹⁴C]oleate or [³H]triolein as substrates, respectively. Substrates were presented either as micelles containing phosphatidyl choline and taurocholate (8) or as acetone droplets (4, 6), and, under the assay conditions used, the activities of all CE hydrolases were measured. In some experiments, cell extracts were incubated with 100 μM benzil (a specific CES inhibitor) before the addition of the substrate (3).

Generation of THP1 cells stably expressing CES1-specific shRNA.

A set of four shRNA vectors specific for human CES1 (p81–p84) as well as a control vector (pRSL) were purchased from Origene. THP1 monocytes (10⁶ cells) were transfected with 0.5 μg vector DNA using Amaxa nucleofection kit V5 and program U-01. Stably transfected cells were selected for puromycin (2 μg/ml) resistance. Since THP1 monocytes grow in a suspension culture precluding clonal selection, the stably transfected cells used in this study represent a mixture of all stable clones.

Isolation of human blood-derived monocytes.

BD Vacutainer CPT cell preparation tubes with sodium citrate were used to isolate total leukocytes from freshly collected blood. Isolated cells were either used to extract total RNA (day 0) or cultured in complete growth medium (RPMI-1640 supplemented with 10% FBS) containing 15 ng/ml macrophage colony-stimulating factor (10). The medium was replaced every 2 days, and total RNA was extracted on days 3 and 5.

Cholesterol efflux from macrophages.

THP1 monocytes stably transfected with control vector (pRSL) or with CES1-specific shRNA vectors (p81–84) were differentiated into macrophages by the addition of 100 nM PMA to the growth medium. Differentiated macrophages were loaded with CE using acetylated LDL (50 μg/ml) and labeled with [³H]cholesterol. Total FC efflux to HDL (25 μg/ml) or 10% FBS was monitored as previously described (26).

Cloning of CES3.

The following oligonucleotide primers were designed based on the sequence of human CES3 (Accession No. NM_024922): sense primer 5′-GAGATCGGTGGTGCTGAAGGGCAGGG-3′ (−73 to −48), nested sense primer 5′-CTTATTCCACCTTCTGAAGCTTCTGTCG-3′ (−45 to −18), antisense primer 5′-CCACCACTTGAAGAGTGGTTTGCC-3′ (1743–1766), and nested antisense primer 5′-GAGGTCCTCCTGGGCCTTCC-3′ (1693–1713).

A c-myc tag was attached to the nested antisense primer to facilitate production of a c-myc-tagged protein. Two-step nested PCR was performed using cDNA obtained from human THP1 macrophage RNA. The PCR product obtained with nested primers was directly cloned into the pcDNA3.1-TOPO vector, where expression is under the control of a cytomegalovirus (CMV) promoter. Plasmid DNA from three different colonies containing the insert was sequenced in both directions. The recombinant plasmid was labeled pCMV-CES3, and the vector without an insert was labeled pCMV.

Transfection of COS-7 cells.

Optimized conditions, as previously described (6), were used to transfect COS-7 cells with the control vector (pCMV) or the CES3 expression vector pCMV-CES3. In some experiments, cells were also transfected with the CES1 (CEH) expression vector pCMV-CEH. Cells were harvested, and total cell homogenates were used to determine enzyme activity as described above and protein expression by Western blot analyses.

CE mobilization from agmACAT1 cells.

Constitutive expression of acyl CoA-cholesterol acyltransferase 1 (ACAT1) in agmACAT1 cells leads to the visible accumulation of cytoplasmic lipid droplets in these cells (13) and provides a suitable model system to monitor CE mobilization by transient overexpression of a CEH, as previously described (5). agmACAT1 cells were maintained in Ham's F-12 medium supplemented with 10% FBS, penicillin-streptomycin, and 200 μg/ml geneticin. These cells were transiently transfected with the control vector pCMV or the CES3 expression vector pCMV-CES3 using optimized conditions (5). The growth medium containing 10% FBS was replaced twice (24 and 48 h after transfection), and analyses were performed 72 h after transfection. In one set of experiments, cellular lipid droplets were stained with oil red O, and, after an extensive wash (≥3), cells were imaged using an Olympus inverted microscope fitted with a digital camera operated with AxioVision software. In another set of experiments, at the end of 72 h, cells were washed three times with PBS, and total lipids were extracted with isopropanol. Total cholesterol and FC were determined by gas chromotography, and esterified cholesterol was determined as the difference between total cholesterol and FC; where needed, this value was multiplied by 1.67 to convert to CE mass. After lipid extraction, cellular proteins were solubilized in 1 N NaOH and quantified using the Pierce BCA kit.

Western blot analyses.

Proteins (20 μg) were separated by 10% SDS-PAGE (Bio-Rad Laboratories), transferred to polyvinylidene difluoride membranes, and immunoblotted with primary antibodies followed by species-specific fluorescently labeled secondary antibodies (LI-COR). Positive immunoreactivity was detected by scanning in the appropriate channels by an Odyssey infrared imaging system (LI-COR) and quantified by densitometry using Quantity One 4.4.0 software (Bio-Rad Laboratories).

Real-time PCR.

Total RNA was extracted using the RNeasy kit (Qiagen). cDNA was synthesized using a high-capacity cDNA reverse transcription kit (Applied Biosystems). Real-time PCR was performed using a Stratagene Mx3000P machine using the TaqMan Universal PCR Master Mix and optimized probe and primer sets from Applied Biosystems. The following probes were used: human CES1 (Hs00275607_m1), CES3 (Hs00227775_m1), KIAA1363 (Hs00736941_m1), and CD68 (Hs02836816_g1). Data were normalized to β-actin. The CES1 copy number was determined using a standard curve as previosly described (24).

RESULTS

CES1 accounts for >70% of the total intracellular CE hydrolase activity.

The CES1 antibody used in this study was initially tested for cross-reactivity and shown to react only with human CES1 and its mouse ortholog Ces3. No cross-reactivity was noted with recombinant CES2 or CES5 (data not shown). Furthermore, in human macrophages, a single immunoreactive band corresponding to a molecular weight of ∼60 kDa was obtained, indicating a lack of reactivity to KIAA1363 or NCEH1 with a molecular weight of 45–50 kDa. To determine the contribution of CES1 to total cellular CE hydrolytic activity, CES1 protein was immunoprecipitated from THP1 cell lysates. As shown in Fig. 1A, no immunoreactive CES1 protein was detected after IP, demonstrating successful immune depletion of CES1. Pre- and post-IP fractions were assayed for CE hydrolytic activity, and the data are shown in Fig. 1B. Specific CES1 depletion reduced the CE hydrolytic activity by >70%, indicating that in human THP1 macrophages, CES1 protein accounts for >70% of the total intracellular CE hydrolysis.

Fig. 1.

Immunoprecipitation (IP) of carboxylesterase 1 [CES1; cholesterol ester (CE) hydrolase (CEH)]. Cell lysates (pre-IP) were incubated with CES1 antibody-conjugated Dynabeads for 24 h at 4°C. Dynabead-antibody-antigen complexes were removed, and the post-IP supernatant was collected. Twenty micrograms of protein from pre-IP and post-IP fractions were analyzed by Western blot analyses. A: representative Western blot. Lanes 1 and 3 are pre-IP; lanes 2 and 4 are post-IP. CES1- and β-actin-immunoreactive bands are marked. B: CEH activity was determined as described in experimental procedures. Data are presented as means ± SD for n = 3, and each assay was performed using three protein concentrations in duplicate. The P value indicating a statistically significant decrease in CE hydrolytic activity in the post-IP fraction is shown.

Significant knockdown of CES1 in THP1-shRNA cell lines.

Four different THP1-shRNA cell lines were generated in addition to a control line stably transfected with the empty parent vector pRSL. CES1 mRNA expression was significantly attenuated in all four lines, namely, p81, p82, p83, and p84; expression of CES1 mRNA was only reduced by ∼50% in p84 (Fig. 2A). There was no significant difference in the expression of macrophage marker CD68 (data not shown). Consistently, CES1 protein expression was reduced by >90% in p81, p82, and p83, and only a 30% decrease was observed in p84. These data confirm the development of several THP1-shRNA cell lines with significantly reduced expression of CES1 at both RNA and protein levels.

Fig. 2.

Expression of CES1 (CEH) mRNA and protein levels in THP1-short hairpin (sh)RNA cell lines. THP1-shRNA cell lines (p81–p84) were generated and differentiated into macrophages as described in experimental procedures. Total RNA as well as total protein extracts were prepared, and CES1 mRNA and protein expression were determined by real-time RT-PCR and Western blot analyses, respectively. A: CES1 copy number was determined using a standard curve. Data are presented as means ± SD for n = 3, and each determination was performed in duplicate. B: representative Western blot (top) and densitometric quantification (bottom). CES1 levels were normalized to β-actin, and protein expression was plotted as a percentage of the control vector (pRSL). *P < 0.05.

Unaltered FC efflux despite attenuated CES1 expression in THP1-shRNA cell lines.

FC, generated by CES1-mediated hydrolysis of intracellular CE, is effluxed by multiple pathways dependent on the nature of the extracellular cholesterol acceptors, and total FC efflux to HDL or 10% FBS was monitored. As shown in Fig. 3, efflux of FC from the different THP1-shRNA cell lines (p81–p84) to HDL (A) or FBS (B) was not significantly different from that observed with control THP1-pRSL cells. In addition, a longer incubation time (24 vs. 4 h) increased FC efflux in all cell lines, and, consistent with cholesterol acceptor efficiency, higher efflux was seen with FBS compared with HDL. These data suggest that despite the significant decrease in CES1 expression, there was no change in FC efflux from these cells.

Fig. 3.

Free cholesterol (FC) efflux from THP1-shRNA cell lines. THP1-shRNA cell lines were differentiated into macrophages and loaded with lipids using acetylated LDL as described in experimental procedures. Cellular cholesterol pools were also radiolabeled by including [³H]cholesterol in the loading medium. After a 24-h equilibration, FC efflux was initiated by the addition of HDL (25 μg/ml; A) or 10% FBS (B) and monitored by determining the radioactivity associated with the culture medium at 4 and 24 h. Cells were lysed at 24 h, and total cellular radioactivity was determined. The percent FC efflux was then calculated. Data are presented as mean ± SD for n = 3, and each determination was performed in duplicate.

Intracellular CE hydrolytic activity is not significantly affected by knockdown of CES1 expression.

Intracellular CE hydrolysis is regarded as a rate-limiting step in cholesterol efflux and the process of reverse cholesterol transport. To explore the mechanisms underlying the observed unchanged FC efflux in THP1-shRNA cells with significantly reduced CES1 expression, total CE hydrolytic activity was determined. CE was presented as micelles as well as droplets, and CE hydrolytic activity in different THP1-shRNA cell lines (p81–p84) was not significantly different from that observed with control THP1-pRSL cells (Fig. 4). Similar activity was observed regardless of the physical state of substrate presentation. Intracellular CEHs have broad substrate specificities, and triolein (triglyceride) presented as micelles was also used to determine enzyme activity. Similar results were obtained when triolein was used as the substrate; there was no change in triolein hydrolytic activity with shRNA-mediated knockdown of CES1 in the different THP1-shRNA cell lines. CE and triolein hydrolytic activities were, however, significantly reduced when THP1 cell extracts were treated with a CES-specific inhibitor (benzil); only 33.12 ± 7.21% and 51.19 ± 3.05% activity toward CE and triolein, respectively, was observed in the presence of benzil.

Fig. 4.

Intracellular CE hydrolytic activity. Total cell lysates were prepared from THP1-shRNA macrophages and used to determine hydrolytic activity using CE (as micelle or droplets) and triolein [triglyceride (TG)] as substrates. Data were normalized to the activity obtained with THP1-pRSL cell lysates for the respective substrates and are expressed as means ± SD for n = 3; each determination was performed in duplicate. To ensure linearity, three different protein concentrations were used for enzyme assays. In the presence of the specific CES inhibitor benzil, only 33.12 ± 7.21% and 51.19 ± 3.05% activity was observed with CE and TG substrates, respectively.

Compensatory increase in the expression of CES3 in THP1-shRNA cells.

The unchanged FC efflux and CE hydrolytic activity in the face of an almost 90% decrease in CES1 expression, which accounts for >70% of the total CE hydrolytic activity in THP1 cells (Fig. 1), suggests that additional enzyme(s) may contribute to total intracellular CE hydrolysis. HSL and KIAA1363 (NCEH1) are two other enzymes that are identified as macrophage CEHs, and while no expression of HSL was detected, KIAA1363 was expressed in THP1 cells, albeit at a very low level (2.7 ± 0.63% of CES1 mRNA). However, expression of KIAA1363 was reduced by ∼40% in THP1-shRNA cell lines (p82–p84; Fig. 5), indicating that CES1 silencing does not lead to any compensatory increase in the expression of KIAA1363 to account for the observed lack of decrease in CE hydrolytic activity in THP1-shRNA cells.

Fig. 5.

Silencing of CES1 does not lead to a compensatory increase in KIAA1363 expression in THP1-shRNA cell lines. Total RNA was prepared from THP1-shRNA macrophages, and mRNA levels of KIAA1363 were determined by real-time PCR. Data are means ± SD for n = 3 and are expressed as a percentage of pRSL.

To further examine the mechanism(s) underlying the observed discordance between CES1 expression and cellular CE hydrolytic activity as well as FC efflux, the expression of another member of the CES family, CES3, was evaluated, and the data are shown in Fig. 6A. In THP1-shRNA cell lines (p81 and p84), there was an almost twofold increase in the expression of CES3 compared with the control cell line THP1-pRSL. It is noteworthy that in untransfected wild-type THP1 cells, CES3 expression is only 29.2 ± 6.08% of CES1 mRNA expression.

Fig. 6.

CES1 silencing results in a compensatory increase in CES3 expression in THP1-shRNA cell lines. A: CES1 and CES3 mRNA levels were determined by real-time RT-PCR. After normalization to β-actin, relative expression was plotted as a percentage of the THP1-pRSL control. Data are presented as means ± SD for n = 3, and each determination was performed in duplicate. *P < 0.05. B: the increase in CES3 expression above that seen in control THP1-pRSL cells was calculated (CES3 expression as %pRSL control − 100) and plotted as superimposed stacked bars to demonstrate the complete restoration of total CES expression (CES1 + CES3) to normal (100%) levels. C: relative expression of CES1 and CES3 in primary human blood-derived monocytes (day 0) and macrophage colony-stimulating factor-differentiated macrophages (days 3 and 5). Data are presented as means ± SD for n = 4, and each determination was performed in duplicate.

To examine the relative changes in CES3 expression after CES1 silencing, mRNA expression of CES3 was also plotted as the percent increase over pRSL (CES3 expression − 100%) as superimposed stacked bars in Fig. 6B. In all THP1-shRNA cell lines (p81–p84), the increase in CES3 expression almost completely restored the expression back to control levels (pRSL or 100%), suggesting that when CES1 is knocked down via shRNA, there is a compensatory increase in CES3.

To further establish relevance to human physiology, the relative expression of CES1 and CES3 was also examined in primary human blood-derived monocytes/macrophages. CES3 was expressed in primary human monocytes as well as macrophage colony-stimulating factor-differentiated macrophages. Consistent with the relative expression of CES1 and CES3 in THP1, CES3 expression was <30% of that of CES1 (Fig. 6C).

Cloning of human CES3 from human THP1 macrophages.

The nucleotide sequence of human CES3 has been reported, and a two-step nested PCR with primers based on the available CES3 sequence was used to clone CES from human THP1 macrophages. Three different clones were sequenced in both directions and compared with the reported CES3 sequence (NM_024922.4). All clones sequenced lacked exon 8, and the predicted protein sequence was identical to that of the reported human brain CES3 sequence (NP_079198). There were no other differences noted between the two sequences. The same strategy was used to clone CES3 from human liver RNA, and the clones obtained contained exon 8 and were identical to the reported CES3 sequence (data not shown). A comparison of the protein sequence of CES1 and macrophage CES3 identified here is shown in Fig. 7. While there was only a 44% identity between the two proteins, the active site serine was present in the conserved GXSXG motif. The acidic (glutamic acid) and histidine residue that together with active site serine form the catalytic triad as well as the SEDCL motif containing the cysteine involved in the formation of the disulfide bridge and the putative loop domain are also conserved.

Fig. 7.

Comparison of human macrophage (HMac)CES1 and HMacCES3 protein sequences. The predicted protein sequence based on the nucleotide sequence of HMacCES3 was aligned with the reported human CES1 sequence (NP_001020366.1). The conserved catalytic triad consisting of the active site serine in the GXSXG motif and glutamic acid and histidine residues are boxed. The two conserved cysteine residues (one in the SEDCL motif) are also boxed.

Newly cloned CES3 catalyzes the hydrolysis of CE.

To determine whether CES3 codes for a putative CEH, total cell lysates of COS-7 cells transfected with the control vector pCMV or the CES3 expression vector pCMV-CES3 were assayed for CE hydrolysis using micelle and droplet substrates. As shown in Fig. 8A, CES3 expression increased the intracellular CE hydrolytic activity by almost fourfold. Expression of the protein was determined by Western blot analysis, and c-myc-tagged CES3 was present only in cells transfected with CES3. It should be noted that no antibody against human CES3 is currently available. As a positive control, some cells were also transfected with the pCMV-CEH vector previously developed in the laboratory (11). As expected, transfection with pCMV-CEH increased the intracellular CE hydrolytic activity (data not shown) (5), and expression of the protein was confirmed by Western blot analyses (Fig. 8B). It is noteworthy that only a single immunoreactive band was obtained in cells transfected with pCMV-CEH and that CES1 antibody did not cross react with CES3; no immunoreactive bands were seen in cells transfected with pCMV-CES3. Cells transfected with the CES3 clone obtained from the liver, identical to reported CES3 and containing exon 8, also showed a comparable increase in CE hydrolytic activity (data not shown), suggesting that the presence or absence of exon 8 does not affect CE hydrolysis.

Fig. 8.

CES3 catalyzes CE hydrolysis. COS-7 cells were transfected with control vector (pCMV) or CES3 expression vector (pCMV-CES3), and cell lysates were assayed for CE hydrolytic activity using either micelle or droplet substrates. All assays were performed in duplicate, and three protein concentrations were used to ensure linearity. A: the percent increase in enzyme activity over that observed in cells transfected with the control vector pCMV is shown (means ± SD, n = 3). The specific enzyme activity obtained with cells transfected with pCMV was 213 ± 37 and 156 ± 137 pmol·h⁻¹·mg⁻¹, respectively, for micelle and droplet substrates. *P < 0.05. B: cell lysates were subjected to Western blot analyses to determine the expression of c-myc-tagged CES3 protein in transfected cells. Some cells were transfected with pCMV-CEH, and expression of the protein was confirmed by positive immunoreactivity to CES1 antibody.

Overexpression of CES3 mobilizes intracellular CE and decreases cellular cholesterol content.

To further evaluate the role of CES3 in the mobilization of intracellular/stored CE, CES3 was overexpressed in agmACAT1 cells, which accumulate reproducible amounts of CE due to stable overexpression of ACAT1. Intracellular CE containing lipid droplets was stained with oil red O, and, as shown in Fig. 9A, cells transfected with pCMV-CES3 contained noticeably fewer numbers of oil red O-stained lipid droplets. CES3-mediated mobilization of intracellular lipid droplets was confirmed by an analysis of cellular lipids. Significant decreases in cellular total cholesterol, FC, and CE levels were observed in cells transfected with the CES3 expression vector pCMV-CES3 (Fig. 9B). Taken together with the data demonstrating the CE hydrolytic activity of CES3, these data clearly establish that CES3 can mobilize intracellular CE and thereby reduce cellular CE content.

Fig. 9.

Transient overexpression of CES3 in agmACAT1 cells results in the increased mobilization of lipid droplets and decreases cellular lipid content. agmACAT1 cells were transiently transfected with either control (pCMV) or CES3 expression vector (pCMV-CES3). Seventy-two hours afte transfection, one set of cells was fixed and stained with oil red O. Representative images are shown in A. Total lipids were extracted from the second set of cells and analyzed by gas chromotography. Cellular total cholesterol (TC), FC or unesterified cholesterol, and CEs were normalized to cellular protein, and the data (means ± SD, n = 6) are shown in B. Individual P values are shown above the bars.

DISCUSSION

The critical role of intracellular CE hydrolysis in FC efflux and reverse cholesterol transport is well documented and supported by the observed increase in susceptibility to atherosclerosis in species with low CE hydrolytic activity in macrophages (15, 23). Although several different enzymes, namely, HSL, CEL, CEH (CES1), and KIAA1363 [NCEH1 (also called AADACL1)], have been reported to catalyze CE hydrolysis, the relative contribution of these enzymes to total cellular CE hydrolytic activity has not been systematically evaluated. Furthermore, despite the observed discordance between gene expression and residual CE hydrolytic activity in the various gene ablation models, no explanation is currently available to explain these discrepancies. In the present study, we used two different approaches to determine the contribution of CES1 to CE hydrolysis in human THP1 macrophages. In the first approach, CES1 protein was specifically depleted from cell lysates using CES1-specific antibodies wherein compensation by any cellular/biological process is not applicable. Our results (Fig. 1) demonstrate that CES1 contributes to >70% of the total cellular CE hydrolytic activity. These data are consistent with that reported by Crow et al. (3), who also demonstrated complete immune depletion of CES1 leading to a >85% decrease in enzyme activity measured using p-nitrophenyl valerate as an esterase substrate. In the second approach, shRNA-mediated CES1 knockdown cell lines were generated where CES1 is silenced in live cells and, therefore, compensatory mechanisms, if any, will still be operative. The data presented here demonstrate that shRNA-dependent knockdown of CES1 affected neither total CE hydrolase activity (Fig. 4) nor FC efflux (Fig. 3) due a compensatory increase in another gene (CES3) that is expressed only at the <30% level in wild-type THP1 cells (Fig. 5) as well as primary human blood-derived monocytes/macrophages (Fig. 6C). Evaluation of the relative expression of CES1 and CES3 in control (pRSL) and THP1-shRNA cell lines (p81–p84) further demonstrated that compensatory increase in CES3 restored the total expression (CES1 + CES3) to control levels in all lines (Fig. 6). It is noteworthy that HSL expression was undetectable in THP1 human macrophages, and, although the expression of KIAA1363 was detected in all cell lines used here, CES1 silencing did not lead to any detectable increase in the expression of this enzyme (Fig. 6).

In contrast to the two-approach (immune depletion as well as gene silencing) strategy used in the present study to evaluate the relative contribution of CES1 to total cellular CE hydrolytic activity, earlier studies have solely focused on single gene ablation. HSL deficiency in macrophages was originally thought to have no measurable effect on CE hydrolysis (16, 2), and, although a decrease in activity was later observed, it did not affect cellular CE turnover (1). Macrophages from KIAA1363-deficient mice showed only a 40–50% decrease in total CE hydrolytic activity despite the complete loss of mRNA and protein (21). Loss of TGH/Ces3 in mice also led to only a 30% decrease in hepatic CEH activity (22). However, the mechanism(s) underlying this discordance between enzyme activity and expression of the targeted gene was not evaluated in these studies. It remains to be seen whether a compensatory increase in known or as-yet-unidentified CEHs could account for the incomplete loss in CE hydrolytic activity or CE turnover despite gene ablation of KIAA1363 or HSL.

Human CES3 was initially cloned and characterized by Sanghani et al. as a novel enzyme involved in the metabolism of xenobiotics (19), which are established substrates for CES1, and it was shown to contain the conserved catalytic and structural residues leading to broad yet similar substrate specificities (20). Homes et al. (9) compared human CES3 and CES1 and concluded that the predicted three-dimensional structure of CES3 is similar to that of CES1, suggestive of similar enzymatic activities. It is noteworthy that the CES family of enzymes was originally characterized as enzymes involved in xenobiotic metabolism and has only recently being identified to play a significant role in metabolism of endogenous lipids. The new variant of macrophage CES3 identified here also encodes for a bona fide lipid hydrolase, as shown by the increase in CE as well as triolein hydrolytic activity in COS-7 cells transiently transfected with the CES3 expression vector (pCMV-CES3; Fig. 8). These data also suggest that the absence of exon 8 did not affect the ability of this new variant of CES3 to hydrolyze neutral lipids. Furthermore, the observed increase in CES3 in THP1-shRNA cells maintained the normal CE hydrolytic activity as well as FC efflux, suggesting that CES3 is involved in intracellular CE mobilization. This role of CES3 in cellular lipid mobilization was further confirmed by the demonstration of a decrease in cellular lipid droplets and CE content in agmACAT1 cells transiently transfected with the CES3 expression vector (Fig. 9). While the present study demonstrates, for the first time, that a compensatory increase in CES3 expression is likely involved in the observed lack of change in intracellular CE hydrolysis and FC efflux, future studies will determine the physiological conditions that can potentially induce CES3 expression.

Gene compensation is an established mechanism where loss of one gene results in a lack of a phenotypic change, and it was specifically predicted in enzymes belonging to gene families and with overlapping substrate specificities (7), as is the case for CES1 and CES3, both of which belong to the CES family of enzymes. These redundant or backup genes have been shown to be transcriptionally responsive to the intactness of their redundant partner and are upregulated if the latter is inactivated, and this new paradigm of “responsive backup circuits” has been validated by numerous examples (11, 12, 17). The data presented here support this concept where an increase in CES3 expression occurs after silencing of CES1. The mechanism(s) involved in the observed upregulation of CES3 in THP1-shRNA cells is currently under investigation.

In conclusion, the data presented here show that CES1 contributes to >70% of the CE hydrolytic activity in human THP1 macrophages. Furthermore, the experiments described here also demonstrate functional redundancy in human macrophage CEHs where shRNA-mediated knockdown of CES1 led to a compensatory increase in otherwise less expressed CES3, resulting in no apparent change in intracellular CE hydrolysis and FC efflux. These data not only provide the first evidence of functional compensation but also offer a novel perspective for understanding macrophage cholesterol homeostasis. Thus, it is imperative to identify the compensatory mechanisms that become operational pursuant to ablation of any single CEH (instead of the currently prevailing emphasis on establishing the “major CEH”) to extend our contemporary understanding of this obligatory first and rate-limiting therapeutically relevant step of reverse cholesterol transport.

GRANTS

This work was supported in part by National Heart, Lung, and Blood Institute Research Grant HL-069946 (to S. Ghosh).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: B.Z., J.B., J.W., S.A.M., and S.G. performed experiments; B.Z., J.B., J.W., and S.G. analyzed data; B.Z. and S.G. approved final version of manuscript; S.G. conception and design of research; S.G. interpreted results of experiments; S.G. prepared figures; S.G. drafted manuscript; S.G. edited and revised manuscript.