Biochimica et Biophysica Acta 1821 (2012) 1003–1011

Contents lists available at SciVerse ScienceDirect

Biochimica et Biophysica Acta

j ourna l homepage: www.e lsev ie r .com/ locate /bba l ip

Impact of endothelial lipase on cellular lipid composition

Monika Riederer a,1, Harald Köfeler b, Margarete Lechleitner a, Michaela Tritscher a, Saša Frank a,⁎a Institute of Molecular Biology and Biochemistry, Center of Molecular Medicine, Medical University Graz, Graz, Austriab Center for Medical Research, Core Facility Mass Spectrometry, Medical University Graz, Graz, Austria

Abbreviations: EL, endothelial lipase; LPC, lysophpalmitoyl-lysophosphatidylcholine; 18:2 LPC, linoleoylLPC; 18:1 LPC, oleoyl-LPC; HAEC, human aortic endoacid; PL, phospholipids; LysoPL, lysophospholipids; HATGL, adipose triglyceride lipase; DGAT, diacylglycerolphocholine cytidylyltransferase; ET, CTP:phosphoethanLPLATs, lysophospholipid acyltransferases; BSA, bovineserum; MS, mass spectrometry⁎ Corresponding author at: Institute of Molecular Bio

of Molecular Medicine, Medical University Graz, Harractria. Tel.: +43 316 380 4193; fax: +43 316 380 9615.

E-mail address: sasa.frank@medunigraz.at (S. Frank)1 Present address: University of Applied Sciences, Bio

Allee 13, 8020 Graz, Austria.

1388-1981 © 2012 Elsevier B.V.doi:10.1016/j.bbalip.2012.03.006

Open access under CC BY

a b s t r a c t

a r t i c l e i n f o

Article history:Received 10 May 2011Received in revised form 9 February 2012Accepted 5 March 2012Available online 6 April 2012

Keywords:Mass spectrometryEndothelial cellLysophospholipidHDLAdenovirus

Using mass spectrometry (MS), we examined the impact of endothelial lipase (EL) overexpression on the cel-lular phospholipid (PL) and triglyceride (TG) content of human aortic endothelial cells (HAEC) and of mouseplasma and liver tissue. In HAEC incubated with the major EL substrate, HDL, adenovirus (Ad)-mediated ELoverexpression resulted in the generation of various lysophosphatidylcholine (LPC) and lysophosphatidyletha-nolamine (LPE) species in cell culture supernatants. While the cellular phosphatidylethanolamine (PE) contentremained unaltered, cellular phosphatidylcholine (PC)-, LPC- and TG-contents were significantly increasedupon EL overexpression. Importantly, cellular lipid composition was not altered when EL was overexpressedin the absence of HDL. [14C]-LPC accumulated in EL overexpressing, but not LacZ-control cells, incubatedwith [14C]-PC labeled HDL, indicating EL-mediated LPC supply. Exogenously added [14C]-LPC accumulated inHAEC as well. Its conversion to [14C]-PC was sensitive to a lysophospholipid acyltransferase (LPLAT) inhibitor,thimerosal. Incorporation of [3H]-Choline into cellular PC was 56% lower in EL compared with LacZ cells, indi-cating decreased endogenous PC synthesis. In mice, adenovirus mediated EL overexpression decreased plasmaPC, PE and LPC and increased liver LPC, LPE and TG content. Based on our results, we conclude that EL not onlysupplies cells with FFA as found previously, but also with HDL-derived LPC and LPE species resulting in in-creased cellular TG and PC content as well as decreased endogenous PC synthesis.

© 2012 Elsevier B.V. Open access under CC BY-NC-ND license.

1. Introduction

EL is a member of the triglyceride (TG) lipase gene family [1,2]. Adistinct feature of EL, compared with other family members is its endo-thelial expression. EL is synthesized as a 55 kDa protein that is secretedas a 68 kDa protein upon maturation. After secretion, a portion of EL iscleaved and inactivated bymembers of themammalian proprotein con-vertase family [3,4]. Uncleaved EL is bound to heparan-sulfate proteo-glycans on the cell surface. Here, by its bridging function, EL facilitatesHDL particle binding and uptake [5,6], as well as selective uptake ofHDL cholesteryl esters [6].

osphatidylcholine; 16:0 LPC,-LPC; 20:4 LPC, arachidonoyl-thelial cells; AA, arachidonicDL, high-density lipoprotein;acyltransferase; CT, CTP:phos-olamine cytidylyltransferase;serum albumin; FCS, fetal calf

logy and Biochemistry, Centerhgasse 21/III, 8010 Graz, Aus-

.medical Science, Eggenberger

-NC-ND license.

By virtue of its pronounced sn-1 phospholipase activity, ELcleaves lipoprotein-associated phospholipids (PL), primarily HDL-phosphatidylethanolamine (HDL-PE) and -phosphatidylcholine (HDL-PC), generating mostly saturated free fatty acids (FFA) [7–9] and sn-1 lyso sn-2 acyl lysophospholipids (LysoPL) [7,9]. By its rather weaklysophospholipase activity, EL further hydrolyzes some of those EL-generated LPL, liberating unsaturated FFA [7].

EL expression is increased under inflammatory conditions as ob-served in cultured endothelial cells exposed to interleukin-1ß (IL-1ß) or tumor necrosis factor α (TNFα) as well as in mice and humansupon LPS injection [10,11]. Furthermore, EL protein mass and activityhave been found to be elevated in patients with subclinical inflamma-tion as encountered in the metabolic syndrome and coronary athero-sclerosis [12,13].

The capacity of EL to generate FFA and LysoPL [7] and to supplycells with those lipids [7,14], strongly argues for a role of EL in themodulation of cellular lipid content and composition.

Two endogenous biosynthetic pathways, the de novo (KennedyPathway) [15] and the remodeling pathway (Lands' Cycle) [16] areresponsible for the maintenance of the physiological lipid balance ofthe cell [17]. Most abundant glycero(phospho)lipids, generated inthe de novo pathway are derived from phosphatidic acid (PA)-deriveddiacylglycerol (DAG). TG is formed from DAG in a reaction catalyzedby diacylglycerol acyltransferase (DGAT) [18] and can be converted

1004 M. Riederer et al. / Biochimica et Biophysica Acta 1821 (2012) 1003–1011

back to DAG and FFA by adipose triglyceride lipase (ATGL) [19]. PCand PE are formed in the CDP-choline and CDP-ethanolamine path-way, whereby the rate limiting steps are catalyzed by the enzymesCTP:phosphocholine cytidylyltransferase (CT) and CTP:phosphoetha-nolamine cytidylyltransferase (ET) [15,20]. The de novo generated PLin turn undergo remodeling (Lands' pathway) through the concertedactions of cellular phospholipases A2 (PLA2) and lysophospholipidacyltransferases (LPLAT), resulting in alteration of PL acyl composi-tion primarily at the sn-2 position [16].

In the present study, we addressed the impact of EL overexpres-sion on the content and composition of the endothelial TG- and PL-pools, both in vitro in EL-overexpressing HAEC as well as in vivo inmice overexpressing EL upon adenoviral infection.

2. Materials and methods

2.1. Cell culture

Human primary aortic endothelial cells (HAEC) were obtainedfrom Lonza (Cologne, Germany) and maintained in endothelial cellgrowth medium [EGM-MV Bullet Kit=EBM medium+growth sup-plements+FCS (Lonza)] supplemented with 50 IU/ml penicillin,and 50 μg/ml streptomycin. Cells were cultured in gelatine-coateddishes at 37 °C in a 5% CO2 humidified atmosphere and were usedfor experiments from passages 5 to 10. Cells were seeded (75,000/well) in 12-well plates 48 h before exposure to HDL.

2.2. Isolation of human HDL

HDL (subclass 3, d=1.125–1.21 g/ml) was prepared by sequentialultracentrifugation of plasma obtained from normolipidemic blooddonors as described previously [6].

2.3. Adenovirus generation

Adenoviruses (Ad) encoding human EL (EL) and bacterial ß-galac-tosidase (LacZ) were prepared exactly as described previously [6].

2.4. HDL labeling

Labeling of HDL with [14C]-PC (NEN) was performed as described[7]. Briefly, 2 μCi of 1-palmitoyl-2-[14C] arachidonyl-PC was driedunder nitrogen, redissolved in 30 μl ethanol, and added to a solutioncontaining HDL (3 mg protein) and lipoprotein-deficient serum(700 μl) in a final volume of 1.7 ml in PBS. Subsequently, this mixturewas incubated under argon in a shaking water bath at 37 °C for 16 h.Thereafter, labeled HDL ([14C]-HDL) was reisolated by density gradi-ent ultracentrifugation in a TLX120 bench-top ultracentrifuge in aTLA100.4 rotor (Beckman). The HDL band was aspirated and desaltedby size-exclusion chromatography using 10DG columns (BioRad).

2.5. Adenoviral infection of HAEC and treatments with HDL and [14C]-HDL

HAEC (75000/well or 150000/well) were seeded in 12- or 6-wellplates. Forty eight hours later, confluent cells were infected with EL-Ad or LacZ-Ad at multiplicity of infection (MOI) of 50 in endothelialbasal medium (EBM) without fetal calf serum (FCS). Following a2 h-infection period, cells were grown in complete EBM for additional24 h. Thereafter, cells were washed with PBS and incubated with EBMmedium without supplements and without serum (serum-free medi-um) for 3 h. Then, serum-free medium was removed and replacedwith fresh serum-free medium supplemented with either: i) PBS or300 μg/ml HDL in PBS for 5 h, followed by lipid extraction, HPLC andMS or ii) 100 μg/ml [14C]-HDL for up to 5 h, followed by thin layerchromatography (TLC).

2.6. Lipid extraction, HPLC and MS

Following materials: i) the cell layer of one 6-well, ii) 1.5 ml of cellculture supernatant, iii) 450 μg of HDL (protein) re-isolated from cellincubation media as described above in Section 2.4, iv) 100 μl ofmouse plasma, and v) 200 mg of liver tissue, were extracted accord-ing to Bligh and Dyer [21] in the presence of an internal standard(PC 12:0/12:0) and dried under a stream of nitrogen. Dried lipid ex-tracts were resuspended in 200 μl CHCl3/MeOH (1:1, v/v) containing1 pmol/μl of each LPC 17:1, LPE 17:1, PE 12:0/13:0 and PC 12:0/13:0respectively, serving as internal standards. Chromatographic separa-tion of lipids was performed by an Accela HPLC (Thermo Scientific)on a Thermo Hypersil GOLD C18, 100×1 mm, 1.9 μm column. SolventA was a water solution of 1% ammonia acetate (v/v) and 0.1% formicacid (v/v) and solvent B was acetonitrile/2-propanol (5:2, v/v) sup-plemented with 1% ammonia acetate (v/v) and 0.1% formic acid(v/v), respectively. The gradient was run from 35% to 70% B for4 min, then to 100% B in additional 16 min with subsequent hold at100% for 10 min. The flow rate was 250 μl/min. Phospholipid specieswere determined by a TSQ Quantum ultra (Thermo Scientific) triplequadrupole instrument in positive ESI mode. The spray voltage wasset to 4500 V and capillary voltage to 35 V. The PC species weredetected in a precursor ion scan on m/z 184 at 34 eV. Acquisition oftriglyceride species was performed on a LTQ-FT in FT full scan modeat a resolution of 200,000. Phospholipid peak areas were calculatedby QuanBrowser for all lipid species and the calculated peak areasfor each species were expressed according to the amount of internalstandard. TG was quantified by Lipid Data Analyzer [22].

2.7. Monitoring of [14C]-LPC in cells by TLC

Following incubation with [14C]-HDL for the indicated time pe-riods, lacZ- and EL-Ad infected cells (12-well dishes) were extensive-ly washed with PBS and extracted with hexane/isopropanol (3:2, v/v), extracts evaporated in the SpeedVac and redissolved in chloro-form before TLC using a solvent system for the separation of phospho-lipids (propionic acid/chloroform/1-propanol/water (3:2:2:1; v/v)).The signals corresponding to [14C]-LPC were visualized upon expo-sure of the TLC plates to a tritium screen (GE Healthcare) on theSTORM imager.

2.8. Conversion of [14C]-LPC to [14C]-PC in HAEC

HAEC were incubated with 1 μM [14C]-16:0 LPC (NEN) in full me-dium containing 5% FCS in the absence or presence of indicated con-centrations of thimerosal (Sigma) for 30 min, followed by theanalysis of [14C]-LPC and [14C]-PC by TLC as described above.

2.9. [3H]-Choline incorporation into cellular PC

Twenty-four hours after infection with LacZ- or EL-Ad (in 12-well dishes), HAEC were washed with PBS and incubated withserum-free medium for 3 h followed by incubation with serum-free medium supplemented with 300 μg/ml HDL and 1 μCi/ml[3H]-Choline (NEN) for 5 h. After the incubation, media were re-moved and cells were washed 3 times with PBS, followed by extrac-tion and TLC as described above. The signals corresponding to PCwere visualized by I2 and lipid spots were cut out of the TLC plateand measured by scintillation counting. The delipidated cell layerswere lysed with 1 ml of 0.3 M NaOH/0.1% SDS and the protein con-tent was determined according to [23].

2.10. Incorporation of [14C]-arachidonic acid into cellular TG

Eight hours after infection with LacZ- or EL-Ad, cells were washedand incubated with 1 μCi/ml [14C]-arachidonic acid in full medium

BA

DC

E

F

Fig. 1. Overexpression of EL in the presence of HDL increases LPC and LPE content in cell medium.LacZ and EL-overexpressing HAEC were incubated in serum-free medium supple-mented with 300 μg/ml HDL for 5 h. Subsequently, A) PC- C) PE- E) LPC- and F) LPE-content/profile of cell culture supernatants as well as B) PC- and D) PE-content of re-isolatedHDL (450 μg protein) were determined by MS. Results shown are mean±SD of two experiments performed in triplicates.

1005M. Riederer et al. / Biochimica et Biophysica Acta 1821 (2012) 1003–1011

containing 5% FCS for 18 h. After extensive washing with PBS, cellswere extracted with hexane/isopropanol (3:2, v/v), evaporated inthe SpeedVac and redissolved in chloroform before TLC using a sol-vent system for the separation of neutral lipids (hexane/diethylether/acetic acid, 70:29:1, vol/vol). The signals corresponding to[14C]-TG were visualized upon exposure of the TLC plates to a tritiumscreen (GE Healthcare) on the STORM imager. Quantification wasperformed by densitometric volume report analysis.

2.11. Western blot

HAEC infected with LacZ- or EL-Ad (MOI 50; 12-well dishes), werewashed with PBS and collected in 200 μl/well of loading buffer [20%(w/v) glycerol, 5% (w/v) SDS, 0.15% (w/v) bromophenol blue,63 mmol/I Tris–HCl, pH 6.8, and 5% (v/v) ß-mercaptoethanol] fol-lowed by boiling for 10 min. 40 μl of the lysate was applied to eachlane and analyzed by SDS-PAGE (10% gel) and subsequent

BA

C

D E

F

LacZ

phase contrast

Nile Red

EL

Fig. 2. Overexpression of EL in the presence of HDL alters lipid composition of HAEC.LacZ and EL-overexpressing HAEC were incubated in serum-free medium supplemented with300 μg/ml HDL for 5 h. Subsequently, A) PC- B) PE- C) LPC- and D) TG-content/profile was determined in lipid extracts of cell layers using MS. Results represent the mean of trip-licate determinations±SD of a representative experiment performed six times. E) LacZ or EL-overexpressing HAEC were incubated in serum-free medium supplemented with300 μg/ml HDL for 20 h. The lower panels show lipid accumulation determined by fluorescence microscopy of cells stained with Nile Red. In the upper panels, the same visualfield is shown in phase contrast. F) LacZ and EL-overexpressing cells were incubated with 1 μCi/ml [14C]-arachidonic acid in full medium containing 5% FCS for 18 h. Followingthe separation of neutral lipids by TLC (hexane/diethyl ether/acetic acid, 70:29:1, vol/vol), the signals corresponding to [14C]-TG were quantified by densitometry. Results representthe mean of duplicate determinations±SD of three experiments.

1006 M. Riederer et al. / Biochimica et Biophysica Acta 1821 (2012) 1003–1011

Fig. 3. EL supplies cells with HDL-derived LPC.A) LacZ- and EL-Ad infected HAEC wereincubated with [14C]-PC labeled HDL in serum-free medium for 5 h. Following exten-sive washing, lipids were extracted from cell layers and separated by TLC. The signalscorresponding to [14C]-LPC were visualized upon exposure of the TLC plates to a tritiumscreen (GE Healthcare) on the STORM imager.B) HAEC were incubated with 1 μM [14C]-LPC in full medium containing 5% FCS in the absence or presence of indicated concen-trations of thimerosal for 30 min, followed by extraction and analysis of [14C]-LPC and[14C]-PC by TLC as described in A).

Fig. 4. EL overexpression decreases endogenous PC synthesis.Following incubation inserum-free medium for 3 h, LacZ- and EL-Ad infected HAEC (in 12-well dishes), wereco-incubated in serum-free medium with 300 μg/ml HDL and 1 μCi/ml [3H]-Cholinefor 5 h. Following three washes of cell layers with PBS and extraction by hexane/iso-propanol (3:2, v/v), the cell extracts were subjected to TLC (propionic acid/chloro-form/1-propanol/water (3:2:2:1; v/v). The signals corresponding to PC weremeasured by scintillation counting. Results represent the mean of a triplicate experi-ment±SD of three experiments.

1007M. Riederer et al. / Biochimica et Biophysica Acta 1821 (2012) 1003–1011

immunoblotting using an EL-specific antibody [6] and a HRP-labeledanti rabbit secondary antibody. Protein signals were detected by anenhanced chemiluminescent substrate of HRP (SuperSignal WestPico, Pierce).

2.12. Nile Red staining

LacZ- or EL-infected HAEC were incubated with 300 μg/ml HDL for20 h and incubated with Nile Red (100 ng/ml) for 15 min. Lipid stain-ing was evaluated by fluorescence microscopy (Zeiss).

2.13. Phospholipase activity assay

Twenty-four hours after infection (in 12-well dishes), LacZ- or EL-Adcells were washed with PBS and incubated with 300 μl of serum-freemedia supplemented with 10 U/ml of heparin for 30 min. Theheparin-release media were collected, frozen at −80 °C and subse-quently used for phospholipase activity assay. The assay was done asdescribed previously [6]. Briefly, the PC substrate was made by mixing[14C]-dipalmitoyl-phosphatidylcholine (PC) (NEN, Boston, MA, U.S.A.),lecithin (1 mg/ml) and substrate buffer Tris/TCNB [100 mM Tris/HCL,pH 7.4, 1% (v/v) Triton X-100, 5 mM CaCl2, 200 mM NaCl, 0.1% (w/v)NEFA-free BSA] and subsequently dried under nitrogen. Substrate wasdriedunder a streamof nitrogen and reconstituted in the substrate buff-er. Aliquots (190 μl) of the heparin-release media were mixed with10 μl of the substrate and incubated at 37 °C for 1 h. The reaction wasterminated by addition of 1 ml of 0.2 M HCl and extraction with hex-ane/isopropanol (3:2, v/v, 0.1% HCl). Aliquots (500 μl) of the upperphase were dried in a speed vac and reconstituted in 100 μl of hex-ane/isopropanol (3:2, v/v). After separation by TLC (hexane/diethylether/acetic acid, 70:29:1, vol/vol), the liberated 14C-NEFAwere quanti-tated in a scintillation counter (Beckman). The activitywas expressed ascpm/mg cell protein.

2.14. Intravenous (i.v.) injection of adenovirus and perfusion

10–12 weeks old male C57Bl/6J mice (5 per group) were anesthe-tized with Isofluran (Pharmacia & Upjohn SA, Guyancourt, France)and injected with 3×109 particles of LacZ- or EL-Ad in 100 μl of0.9% NaCl via tail vein injection. 24 h later, blood was collected fromthe retroorbital plexus into EDTA-containing tubes and spun. Plasmawas stored at −70 °C before lipid extraction. For the analysis ofliver samples, anesthetized mice were perfused with PBS throughthe left ventricle before tissue sampling. Animal experiments were

conducted in conformity with the Public Health Service Policy onHuman Care and Use of Laboratory Animals and were approved bythe Austrian Ministry of Science and Research according to the Regu-lations for Animal Experimentation.

2.15. Statistical analysis

Cell culture experiments were performed at least three times andvalues are expressed as mean±SD. Statistical analyses were per-formed in GraphPad Prism 5 using the paired t test, for experimentscomparing 2 groups; and two-way ANOVA with Bonferroni-correction for experiments comparing 3 or more groups. Pb0.05was considered statistically significant (*).

3. Results

3.1. EL releases LPC- and LPE-species from HDL-PL

To examine the impact of EL on cellular PC, PE and TG, human ELwas overexpressed in HAEC by adenoviral gene transfer (EL-Ad).The applied amount of EL-Ad (MOI 50) yielded a moderate EL overex-pression: a 1.7-fold increase in the heparin-releasable phospholipaseactivity, compared with LacZ-Ad infected cells (Supplementary Fig.1A) and EL signals of moderate intensity detected by Western blot-ting of EL-Ad cell lysates, compared with no signals in cell lysates ofLacZ-Ad infected cells (Supplementary Fig. 1B). Following incubationof EL- and LacZ-Ad infected cells with HDL (in serum-free medium),we analyzed the phospholipid profile of the incubation media byMS. The analysis revealed a slight but not significant EL-mediated de-crease in total PC content (Fig. 1A) as well as in the majority of dis-tinct PC species (Supplementary Fig. 2A), when compared withLacZ-adenovirus infected control cell media. In contrast, the decreasein PC content was significant in HDL, re-isolated from incubationmedia of EL overexpressing cells compared with LacZ samples(Fig. 1B; Supplementary Fig. 2B). The total PE as well as all distinctPE species were significantly decreased in both incubation media(Fig. 1C; Supplementary Fig. 2C) as well as in re-isolated HDL(Fig. 1D; Supplementary Fig. 2D) of EL compared with LacZ incuba-tions. The alterations in HDL-PC and -PE species were accompaniedby profound increases in several LPC and LPE species (Fig. 1E and F,respectively).

A

B

DC

Fig. 5. EL overexpression alters plasma PL in mice.Blood was collected from fed, male C57Bl/6J mice, 24 h after i.v. injection of 3×109 virus particles of LacZ- or EL-Ad. The A) PC-B) PE- C) LPC-content/profile was determined in lipid extracts of 100 μl plasma. D) Shows the ratio of LPC normalized to total PC content. Results represent the mean±SD of arepresentative experiment with 5 animals per group. Two additional experiments with 5 animals per group gave similar results.

1008 M. Riederer et al. / Biochimica et Biophysica Acta 1821 (2012) 1003–1011

3.2. EL alters the PC- and TG-profile of HAEC

Next, we examined the impact of EL overexpression in the pres-ence of HDL on the cellular PC, PE and TG profile of HAEC. Both thetotal PC content as well as various PC species were increased in EL-overexpressing cells compared with LacZ controls (Fig. 2A; Supple-mentary Fig. 3A). In contrast to PC, both total PE (Fig. 2B) as well asthe majority of various PE species (Supplementary Fig. 3B) were unal-tered in EL overexpressing HAEC compared with LacZ-infected con-trol cells. The total LPC content as well as various LPC species weremore abundant in EL overexpressing HAEC than in control cells(Fig. 2C). Unfortunately, LPE species were below the detection limitof the applied analytical approach. Importantly, cellular PC and PEwere unaltered upon EL overexpression in the absence of HDL (Sup-plementary Fig. 4A,B,C,D). EL overexpression also markedly increasedthe cellular TG content, compared with LacZ-infected control cells(Fig. 2D, and Supplementary Table 1 for detailed species distribution).This increase in TG could also be demonstrated by staining of cellswith the lipophilic dye Nile Red (Fig. 2E). Additionally, duringincubation in full medium containing 5% FCS, EL overexpressingcells incorporated significantly more of exogenously added [14C]-arachidonic acid into TG, compared with LacZ-Ad control cells(Fig. 2F).

3.3. EL supplies cells with HDL-derived LPC

The accumulation of LPC species in EL overexpressing cells(Fig. 2C) suggested the EL-mediated supply with HDL-derived LPC.To directly demonstrate the ability of EL to supply cells with HDL-derived LPC, HAEC were incubated with HDL labeled within its phos-pholipid moiety with [14C]-PC. As shown in Fig. 3A, a time-dependentincrease in cellular [14C]-LPC was observed in EL-overexpressing butnot in LacZ-control cells.

To further examine the fate of exogenously added LPC in cells,HAEC were incubated with [14C]-LPC in the absence or presence ofthimerosal (a general LPCAT inhibitor). As indicated in Fig. 3B, exog-enous [14C]-LPC rapidly accumulated in cells, and its conversion to[14C]-PC could be inhibited by thimerosal in a dose-dependent man-ner (Fig. 3B).

3.4. EL overexpression diminishes endogenous PC synthesis

To examine the impact of EL overexpression on the rate of endog-enous PC synthesis, we measured [3H]-Choline incorporation into en-dogenous PC in EL- and LacZ-Ad infected cells in the presence of HDL.As shown in Fig. 4, EL overexpressing cells incorporated 56% less [3H]-

A

B

C

ED

Fig. 6. EL overexpression promotes accumulation of TAG, LPC and LPE in mouse liver.Perfused liver tissue was collected from fed, male C57Bl/6J mice, 24 h after i.v. injection of3×109 virus particles of LacZ- or EL-Ad. The A) PC- B) PE- C) LPC- D) LPE- and E) TG-profile was determined in lipid extracts of 200 mg tissue. Results represent the mean±SDof a representative experiment with 5 animals per group. Two additional experiments with 5 animals per group gave similar results.

1009M. Riederer et al. / Biochimica et Biophysica Acta 1821 (2012) 1003–1011

1010 M. Riederer et al. / Biochimica et Biophysica Acta 1821 (2012) 1003–1011

Choline compared with LacZ-Ad infected control cells, indicating at-tenuation of endogenous PC synthesis by EL overexpression.

3.5. EL alters the lipid profile of mouse plasma and liver

To address the impact of EL on plasma and tissue lipid profile invivo, EL was overexpressed in mice by i.v. EL-Ad injection. In plasma,both the PC and the PE content were strikingly decreased upon ELoverexpression (Fig. 5A,B). While the plasma concentration of LPCwas markedly decreased upon EL overexpression (Fig. 5C), the rela-tive amounts of LPC, normalized to total plasma PC, were highly in-creased upon EL overexpression (Fig. 5D). Very low LPE plasmalevels precluded their accurate quantification (not shown). In theliver, EL overexpression resulted in unaltered PC (Fig. 6A) and PE(Fig. 6B), increased LPC species 22:6, 18:2, and 20:4 (Fig. 6C) andLPE species 20:4 and 18:2 (Fig. 6D) as well as increased TG (Fig. 6E,Supplementary Table 2 for detailed list of species).

4. Discussion

reviously, we demonstrated that EL by virtue of its sn-1 phospholi-pase and lysophospholipase activity cleaves HDL-PC, thereby generat-ing saturated and unsaturated FFA [7], which are efficientlyincorporated into TG and PL of EL overexpressing cells [7,14]. Due toits rather weak lysophospholipase activity, compared to its sn-1 phos-pholipase activity, ELs' action on HDL-PC also generates substantialamounts of various, mostly unsaturated sn-1-lyso, sn-2-acyl-lysoPL[7]. In the present study, we monitored the generation of LPC and LPEin EL overexpressing cells as well as in EL overexpressing mice and ex-amined the impact of EL overexpression on cellular PL and TG contentand composition by MS.

The action of EL on HDL yielded a variety of LPC and LPE species inmedium. Due to an approximately 60-fold higher amount of PC thanPE in HDL3 (Supplementary Fig. 5A,B), the amounts of EL-generatedLPC exceeded that of LPE by approximately 30-fold (compare Fig. 1Eand F). While the EL-mediated decrease in the PC content of the incu-bation media did not reach statistical significance (Fig. 1A), the PCcontent was significantly decreased in HDL re-isolated from EL incu-bations compared with LacZ (Fig. 1B). This more pronounced effectof EL on the reisolated HDL-PC content might be due to the presenceof phospholipids not associated with HDL (soluble or bound to pro-teins secreted by HAEC), and accordingly less accessible for EL, inthe incubation media. The more pronounced EL-mediated decreasein HDL-PE compared to -PC might be due to the fact that PE is a bettersubstrate for EL than PC [9]. Due to endpoint measurements, our ex-perimental approach excludes the possibility to monitor dynamic bi-directional flux of lipids between HDL and cells. However, it is likelythat EL-modified HDL, depleted in PL, exhibits increased capacity toacquire lipids, including PL, from cells. Furthermore, in addition to hy-drolysis of HDL-PC and -PE, the EL-mediated bridging of HDL to thecell surface [6], might further contribute to the exchange of lipids be-tween HDL and cells.

In line with a profound EL-mediated provision of LPC, the total cel-lular PC content was increased (Fig. 2A), despite decreased endoge-nous PC synthesis (Fig. 4). Compromised de novo PC synthesis in ELoverexpressing cells might be due to FFA provision by EL [7,24]with concomitantly increased TG synthesis and DAG depletion [18].Indeed, both incorporation of exogenous FFA into TG (Fig. 2F) andthe TG content (Fig. 2D,E; Supplementary Table 1) were markedly in-creased in EL overexpressing HAEC. In addition to supplying FFA, theEL-mediated HDL holoparticle uptake [6,25] might supply cells withHDL-TG, thus contributing to the expansion of the TG pool in EL over-expressing cells. Moreover, attenuation of CT activity by LPC [26]might, at least in part, contribute to the observed attenuation of thede novo PC synthesis in EL overexpressing cells.

Tracer experiments directly demonstrated the capacity of EL tosupply cells with HDL-derived LPC (Fig. 3A). As found in HUVEC[27] reacylation of exogenous LPC is operative in HAEC as well(Fig. 3B). Considering the fact that in aqueous medium at neutralpH, acyl chains rapidly migrate from the sn-2 to the deacylated sn-1position to give a more stable intermediate [28], we assume that EL-generated sn-2-acyl LPC are rapidly converted to sn-1-acyl isomers,followed by reacylation, as found for exogenously added [14C]-sn-1acyl LPC.

In vivo, in mice injected intravenously with EL-adenovirus, amarked decrease in plasma PC and PE suggests a profound generationof LPC and LPE. However, we observed decreased plasma concentra-tions of LPC and undetectable levels of LPE upon EL overexpression.This is conceivably a consequence of a rapid removal of LPC and LPEfrom plasma to tissues, exemplified by concomitantly increased LPCand LPE in the liver. Furthermore, the substrate availability (PC andPE) for EL and various other plasma phospholipases is markedly de-creased by EL overexpression, resulting in markedly decreasedsteady-state LPC and LPE plasma levels despite their overproduction.Along these lines, the relative abundance of plasma LPC (normalizedto total plasma PC), was markedly increased upon EL overexpression(Fig. 5D).

In contrast to EL overexpressingHAEC, the PC contentwas unalteredinmouse liversmost probably due to the capability of hepatocytes to se-cret excess PC into bile and circulation, upon incorporation intolipoproteins.

Based on our results, it is tempting to speculate that by providingLPE and LPC for LPCAT-catalyzed reacylation reactions, EL helps to cir-cumvent the energy-consuming de novo PL synthesis. Thus, by savingcellular energy and meeting the demand for PL, EL might support en-dothelial cell proliferation, migration and angiogenesis, physiologicalstates with an increased demand for energy and PL [29]. Interestingly,EL expression has been found to be upregulated in HUVEC undergo-ing tube formation on matrigel, compared with growth-arrestedHUVEC in monolayers [2].

We conclude that EL not only supplies cells with FFA as found pre-viously [7,14], but also with HDL-derived LPC and LPE species. This ELmediated supply with HDL-derived lipids results in the accumulationof cellular TG, and PC, accompanied by decreased endogenous PCsynthesis.

Supplementary materials related to this article can be found on-line at doi:10.1016/j.bbalip.2012.03.006.

Grants

This work was supported in part by the Austrian Science Founda-tion (FWF; grant P19473-B05 to S.F.), the Jubilee Foundation of theAustrian National Bank (grant 12778 to S.F.) and the Lanyar Founda-tion (grant 328 to S.F.), which had no roles in the study design, collec-tion, analysis and interpretation of data, writing report or submissionof the article.

Acknowledgements

We thank Isabella Hindler for help with the care of the mice.

References

[1] M. Jaye, K.J. Lynch, J. Krawiec, D. Marchadier, C. Maugeais, K. Doan, V. South, D.Amin, M. Perrone, D.J. Rader, A novel endothelial-derived lipase that modulatesHDL metabolism, Nat. Genet. 21 (1999) 424–428.

[2] K. Hirata, H.L. Dichek, J.A. Cioffi, S.Y. Choi, N.J. Leeper, L. Quintana, G.S. Kronmal,A.D. Cooper, T. Quertermous, Cloning of a unique lipase from endothelial cells ex-tends the lipase gene family, J. Biol. Chem. 274 (1999) 14170–14175.

[3] M. Gauster, A. Hrzenjak, K. Schick, S. Frank, Endothelial lipase is inactivated uponcleavage by the members of the proprotein convertase family, J. Lipid Res. 46(2005) 977–987.

1011M. Riederer et al. / Biochimica et Biophysica Acta 1821 (2012) 1003–1011

[4] W. Jin, I.V. Fuki, N.G. Seidah, S. Benjannet, J.M. Glick, D.J. Rader, Proprotein conver-tases [corrected] are responsible for proteolysis and inactivation of endothelial li-pase, J. Biol. Chem. 280 (2005) 36551–36559.

[5] I.V. Fuki, N. Blanchard, W. Jin, D.H. Marchadier, J.S. Millar, J.M. Glick, D.J. Rader,Endogenously produced endothelial lipase enhances binding and cellular pro-cessing of plasma lipoproteins via heparan sulfate proteoglycan-mediated path-way, J. Biol. Chem. 278 (2003) 34331–34338.

[6] J.G. Strauss, R. Zimmermann, A. Hrzenjak, Y. Zhou, D. Kratky, S. Levak-Frank, G.M.Kostner, R. Zechner, S. Frank, Endothelial cell-derived lipase mediates uptake andbinding of high-density lipoprotein (HDL) particles and the selective uptake ofHDL-associated cholesterol esters independent of its enzymic activity, Biochem.J. 368 (2002) 69–79.

[7] M. Gauster, G. Rechberger, A. Sovic, G. Horl, E. Steyrer, W. Sattler, S. Frank, Endo-thelial lipase releases saturated and unsaturated fatty acids of high density lipo-protein phosphatidylcholine, J. Lipid Res. 46 (2005) 1517–1525.

[8] M.G. McCoy, G.S. Sun, D. Marchadier, C. Maugeais, J.M. Glick, D.J. Rader, Character-ization of the lipolytic activity of endothelial lipase, J. Lipid Res. 43 (2002)921–929.

[9] S. Chen, P.V. Subbaiah, Phospholipid and fatty acid specificity of endothelial li-pase: potential role of the enzyme in the delivery of docosahexaenoic acid(DHA) to tissues, Biochim. Biophys. Acta 1771 (2007) 1319–1328.

[10] W. Jin, G.S. Sun, D. Marchadier, E. Octtaviani, J.M. Glick, D.J. Rader, Endothelialcells secrete triglyceride lipase and phospholipase activities in response to cyto-kines as a result of endothelial lipase, Circ. Res. 92 (2003) 644–650.

[11] K.O. Badellino, M.L. Wolfe, M.P. Reilly, D.J. Rader, Endothelial lipase is increased invivo by inflammation in humans, Circulation 117 (2008) 678–685.

[12] K.O. Badellino, M.L. Wolfe, M.P. Reilly, D.J. Rader, Endothelial lipase concentra-tions are increased in metabolic syndrome and associated with coronary athero-sclerosis, PLoS Med. 3 (2006) e22.

[13] M.E. Paradis, K.O. Badellino, D.J. Rader, Y. Deshaies, P. Couture, W.R. Archer, N.Bergeron, B. Lamarche, Endothelial lipase is associated with inflammation inhumans, J. Lipid Res. 47 (2006) 2808–2813.

[14] J.G. Strauss, M. Hayn, R. Zechner, S. Levak-Frank, S. Frank, Fatty acids liberat-ed from high-density lipoprotein phospholipids by endothelial-derived li-pase are incorporated into lipids in HepG2 cells, Biochem. J. 371 (2003)981–988.

[15] E.P. Kennedy, S.B. Weiss, The function of cytidine coenzymes in the biosynthesisof phospholipides, J. Biol. Chem. 222 (1956) 193–214.

[16] W.E. Lands, Metabolism of glycerolipides; a comparison of lecithin and triglycer-ide synthesis, J. Biol. Chem. 231 (1958) 883–888.

[17] H. Shindou, T. Shimizu, Acyl-CoA:lysophospholipid acyltransferases, J. Biol. Chem.284 (2009) 1–5.

[18] C.L. Yen, S.J. Stone, S. Koliwad, C. Harris, R.V. Farese Jr., Thematic review series:glycerolipids. DGAT enzymes and triacylglycerol biosynthesis, J. Lipid Res. 49(2008) 2283–2301.

[19] R. Zimmermann, J.G. Strauss, G. Haemmerle, G. Schoiswohl, R. Birner-Gruenberger, M. Riederer, A. Lass, G. Neuberger, F. Eisenhaber, A. Hermetter, R.Zechner, Fat mobilization in adipose tissue is promoted by adipose triglyceride li-pase, Science 306 (2004) 1383–1386.

[20] P.H. Pritchard, D.E. Vance, Choline metabolism and phosphatidylcholine biosyn-thesis in cultured rat hepatocytes, Biochem. J. 196 (1981) 261–267.

[21] E.G. Bligh, W.J. Dyer, A rapid method of total lipid extraction and purification, Can.J. Biochem. Physiol. 37 (1959) 911–917.

[22] J. Hartler, M. Trotzmuller, C. Chitraju, F. Spener, H.C. Kofeler, G.G. Thallinger, Lipiddata analyzer: unattended identification and quantitation of lipids in LC–MS data,Bioinformatics 27 (2011) 572–577.

[23] O.H. Lowry, N.J. Rosebrough, A.L. Farr, R.J. Randall, Protein measurement with theFolin phenol reagent, J. Biol. Chem. 193 (1951) 265–275.

[24] T. Yasuda, T. Ishida, D.J. Rader, Update on the role of endothelial lipase inhigh-density lipoprotein metabolism, reverse cholesterol transport, and ath-erosclerosis, Circ. J. 74 (2010) 2263–2270.

[25] N. Nijstad, H. Wiersma, T. Gautier, M. van der Giet, C. Maugeais, U.J. Tietge, Scav-enger receptor BI-mediated selective uptake is required for the remodeling of highdensity lipoprotein by endothelial lipase, J. Biol. Chem. 284 (2009) 6093–6100.

[26] K.P. Boggs, C.O. Rock, S. Jackowski, Lysophosphatidylcholine and 1-O-octadecyl-2-O-methyl-rac-glycero-3-phosphocholine inhibit the CDP-choline pathway ofphosphatidylcholine synthesis at the CTP:phosphocholine cytidylyltransferasestep, J. Biol. Chem. 270 (1995) 7757–7764.

[27] L.L. Stoll, H.J. Oskarsson, A.A. Spector, Interaction of lysophosphatidylcholine withaortic endothelial cells, Am. J. Physiol. 262 (1992) H1853–1860.

[28] A. Pluckthun, E.A. Dennis, Acyl and phosphoryl migration in lysophospholipids:importance in phospholipid synthesis and phospholipase specificity, Biochemis-try 21 (1982) 1743–1750.

[29] J.M. Carter, L. Demizieux, R.B. Campenot, D.E. Vance, J.E. Vance, Phosphatidylcho-line biosynthesis via CTP:phosphocholine cytidylyltransferase 2 facilitates neuriteoutgrowth and branching, J. Biol. Chem. 283 (2008) 202–212.