3208 Mol. BioSyst., 2012, 8, 3208–3215 This journal is c The Royal Society of Chemistry 2012

Cite this: Mol. BioSyst., 2012, 8, 3208–3215

A cellular lipidomic study on the Ab-induced neurotoxicity and

neuroprotective effects of EGCG by using UPLC/MS-based glycerolipids

profiling and multivariate analysisw

Hongyang Zhang,zab Jing-Rong Wang,zac Lee Fong Yau,c Hing Man Ho,c

Chi Leung Chan,cPing Hu,

bLiang Liu*

acand Zhi-Hong Jiang*

ac

Received 1st April 2012, Accepted 12th September 2012

DOI: 10.1039/c2mb25126d

The aim of this study was to investigate the cellular lipid metabolism associated with b-amyloid

peptide (Ab)-induced neurotoxicity as well as the neuroprotective effect of (�)-epigallocatechingallate (EGCG), a major polyphenol in green tea. An ultra-performance liquid chromatography-

quadrupole-time-of-flight mass spectrometry (UPLC-Q-TOF-MS)-based lipidomic approach was

developed to screen and identify changes of the glycerolipids (GL) upon Ab treatment with or

without the presence of EGCG in PC12 cells. Principle component analysis (PCA) showed that

the Ab-treated group was well separated from the control group, whereas the EGCG group was

closer to the control group. The GL levels were significantly elevated in Ab-treated cells

compared with the control group, but were restored near to normal levels after EGCG treatment.

The elevated phosphatidylcholines (PCs) levels observed in the Ab-treated PC12 cells were quite

probably the integrated results of the reduced phospholipase A2 (PLA2) activity and the enhanced

activity of lysophospholipid acyltransferases. Moreover, an increased liberation of arachidonic

acid (AA) from PCs was observed as another important response of PC12 cells to the Abaggregates, implying an active inflammatory process occurring in Ab induced neurotoxicity.

EGCG treatment can reverse the deregulated metabolism of PCs, which might be one of the

biochemical mechanisms contributing to its neuroprotective effect. Collectively, results obtained

from the current lipidomic analyses of PC12 cells provided important insight into the biochemical

mechanisms underlying Ab-induced neurotoxicity and neuro protective effects of EGCG. This is

the first report of the lipidomic study on the neuroprotective effect of EGCG.

Introduction

Alzheimer’s disease (AD) is the most common neurodegenerative

disorder that affects a large number of the elderly population

worldwide.1 It is estimated that about 20 million people suffer from

AD-attributable dementia, with 20% of the individuals above

75 years old at risk of developing the disease. Evidence drawn

primarily from neuropathology, genetics and biophysics has

shaped the well-known ‘‘amyloid hypothesis’’, which implicates

the causative role of b-amyloid (Ab) peptide aggregation in AD.1

Numerous studies have aimed at interfering with amyloid

aggregation as an approach for the prevention or treatment

of AD.2

Recently, substantial evidence from epidemiological, animal,

and cellular studies has suggested a close association between lipid

metabolism and neurodegeneration.3–5 Biochemical studies have

revealed decreased levels of one or more classes of brain phospho-

lipids and increased levels of phospholipid breakdown products in

cerebral cortical areas of postmortem AD brain, suggesting an

important role of damage to brain membrane phospholipids in the

pathogenesis of AD.6 Further studies have suggested alterations in

membrane lipid composition associated with Ab.7 On the other

hand, amyloid peptide has a physiological function in regulating

lipid homeostasis in return.8 These studies provided evidence for a

complex and dual lipid-AD link, suggesting a potential lipid-based

causal approach for treating AD.

Human epidemiological and in vivo data suggested that

drinking green tea may decrease the incidence of AD.9

(�)-Epigallocatechin gallate (EGCG, Scheme 1), the polyphenol

found in large amounts in green tea, has been shown to possess

a State Key Laboratory of Quality Research in Chinese Medicine,Macau Institute for Applied Research in Medicine and Health,Macau University of Science and Technology, Macau, China.E-mail: zhjiang@must.edu.mo, lliu@must.edu.mo;Fax: +86 853 28825886; Tel: +86 853 88972777

b School of Chemistry & Molecular Engineering, East-ChinaUniversity of Science & Technology, Shanghai, China

c School of Chinese Medicine, Hong Kong Baptist University,Kowloon Tong, Hong Kong, Chinaw Electronic supplementary information (ESI) available. See DOI:10.1039/c2mb25126dz These authors contributed equally to the work.

MolecularBioSystems

Dynamic Article Links

www.rsc.org/molecularbiosystems PAPER

Dow

nloa

ded

by L

aure

ntia

n U

nive

rsity

on

12 M

arch

201

3Pu

blis

hed

on 0

3 O

ctob

er 2

012

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/C2M

B25

126D

View Article Online / Journal Homepage / Table of Contents for this issue

This journal is c The Royal Society of Chemistry 2012 Mol. BioSyst., 2012, 8, 3208–3215 3209

neuroprotective activities in a wide array of cellular and animal

models of neurological disorders.10 It has become clear that the

mechanism underlying the neuroprotective effect of EGCG goes

beyond its antioxidant activity, and might be at least partially

associated with their protection against Ab-induced toxicity.11–13

However, although several studies on the mechanisms of the

protective effect of EGCG against Ab-induced toxicity were

carried out,14 no attempt has been made to reveal the association

of such neuroprotective effect with the modulating effect of

EGCG on lipid metabolism, which has been recognized and

demonstrated by recent experimental studies.15,16

Lipidomics, a newly and rapidly expanding research field

that studies lipidomes and the organizational hierarchy of

lipids, has been proven to be a global approach for lipid

analysis in biological systems.17 It has been driven by rapid

advances in technologies including liquid chromatography

coupled with mass spectrometry (LC-MS), nuclear magnetic

resonance (NMR), and computational methods, as well as

increasingly appreciated role of lipids in many diseases such as

obesity,18 atherosclerosis,19 hypertension,20 and diabetes.21

Several lipidomic studies on AD have been carried out and

some potential lipid markers have been discovered.22,23 However,

since most of these studies used the brain of an AD patient for

lipid analysis, several blemishes still existed. First, due to multiple

cell types in the brain, varying degree of the cross presence of gray

and white matter during sampling may lead to an unpredictable

variable, which may overshadow real differences between the

samples from diseased and normal states.22 Second, metabolomics

of whole organism reflected holistic biology of human, thus

understanding of functions of specific cell types cannot be

obtained. Cellular lipidomics, which provide a metabolic profile of

individual cell types, are much easier to control, with less individual

variations, thus can provide more constant backgrounds against

which more subtle metabolic changes become apparent.24 Further-

more, cellular lipidomics provide relevant information about cell

functions of specific cell types, and therefore can be complementary

to the brain lipid studies on AD.

With an aim at addressing the global alteration of GL levels

in the Ab-treated neuron-like cells, and understanding the

mechanisms underlying Ab-induced neurotoxicity and neuro-

protective effect of natural products, a cellular lipidomics

investigation was carried in the current study. GL profiles

in the Ab-treated PC12 cells with or without the presence

of EGCG were characterized by employing the newly devel-

oped ultra-performance liquid chromatography-quadrupole-

time-of-flight mass spectrometry (UPLC-Q-TOF-MS) based

lipidomic approach. Multivariate statistical methods, including

principle component analysis (PCA) and partial least squares-

discriminant analysis (PLS-DA) were further employed for

differentiating the different groups and screening of potential

lipid markers.

Materials and methods

Chemicals and materials

The b-amyloid peptide (25–35, Ab25–35) trifluoroacetate salt

was obtained from Bachem (Bubendorf, Switzerland). 1,2-

Ditetradecanoyl-sn-glycero-3-phosphocholine (DMPC, PC

14 : 0/14 : 0) was purchased from Avanti Polar Lipids

(Alabaster, AL, USA). HPLC-grade acetonitrile, methanol,

chloroform, and isopropanol were purchased from Merck

(Darmstadt, Germany). Ultrapure water (18.2 MO) was

purified with a Milli-Q system (Millipore, MA, USA). EGCG

was obtained from Shanghai Ronghe Science Technology Ltd

(Shanghai, China). 3-[4,5-Dimethylthiazol-2-yl]-2,5-diphenyl-

tetrazolium bromide (MTT) was obtained from Sigma-Aldrich

(St. Louis, MO, USA). Rat pheochromocytoma PC12 cells

were obtained from the American Type Tissue Collection

(ATCC, Manassas, VA, USA). All other chemicals were of

analytical grade.

Cell culture and treatment

Rat pheochromocytoma PC12 cells were cultured in F-12K

medium (ATCC, Manassas, VA, USA) containing 15% horse

serum (Gibcos, New Zealand) and 2.5% FBS (Gibcos, New

Zealand) at 37 1C in a 5% CO2/95% air atmosphere. For

assessment of cell viability, PC12 cells were seeded in a 96-well

plate coated with collagen I (Gibcos, New Zealand) at a

density of 6 � 104 cells per well and allowed to adhere for 16 h

before treatment. Different concentrations of EGCG (1, 5, 10,

25, 50, 100 mM) were added to the cells of EGCG group 2 h

prior to the addition of Ab25–35 which was freshly dissolved in

distilled water on the day of the experiment.9 After subsequent

incubation for 24 h, MTT solution (10 mL per well, 5 mg ml�1

solution) was added to each well and then incubated for 4 h at

37 1C, after that 100 mL lysing sodium dodecyl sulfate (SDS)

was added and kept at room temperature overnight. The

optical densities were determined at 570 nm using a microplate

reader. Each experiment was carried out for three times.

For lipidomic analysis, PC12 cells were seeded in a 12-well

plate under the same conditions and divided into the following

three groups (each n = 7): the control, Ab-treated, and

EGCG-treated groups. Lipids were extracted from cells at

24 h after Ab25–35 addition. Plating controls seeded and

cultured at the same time were counted immediately prior to

lipid extraction to establish final cell numbers.

GL extraction

GLs were extracted according to the methods reported by

Figeys et al.25 with minor modifications. Briefly, medium was

removed from each plate and the cells were washed extensively

with 10 mM phosphate buffered saline. Cultures of PC12 cells

were placed on ice, and 1 mL of ice-cold methanol acidified

with 2% acetic acid was added to each well. Cells were then

Scheme 1 Chemical structure of (-)-epigallocatechin gallate (EGCG).

Dow

nloa

ded

by L

aure

ntia

n U

nive

rsity

on

12 M

arch

201

3Pu

blis

hed

on 0

3 O

ctob

er 2

012

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/C2M

B25

126D

View Article Online

3210 Mol. BioSyst., 2012, 8, 3208–3215 This journal is c The Royal Society of Chemistry 2012

scraped and collected in acidified methanol. Before extraction,

10 mL of 25 mg mL�1 internal standard solution (DMPC, non-

naturally occurring in mammalian cells) was added. Seven

wells per condition were harvested for lipid extraction. Lipids

were extracted using 0.95 mL of chloroform and 0.8 mL of

0.1 M sodium acetate in acid-washed glass tubes with PTFE

coated tops. After layer separation by centrifugation, 0.85 mL of

the lower organic fractions were collected, evaporated under a

gentle stream of nitrogen gas, and the residue was reconstituted

in 200 mL of methanol. The solutions were filtered through a

0.22 mm syringe filter (Millipore) into autosampler vials (Waters)

and stored at �80 1C before analysis.

UPLC-MS analysis

Chromatographic separation was performed using a Waters

ACQUITY UPLCt system (Waters Corp., Milford, USA)

equipped with a binary solvent delivery system and an auto-

sampler. A Waters ACQUITY HSS C18 column (100 �2.1 mm, 1.7 mm) was used to separate the endogenous lipids.

The mobile phase consisted of (A) acetonitrile–water (40 : 60,

v/v) and (B) acetonitrile–isopropanol (10 : 90, v/v), both

containing 10 mM ammonium acetate. A linear gradient was

optimized as follows (flow rate, 0.3 mL min�1): 0–4 min, 50%

B to 70% B; 4–15 min, 70% B to 80% B; 15–18 min, 80% B to

100% B; 18–19 min, 100% B to 100% B; 19–20 min, 100% B

to 50% B; 20–25 min, equilibration with 50% B. 2 mL of

sample solution was injected in each run. The column and

auto-sampler were maintained at 50 1C and 4 1C, respectively.

Mass spectrometry was performed using a MicroTOF-Q

mass spectrometer (Bruker Daltonics Inc., MA, USA)

equipped with an electrospray ion source. The system was

operated in positive ion mode and the optimized MS condi-

tions were as follows: capillary voltage 4500 V, end plate offset

voltage �500 V, capillary temperature 200 1C, drying gas flow

4.0 L min�1, drying gas temperature 200 1C, nebulizing gas

pressure 0.4 bar. For targeted MS/MS scan, collision energy

was set as 20 eV. Data were collected in centroid mode from

100–1200 m/z. For accurate mass measurement, the TOF mass

spectrometer was calibrated routinely using sodium formate

solution infused at a flow rate of 2.5 mL h�1. Sodium formate

solution was prepared by mixing 0.05 mM NaOH solution

with 0.05% formic acid in 90 : 10 proportion of 2-propanol/

distilled water.

Samples were analyzed in a random sequence in order to

reduce systemic error. Each sample was analyzed twice and the

average data were used for subsequent processing. Aliquots of all

samples were mixed together to serve as a pooled quality control

(QC) sample. Seven QC samples were evenly inserted into the

analytical sequence to monitor the analytical performance.

Data processing

The obtained CDF format data were firstly converted to

a MarkerLynx RAW format file through Waters Dbridge

software, and then exported intoMarkerlynx software (Version 4.1,

Waters, USA), where the data were applied to external scaling with

cell numbers, normalization with the peak intensity of internal

standard, peak alignment and data filtering. The resultant

three-dimensional matrix comprising compounds indices

(retention time and m/z) and ion intensities (multiplied by

10 000) were then imported into software SIMCA-P + 12.0

(Umetrics, Umea, Sweden). The data were preprocessed using

mean-centering and Pareto-scaling prior to multivariate statistical

analysis. Principal component analysis (PCA) was used to visualize

general clustering among groups. Partial least-squares-discriminant

analysis (PLS-DA) was carried out to identify the differentially

expressed GLs responsible for the separation. Statistical analyses of

the differentially expressed metabolites were performed by using

Microsoft Office Excel 2003.

Results and discussion

Ab-induced toxicity and neuroprotective effect of EGCG

It has been demonstrated that Ab25–35 contributes significantlyto the expression of aggregation and neurotoxicity in the full

length Ab protein due to its putative b-turn site and a

hydrophobic domain.26 As the proposed active fragment of

Ab, this peptide has been extensively applied in assessing the

neurotoxicity associated with Ab aggregation. As shown in

Fig. 2, Ab25–35 dissolved in H2O (10 mM) induced significant

toxicity to PC12 cells (about 60–70% decrease in MTT values

vs. control). The toxic effect of Ab25–35 was reduced in a dose-

dependent manner by a co-treatment with EGCG. The effects

were significant at 5–100 mM and maximal at 25 mM (Fig. 1a).

The decrease in the protective effect at higher dosage resulted

from the cytotoxic of EGCG to PC12 cells, which has been

indicated in previous report and demonstrated in our

Fig. 1 Ab25–35-induced toxicity and protective effect of EGCG in

PC12 cells. Results obtained from MTT assay with (a) or without Abtreatment (b); results obtained from counting of the cells for lipid

extraction (c).

Dow

nloa

ded

by L

aure

ntia

n U

nive

rsity

on

12 M

arch

201

3Pu

blis

hed

on 0

3 O

ctob

er 2

012

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/C2M

B25

126D

View Article Online

This journal is c The Royal Society of Chemistry 2012 Mol. BioSyst., 2012, 8, 3208–3215 3211

experiment (Fig. 1b). To observe the maximal effect of EGCG

on the lipid metabolism, 25 mM was selected for subsequent

lipidomic analysis. In the experiment for lipid extraction, the

cells seeding in a 12-well plate showed similar proliferation

properties with that in a 96-well plate, as confirmed by the cell

counting of the control plate (Fig. 1c).

Cellular GL profiling

High efficient extraction is crucial for the cellular GL analysis.

A classic Bligh/Dyer method modified with acidified methanol

was employed to optimally extract GLs without causing cleavage.

Compared with the Figeys method, we performed single extrac-

tion so as to minimize the variations generated during extraction.

A test of 4 sequential extractions was carried out to evaluate the

maximum recovery. The extracts of each cycle were analyzed and

eight GLs belonging to different lipid classes were selected to

evaluate the extraction efficiencies. For each class of lipids,

representative lipids with different abundance levels (low-, middle-

, and high-abundance) were selected so as to roundly examine the

recovery (Fig. S1 and S2, ESIw). Over 90% of each class were

recovered in the first extraction, with less than 10% being lost in

the subsequent extractions. This result suggests an acceptable

recovery of the initial extract described in our manuscript.

The UPLC system, beneficial from the sub-2 mm hybrid

columns, can produce sharper and more concentrated peaks,

thus providing a rapid, sensitive, and convenient method for

the analysis of a wide range of lipids in cells. To achieve a

better separation and peak shape of these hydrophobic com-

pounds, isopropanol and ammonium acetate in the mobile

phase with gradient elution were used. The optimized base

peak intensity (BPI) chromatogram of GLs extracted from the

PC12 cell was shown in Fig. 2.

For the analysis of GLs, positive mode was selected based

on the general higher abundant ions of lysophosphatidylcho-

lines (LPCs), platelet activating factors (PAFs), phosphatidyl-

cholines (PCs), and triacylglycerols (TGs) than that obtained

in negative mode. More than 1500 peaks were detected

consistently in most cell samples using our UPLC/MS analy-

tical approach. High resolution mass obtained from the TOF-

MS analysis (with resolution >15 000) facilitated tentative

identification of the GLs based on the accurate mass with less

than 5 ppm error. By matching the accurate mass of the GLs in

PC12 with that of GLs recorded in LIPIDMAPS (containing

3036 GLs as of Apr. 21, 2010)27 and published literature,28,29 a

total of 83 GLs were tentatively identified (see Table S1 in the

ESIw). These included LPCs or PAFs, PCs, and TGs.

Multivariate analysis

In this study, a two-component PCA score plot of UPLC/MS

data was used to visualize the general variation of GLs among

three groups. As illustrated in Fig. 3a, the first two-component

PCA model cumulatively accounted for 63.5% of the total

variance. The PCA scores plot showed a clear separation

between the control and the Ab group in the PC1 dimension,

indicating a prominent toxicity induced by Ab25–25. Upon the

treatment of EGCG, the cellular lipidomic profiles were found

to shift back near to the control group, reflecting a potential

protective effect of EGCG against Ab-induced neurotoxicity.

The tight location of QC samples in the scores plot of PCA

Fig. 2 Typical UPLC/MS base peak ion chromatogram (BPI) of the

glycerolipids extract from PC12 cell using electrospray ionization in

positive ion mode.

Fig. 3 Multivariate statistical results of lipidomic data: (a) PCA and

(b) PLS-DA scores plot of the control (blue dot, n = 7), Ab (red

triangle, n = 7), EGCG intervention group (dark diamond, n = 7),

and QC samples (green star, n = 7); (c) loadings plot of PLS-DA

model using Pareto scaling with mean centering (labeled ions were

potential GL markers responsible for group clustering).

Dow

nloa

ded

by L

aure

ntia

n U

nive

rsity

on

12 M

arch

201

3Pu

blis

hed

on 0

3 O

ctob

er 2

012

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/C2M

B25

126D

View Article Online

3212 Mol. BioSyst., 2012, 8, 3208–3215 This journal is c The Royal Society of Chemistry 2012

(also seen in PLS-DA) demonstrated that the reproducibility

of analytical platforms was guaranteed.

To maximize the variation and characterize the differentially

expressed GLs responsible for the separation of control, Aband EGCG groups, the preprocessed data matrices were

further analyzed using a PLS-DA model, a sophisticated

supervised method capable of establishing the optimal position

to place a discriminant surface that separates classes best and

generate the matrices of scores and loadings. A well-fitted two-

component PLS-DA model (R2X = 0.388, R2

Y = 0.965, Q2 =

0.846) was constructed (Fig. 3b) and clear separation of each

group was obtained. In the corresponding loading plot (Fig. 3c),

the distance of individual variables from the main cluster is

positively related to their influence on the group separation, that

means compounds far away from the main cluster would have

greater impact on the classification. To evaluate the contribution

of respective compounds (variable) to the separation, variable

importance in the projection (VIP) value of each compound was

calculated. Variables with larger VIP values were considered as

the candidate markers contributing potentially to the separation.

The 10 potential GL markers identified based on the VIP value

were presented in Table 1.

MS/MS identification of potential GL biomarkers

The targeted MS/MS analysis obtained by QTOF-MS provided

the accurate mass of both molecular and fragment ions, which

are informative for identification of GL markers. Taking m/z

760.6 (PC), 524.4 (LPC), and 874.8 (TG) as examples, the

identification procedures were illustrated step by step as follows.

As shown in Fig. 4a, the molecular ion of the target GL is

m/z 760.5856, suggesting a molecular formula of C42H83NO8P

(error 0.1 ppm). The most abundant daughter ion at m/z

184.0730 (calculate for C5H15NO4P, error 4.9 ppm) was

assigned to be the diagnostic fragment corresponding to the

polar head group of PCs, indicating that the target GLs was a

PC.30 Detailed analysis of the less abundant ions revealed

fragment ions generated from cleavage of sn-2 (m/z 496.3381,

calculated for C24H51NO7P, error 4.4 ppm) and sn-1 (m/z

522.3547, calculated for C26H53NO7P, error 2.5 ppm), as well

as ions of respective dehydration products at m/z 478.3273

(calculated for C24H49NO6P, error 5.2 ppm, dehydrated ion of

sn-2 cleavage) and m/z 504.3444 (calculated for C26H51NO6P,

error 2.0 ppm, dehydrated ion of sn-1 cleavage), suggesting the

Table 1 Identification of GLs contributing to the differentiation of control, Ab and EGCG group on the basis of accurate mass and MS/MS

No. tR (min)

MS fragments

Identified GL VIP value

Peak intensity (�e4)a

MS1 Characteristic MS2 Control Ab Ab + EGCG

1 7.73 760.5856 184.0730, 496.3381,504.3444, 478.3273,522.3547

PC 16 : 0/18 : 1 11.32 13.91 � 2.42 26.02 � 3.81b 12.50 � 1.64c

2 6.81 758.5695 184.0733, 496.3386,502.3282, 520.3373, 478.3276

PC 16 : 0/18 : 2 10.30 12.68 � 1.98 22.38 � 2.66b 11.27 � 1.14c

3 7.68 734.5686 184.0732, 496.3377, 478.3294 PC 16 : 0/16 : 0 8.77 7.84 � 1.32 14.14 � 2.26b 6.35 � 0.79c

4 7.77 786.5990 184.0732, 524.3710, 520.3379,506.3576, 502.3289

PC 18 : 0/18 : 2 8.27 7.23 � 1.21 13.62 � 2.00b 6.39 � 0.83c

5 18.02 874.7842 577.5176, 575.5023, 601.5181 TG 16 : 0/18 : 1/18 : 2 6.59 4.32 � 0.84 8.40 � 1.16b 3.77 � 0.84c

6 17.75 872.7685 599.5024, 573.4862, 577.5171 TG 16 : 0/18 : 1/18 : 3 6.20 3.55 � 0.77 7.31 � 1.34b 3.27 � 0.85c

7 6.85 784.5822 184.0732, 522.3534, 520.3386,502.3307, 504.3399

PC 18 : 1/18 : 2 6.09 4.55 � 0.82 8.17 � 1.42b 4.11 � 0.58c

8 18.26 876.7998 577.5182, 603.5331 TG 16 : 0/18 : 1/18 : 1 5.71 3.73 � 0.60 6.83 � 0.90b 3.41 � 0.65c

9 18.27 850.7872 577.5183, 551.5020 TG 16 : 0/16 : 0/18 : 1 5.38 4.55 � 0.87 7.22 � 0.79b 4.04 � 0.76c

10 2.95 524.3686 184.0726, 258.1088, 240.0991 LPC 18 : 0 5.29 0.30 � 0.04 0.57 � 0.17b 0.24 � 0.06c

a Peak intensities were external scaled and normalized with cell numbers and the intensity of internal standard. b Student’s T-test, P o 0.01, vs.

control group. c Student’s T-test, P o 0.01, vs. Ab-treated group.

Fig. 4 Structures and MS/MS spectra of representative biomarkers:

(a) PC 16 : 0/18 : 1, (b) LPC 18 : 0, and (c) TG 16 : 0/18 : 1/18 : 2.

Dow

nloa

ded

by L

aure

ntia

n U

nive

rsity

on

12 M

arch

201

3Pu

blis

hed

on 0

3 O

ctob

er 2

012

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/C2M

B25

126D

View Article Online

This journal is c The Royal Society of Chemistry 2012 Mol. BioSyst., 2012, 8, 3208–3215 3213

loss of 16 : 0 and 18 : 1 FA. Based on above information,

the target GL was identified as PC 16 : 0/18 : 1. Likewise, the

target GL at m/z 524.3686 (calculated for C26H55NO7P, error

5.7 ppm) was characterized to be a LPC which belongs to the

subclass of monoacylglycero-PCs, as suggested by the diagnostic

fragment ion at m/z 184.0726 and the sole fragment formed by

cleavage of the sn-1 ester bond (Fig. 4b). On the basis of the

fragment ions at m/z 258.1088 (calculated for C8H21NO6P, error

7.4 ppm) and the subsequent ion loss of H2O m/z 240.0991

(calculated for C8H19NO5P, error 4.2 ppm), this GL was identified

to be LPC 18 : 0.

The fragment ions of TGs have been well studied and

may be used to determine the length and composition of the

FA chains.31 As seen in Fig. 4c, the mass spectrum of the target

GL showed quasimolecule ions at m/z 874.7842 (calculated for

C55H104NO6, error 2.5 ppm), corresponding to the ([M+NH4]+

of TG 52 : 3. The fragment ions generated from the loss of 16 : 0

FA, 18 : 2 FA, and 18 : 1 FA can be observed at m/z 601.5181

(calculate for C39H69O4, error 2.5 ppm), 577.5176 (calculate for

C37H69O4, error 3.5 ppm), and 575.5023 (calculate for C37H67O4,

error 2.8 ppm), respectively. It should be noted that the

dissociation of the [M + NH4]+ ion of TG is different from

other lipids in that all their molecular species displayed the

characteristic loss of an acyl side-chain as a neutral free acid and

ammonia.29 For example, the loss of the 16 : 0 FA chain

corresponds to C16H35O2N [CH3(CH2)14CO2H+NH3]. Based

on the above evidence, the target GL was characterized to be

TG 16 : 0/18 : 1/18 : 2. By using this approach, 10 potential

GL biomarkers, including different classes of PCs, TGs, and

LPC, were identified as presented in Table 1.

Lipid alterations in Ab-induced neurotoxicity and

neuroprotection of EGCG

The multivariate analysis based on the lipidomics data showed

significant GL metabolic alterations in PC12 cells after the Abtreatment. In Table 1, 10 GL biomarkers were significantly

elevated in Ab-treated cells as compared to the control cells. It

can be seen that the most significant alterations resulted from Abinsult are the increased levels of PCs. Although some studies have

reported increased levels of glycerophosphocholine, phospho-

choline and choline, which are water-soluble metabolites of PCs, as

a consequence of activation of calcium-dependent phospholipase

A2 (cPLA2),32 evidence from other studies suggested decreased

PLA2 activity in postmortem parietal and frontal cortices of AD

patients, and this reduction was correlated with the severity of

dementia.33,34 Collective data showed that decreased membrane

phospholipid metabolism mediated by reduced activity of PLA2 in

the cerebral cortex and hippocampus is an early event in AD and

may contribute to cognitive dysfunction and neuropathology at

early stages of the disease.35 Ross et al. further demonstrated

the increased activity of lysophospholipid acyltransferases,

which catalyze the re-synthesis of PCs from LPCs, in AD

patients’ brain.6 The decreased activity of PLA2 and increased

activity of lysophospholipid acyltransferases would both result

in the accumulation of PCs (Fig. 5). The elevated PCs level

observed in the Ab-treated PC12 cells was quite possibly be the

integrated consequence of the reduced PLA2 activity and

enhanced activity of lysophospholipid acyltransferases.

Furthermore, recent researches also suggested the critical

role of arachidonic acid (AA) and specific PLA as mediators in

Ab-induced neurodegenerative pathogenesis.36 Various PLA

types, such as PLA1 and PLA2, can catalyze the hydrolysis at

sn-1 (PLA1) or sn-2 (PLA2) position of specific PCs to release

arachidonic acid and LPCs. The activity of PLA can be

estimated by calculating the ratios of LPC 16 : 0 and LPC

18 : 0 to PC 16 : 0/20 : 4 and PC 18 : 0/20 : 4.29,37 Therefore,

we differentiated the isomers of LPC 16 : 0 or PAF 14 : 0, PC

36 : 4, and PC 38 : 4 by targeted MS/MS (Table S2, ESIw)firstly. Then calculated the ratio of normalized response of

LPC 16 : 0 and LPC 18 : 0 to the response of PC 16 : 0/20 : 4

and PC 18 : 0/20 : 4, respectively (Table S2, ESIw). It could be

seen that both two ratios increased in Ab-treated groups

compared to that in controls (Fig. 6), plausibly indicating

enhanced activity of PLA catalyzing AA-containing PCs.38

Fig. 5 Phospholipid metabolism in Lands’ cycle and the possible effect

of Ab to the PLA2 and lysophospholipid acyltransferases (revised based

on D. Hishikawa et al., Proc. Natl. Acad. Sci. U. S. A., 2008, 105,

2830–2835).

Fig. 6 Peak intensity ratios of (LPC 16 : 0)/(PC 16 : 0/20 : 4) and

(LPC 18 : 0)/(PC 18 : 0/20 : 4) in different groups (control, Ab, andEGCG intervention group). The error bar spread indicates the standard

deviation of all results in each group (n = 7). ‘‘*’’ denote P o 0.05

(control vs. Ab-treated group); ‘‘D’’ and ‘‘DD’’ denote P o 0.05 and

P o 0.01, respectively (Ab-treated vs. Ab + EGCG group).

Dow

nloa

ded

by L

aure

ntia

n U

nive

rsity

on

12 M

arch

201

3Pu

blis

hed

on 0

3 O

ctob

er 2

012

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/C2M

B25

126D

View Article Online

3214 Mol. BioSyst., 2012, 8, 3208–3215 This journal is c The Royal Society of Chemistry 2012

This observation is partially consistent with previous results

and suggests that an enhanced PLA activity might be asso-

ciated with the AD process.39 Upon EGCG treatment, the

ratio of (LPC 18 : 0)/(PC 18 : 0/20 : 4) was reversed while the

ratio of (LPC 16 : 0)/(PC 16 : 0/20 : 4) was just opposite.

However, the ratio of (LPC 16 : 0)/(PC 16 : 0/20 : 4) is very

small (o0.10), implying a low-level hydrolysis of PC 16 : 0/

20 : 4, which in turn resulted in the release of only a small

amount of corresponding LPC and AA. In comparison, the

ratio of (LPC 18 : 0)/(PC 18 : 0/20 : 4) was quite high (almost

100 folds of the ratio of (LPC 16 : 0)/(PC 16 : 0/20 : 4)),

which might consequently yield a large amount of LPC

18 : 0 (almost 10 folds of LPC 16 : 0) and AA (Table S2,

ESIw). This phenomenon indicates that different PLA isomers

may be involved in the hydrolysis of PC 16 : /20 : 4 and

PC 18 : 0/20 : 4, and their activity was individually modu-

lated, or certain PLA enzymes involved possess specificity in

the hydrolysis of these two PCs. It can be speculated that the

hydrolysis of PC 18 : 0/20 : 4 is the primary factor of AA

release.

Because of the dominant effect of PC18 : 0/20 : 4 hydrolysis

in the release of AA, the reduced ratio (LPC 18 : 0)/(PC

18 : 0/20 : 4) in the EGCG group might suggest a protective

effect of EGCG attributable to the reduced activity of specific

PLAs.40 However, since more than ten groups of PLAs with

varied substrate specificity are involved in the lipid meta-

bolism,41 and also because the upstream and downstream

enzymes catalyze the synthesis of PC and further metabolism

of LPC, the level of LPC is not always linearly associated with

the activity of certain PLAs and the level of AA. Therefore, the

activity of PLAs and the level of AAs upon Ab treatment and

EGCG intervention need to be further elucidated. Even

though, our result provided further insight into the underlying

mechanism of EGCG’s protective effect and potential thera-

peutic approach for AD.

Conclusion

In the current work, a UPLC-MS-based cellular lipidomic

approach has been developed. Cellular GL profiles in the

Ab-treated PC12 cells with or without the presence of EGCG

were compared. The results showed that GL alterations were

extensively involved in the neurotoxicity induced by Abaggregates as well as protection of EGCG against such

neurotoxicity. We found that the level of PCs significantly

elevated in PC12 cells upon addition of Ab aggregates, this

might be resulted from the decreased activity of PLA2 and

enhanced activity of lysophospholipid acyltransferases

(Fig. 6). This result provided cellular evidence for the previous

studies in AD patients.36,37 In addition, increased liberation of

arachidonic acid from PCs was indicated to be an important

response of PC12 cells to the Ab aggregates in our study. This

is consistent with the hypothesis that there is an active

inflammatory process occurring in AD. EGCG treatment

can reverse the deregulated metabolism of PC, which might

be one of the biochemical mechanisms contributing to its

neuroprotective effect. Collectively, the results obtained from the

lipidomic analyses of PC12 cells provided important insight into

the biochemical mechanisms underlying Ab-induced neurotoxicity

and protective effects of EGCG. This is the first report of the

lipidomic study on the neuroprotective effect of EGCG.

Acknowledgements

This study was financially supported by Macao Science and

Technology Development Fund,MSAR (039/2011/A2 to ZHJ).

Notes and references

1 T. Hartmann, J. Kuchenbecker and M. O. W. Grimm, J. Neurochem.,2007, 103, 159–170.

2 F. G. De Felice, J. C. Houze, J. Garcia-Abreu, P. R. Jr Louzada,R. C. Afonso, M. N. Meirelles, R. Lent, V. M. Neto andS. T. Ferreira, FASEB J., 2001, 15, 1297–1299.

3 O. Ghribi, J. Alzheimer’s Dis., 2008, 15, 673–684.4 M. S. Lin, L. Y. Chen, S. S. Wang, Y. Chang and W. Y. Chen,Colloids Surf., B Biointerfaces, 2008, 65, 172–177.

5 M. Wakabayashi, T. Okada, Y. Kozutsumi and K. Matsuzaki,Biochem. Biophys. Res. Commun., 2005, 328, 1019–1023.

6 R. M. Ross, A. Moszczynska, J. Erlich and S. J. Kish,J. Neurochem., 1998, 70, 786–793.

7 X. L. Qi, J. Xiu, K. R. Shan, Y. Xiao and R. Gu, Neurochem. Int.,2005, 46, 613–621.

8 M. O. Grimm, H. S. Grimm, A. J. Patzold, E. G. Zinser,R. Halonen, M. Duering, J. A. Tschape, B. De Strooper,U. Muller, J. Shen and T. Hartmann, Nat. Cell Biol., 2005, 7,1118–1123.

9 S. Bastianetto, Z. X. Yao, V. Papadopoulos and R. Quirion, Eur.J. Neurosci., 2006, 23, 55–64.

10 S. Mandel, T. Amit, O. Weinreb, L. Reznichenko andM. B. Youdim, CNS Neurosci. Ther., 2008, 14, 352–365.

11 Y. Levites, T. Amit, S. Mandel and M. B. Youdim, FASEB J.,2003, 17, 952–954.

12 S. Bastianetto, Z. X. Yao, V. Papadopoulos and R. Quirion, Eur J.Neurosci., 2006, 23, 55–64.

13 Y. T. Choi, C. H. Jung, S. R. Lee, J. H. Bae, W. K. Baek,M. H. Suh, J. Park, C. W. Park and S. I. Suh, Life Sci., 2001,70, 603–614.

14 C. L. Lin, T. F. Chen, M. J. Chiu, T. D. Way and J. K. Lin,Neurobiol. Aging, 2009, 30, 81–92.

15 S. K. Patra, F. Rizzi, A. Silva, D. O. Rugina and S. Bettuzzi,J. Physiol. Pharmacol., 2008, 59, 217–235.

16 Y. Sun, W. C. Hung, F. Y. Chen, C. C. Lee and H. W. Huang,Biophys. J., 2009, 96, 1026–1035.

17 X. L. Han and R. W. Gross, Mass Spectrom. Rev., 2005, 24,367–412.

18 A. Gonzalez-Periz, R. Horrillo, N. Ferre, K. Gronert,B. Dong, E. Moran-Salvador, E. Titos, M. Martinez-Clemente,M. Lopez-Parra, V. Arroyo and J. Claria, FASEB J., 2009, 23,1946–1957.

19 K. Ekroos, M. Janis, K. Tarasov, R. Hurme and R. Laaksonen,Curr. Atheroscler. Rep., 2010, 12, 273–281.

20 J. Graessler, D. Schwudke, P. E. Schwarz, R. Herzog,A. Shevchenko and S. R. Bornstein, PLoS One, 2009, 4, e6261.

21 G. Schmitz and M. Hanefeld, Diabetes, 2008, 57, A26.22 X. L. Han, S. Rozen, S. H. Boyle, C. Hellegers, H. Cheng,

J. R. Burke, K. A. Welsh-Bohmer, P. M. Doraiswamy andR. Kaddurah-Daouk, PLoS One, 2009, 4, e21643.

23 X. L. Han, Biochim. Biophys. Acta, 2010, 1801, 774–783.24 M. Cuperlovic-Culf, D. A. Barnett, A. S. Culf and I. Chute, Drug

Discovery Today, 2010, 15, 610–621.25 S. N. Whitehead, W. M. Hou, M. Ethier, J. C. Smith,

A. Bourgeois, R. Denis, S. A. L. Bennett and D. Figeys, Anal.Chem., 2007, 79, 8539–8548.

26 C. J. Pike, A. J. Walencewicz-Wasserman, J. Kosmoski,D. H. Cribbs, C. G. Glabe and C. W. Cotman, J. Neurochem.,1995, 64, 253–265.

27 http://www.lipidmaps.org/; accessed Apr 21, 2010.28 C. X. Hu, J. van Dommelen, R. van der Heijden, G. Spijksma,

T. H. Reijmers, M. Wang, E. Slee, X. Lu, G. W. Xu, J. van derGreef and T. Hankemeier, J. Proteome Res., 2008, 7, 4982–4911.

Dow

nloa

ded

by L

aure

ntia

n U

nive

rsity

on

12 M

arch

201

3Pu

blis

hed

on 0

3 O

ctob

er 2

012

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/C2M

B25

126D

View Article Online

This journal is c The Royal Society of Chemistry 2012 Mol. BioSyst., 2012, 8, 3208–3215 3215

29 J. M. Castro-Perez, J. Kamphorst, J. DeGroot, F. Lafeber,J. Goshawk, K. Yu, J. P. Shockcor, R. J. Vreeken andT. Hankemeier, J. Proteome Res., 2010, 9, 2377–2389.

30 B. Brugger, G. Erben, R. Sandhoff, F. T. Wieland and W. D.Lehmann, Proc. Natl. Acad. Sci. U. S. A., 1997, 94, 2339–2344.

31 A. M. McAnoy, C. C. Wu and R. C. Murphy, J. Am. Soc. MassSpectrom., 2005, 16, 1498–1509.

32 A. Walter, U. Korth, M. Hilgert, J. Hartmann, O. Weichel,M. Hilgert, K. Fassbender, A. Schmitt and J. Klein, Neurobiol.Aging, 2004, 12, 1299–1303.

33 W. F. Gattaz, A. Maras, N. J. Cairns, R. Levy and H. Forstl, Biol.Psychiatry, 1995, 37, 13–17.

34 W. F. Gattaz, N. J. Cairns, R. Levy, H. Forstl, D. F. Braus andA. Maras, Neuroscience (Oxford), 1996, 246, 129–131.

35 R. O. Sanchez-Mejia, J. W. Newman, S. Toh, G. Q. Yu,Y. G. Zhou, B. Halabisky, M. Cisse, K. Scearce-Levie,

I. H. Cheng, L. Gan, J. J. Palop, J. V. Bonventre and L. Mucke,Nat. Neurosci., 2008, 11, 1311–1317.

36 R. O. Sanchez-Mejia and L. Mucke, Biochim. Biophys. Acta, 2010,1801, 784–790.

37 C. X. Hu, H. W. Kong, F. X. Qu, Y. Li, Z. Q. Yu,P. Gao, S. Q. Peng and G. W. Xu, Mol. Biosyst., 2011, 7,3271–3279.

38 M. Chalimoniuk, A. Stolecka, M. Caka"a, S. Hauptmann,K. Schulz, U. Lipka, K. Leuner, A. Eckert, W. E. Muller andJ. B. Strosznajder, Acta Biochim. Pol., 2007, 54, 611–623.

39 V. Frisardi, F. Panza, D. Seripa, T. Farooqui and A. A. Farooqui,Prog. Lipid Res., 2011, 50, 313–330.

40 A. A. Farooqui, M. L. Litsky, T. Farooqui and L. A. Horrocks,Brain Res. Bull., 1999, 49, 139–153.

41 J. E. Burke and E. A. Dennis, J. Lipid Res., 2009, 50,S237–S242.

Dow

nloa

ded

by L

aure

ntia

n U

nive

rsity

on

12 M

arch

201

3Pu

blis

hed

on 0

3 O

ctob

er 2

012

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/C2M

B25

126D

View Article Online