doi:10.1111/j.1440-1746.2011.06724.

GASTROENTEROLOGY

jgh_6724

1290..1297

Metabolomic phenotype of gastric cancer and precancerous

stages based on gas chromatography time-of-flight
mass spectrometry
Lianzhen Yu,*,1 Jiye Aa,,1 Jin Xu,* Min Sun, Sixuan Qian, Liping Cheng, Shuping Yang* and
Ruihua Shi*
Departments of *Gastroenterology and Hematology, the First Hospital Affiliated to Nanjing Medical University, and Key Laboratory of Drug
Metabolism and Pharmacokinetics, China Pharmaceutical University, Nanjing, and Department of Gastroenterology, the Zhangjiagang Hospital
Affiliated to Soochow University, Suzhou, and Department of Medicine, the Changzhou Tumor Hospital Affiliated to Soochow University,
Changzhou, China

Key words
gas chromatography time-of-flight mass
spectrometry, gastric cancer, metabolic
phenotype.
Accepted for publication 16 March 2011.
Correspondence
Dr Ruihua Shi, Department of
Gastroenterology, the First Hospital Affiliated
to Nanjing Medical University, 300 Guangzhou
Road, Nanjing, 210029, China. Email:
ruihuashi@126.com
1
These authors contributed equally to this
work.

Abstract
Background and Aim: To study the low-molecular-weight metabolites in blood plasma of
patients with the progressive disease, gastric cancer, and to characterize different stages
from chronic superficial gastritis (CSG) to chronic atrophic gastritis (CAG), intestinal
metaplasia (IM), gastric dysplasia (DYS) and finally gastric cancer (GC).
Methods: We applied gas chromatography time-of-flight mass spectrometry (GC/TOFMS) to determine metabolites levels in plasma obtained from 80 patients including 19 with
CSG, 13 with CAG, 10 with IM, 15 with DYS and 22 with GC (nine preoperation and 13
postoperation). Principal component analysis (PCA) and statistics were used to differentiate the stages and to identify the markers of gastric cancer.
Results: Totally, 223 peaks were detected in GC/TOF-MS and 72 compounds were
authentically identified. CSG showed distinct difference from the other groups of CAG,
IM, DYS and GC, whose plots clustered closely. IM clustered closely to GC, suggesting
similar metabolic patterns of them. Fifteen identified metabolites contributed most to the
differentiating between CSG and GC, and characterized different stages of GC. Statistics
revealed elevated levels of 2-Hydroxybutyrate, pyroglutamate, glutamate, asparagine,
azelaic acid, ornithine, urate, 11-eicosenoic acid, 1-monohexadecanoylglycerol and
g-tocopherol, while downregulation of creatinine, threonate in GC group, indicating that
GC patients were obviously involved in oxidative stress, and perturbed metabolism of
amino acids and fatty acids.
Conclusion: The metabolic phenotype of CSG is significantly different from GC, while
that of IM is similar to it. The discriminatory metabolites characterizing progressive stages
from CSG to GC might be the potential markers to indicate a risk of GC.

Introduction
Although the morbidity and mortality has decreased significantly
in the past few decades, gastric cancer (GC), one of the most
common malignant tumors, remains the fourth most common
malignancy and occupies the second position in terms of the
number of mortalities caused by cancer.1 It was predicted that, in
2005, 0.3 million deaths and 0.4 million new cases from gastric
cancer would rank the third most common cancer in China.2
Exploring of the mechanism of GC and identifying the markers for
early diagnosis is of significant importance and has a great potential to reduce the morbidity and mortality of GC.
To date, little is known of the gastric carcinogenesis since the
development of GC is involved in multi-gene and multi-factor.
1290

Correa3 proposed a model of human intestinal-type gastric carcinogenesis from normal mucosa to chronic superficial gastritis
(CSG), to chronic atrophic gastritis (CAG), to intestinal metaplasia (IM), to dysplasia (DYS), and finally to intestinal-type
gastric cancer, which is widely accepted. Atrophic gastritis and
intestinal metaplasia are considered as important risk factors of
gastric carcinogenesis,4 and early intervention on gastric precancerous lesions and precancerous diseases may reduce the incidence of gastric cancer. Unfortunately, ultimate diagnosis of
CAG, IM and GC primarily relies on endoscopic examination
and pathological section, which is painful and harmful. Identification of biomarkers to indicate a risk of GC and establishment
of a sensitive, reliable routine examination method is greatly
desired.

Journal of Gastroenterology and Hepatology 26 (2011) 12901297

2011 Journal of Gastroenterology and Hepatology Foundation and Blackwell Publishing Asia Pty Ltd

L Yu et al.

Metabolomic phenotype of gastric cancer

The inclusion or exclusion of the patient in the study was

assessed based on the entire body of data available. Inclusion
criteria included: (i) aged between 20 and 60 years, both male or
female; (ii) definite diagnosis pathologically and gastroscopically;
(iii) all patients were positive in Helicobacter pylori infection; and
(iv) the postoperative GC patients had been performed radical
operation for 4 to 6 weeks. Exclusion criteria were: (i) metabolic
disease, such as hyperlipoidemia, diabetes mellitus, gout; (ii) pregnant or lactating female; (iii) any symptom of acute disease during
the last 2 weeks, such as: febrile, cough, headache, diarrhea,
hematuria; (iv) momentous stress reaction during the last 2 weeks,
such as psychic trauma and a large area of empyrosis; (v) hematopathy, including all types of leukemia and anemia; (vi) use of
specific drugs during the last 2 weeks, such as antibiotic (quinolones, torlamician), hormone (deoxycortisone, dexamethasone),
nonsteroid anti-inflammatory drugs (NSAIDs; acenterine, saridon,
contac), among others; and (vii) medical treatments before blood
collection, such as radiotherapy or chemotherapy.
All of the subjects were examined by gastroscope and biopsied
for pathology. Each of the 80 subjects who were under fasting
conditions provided 3 mL venous blood in the First Hospital Affiliated to Nanjing Medical University. The blood was centrifuged at
1600 g for 10 min and 500 mL plasma was transferred to an Eppendorf tube, then stored at -80C until analysis. According to pathologic diagnosis, the 80 samples were divided into six groups: 19
chronic superficial gastritis (CSG), 13 chronic atrophic gastritis
(CAG), 10 intestinal metaplasia (IM), 15 dysplasia (DYS), 9 preoperative gastric cancer (GC) and 13 postoperative gastric cancer
(GC_p). The data for the patients included in this study are
detailed in Table 1.

Metabolomics is a novel technique that can quantitatively determine level change of small molecules in bio-samples resulting
from various pathologies, physiological changes and external
stimulation.58 Recently, metabolomics has been used to characterize the metabolic perturbation and identify potential biomarkers
in body fluids of various cancers.916 Denkert and colleagues9
profiled metabolic phenotype of colorectal cancer tissue and corresponding control by using gas chromatography time-of-flight
mass spectrometry (GC/TOF-MS), and revealed that intermediates
of citric acid cycle were downregulated and many amino acids
were upregulated in colorectal cancer tissue. Chan and colleagues10 compared the metabolic phenotype of biopsied colorectal cancer and their matched normal mucosae obtained from 31
colorectal cancer patients using high-resolution magic angle
spinning-nuclear magnetic resonance and gas chromatography
mass spectrometry (GC/MS), observed the elevated glycolysis,
nucleotide biosynthesis, lipid metabolism, inflammation and
steroid metabolism in malignant samples. Moreover, by applying
capillary electrophoresis time-of-flight mass spectrometry (CE/
TOFMS), Hirayama and colleagues11 determined extremely low
level of glucose, while highly elevated lactate, and glycolytic
intermediate, and most amino acids except glutamine, in both
colon and GC tissues. For identifying diagnostic markers of GC
and to explore the mechanism of the gastric carcinogenesis, in this
study, we investigated the dynamic changes of small molecule
metabolites in blood plasma of different stages in gastric carcinogenesis based on a well established metabolomic platform.1719

Methods
Patients and plasma samples

Chemicals and reagents

Altogether, 80 male and female individuals attending the First

Affiliated Hospital of Nanjing Medical University were eligible
for inclusion if they had been diagnosed with gastric diseases
between January 2007 and September 2008. All of the patients
provided their written informed consent before enrollment in the
study. The study was evaluated and approved by the Institutional
Review Board of the Nanjing Medical University.

Table 1

All the authentic reference compounds used for compound identification were of analytical grade, purchased from Sigma-Aldrich
or Sigma (St Louis, MO, USA), Aldrich (Steinhein, Germany),
Merck (Darmstadt, Germany), or Serva (Heidelberg, Germany).
Methoxyamine hydrochloride and the stable-isotope-labeled internal standard compound [13C2-]-myristic acid were purchased from

Patients characteristics in the study of gastric cancer and precancerous stages

Characteristic

Gender
Male
Female
Age(years)
Mean SD
Range
Drinking
Smoking
Taking medicine
H.pylori infection

Patients of diverse stages

CSG

CAG

IM

DYS

GC

GC_p

9
10

7
6

5
5

8
7

4
5

7
6

37 12.2
2060
+

42 14.1
2257
+

40 14.8
2155
+

38 16.3
2458
+

40 13.2
2255
+

43 13.0
2560
+

Not during the last 2 weeks. CAG, chronic atrophic gastritis; CSG, chronic superficial gastritis; DYS, gastric dysplasia; GC, gastric cancer; GC_p,
postoperative gastric cancer; IM, intestinal metaplasia.

Journal of Gastroenterology and Hepatology 26 (2011) 12901297

2011 Journal of Gastroenterology and Hepatology Foundation and Blackwell Publishing Asia Pty Ltd

1291

Metabolomic phenotype of gastric cancer

L Yu et al.

Sigma-Aldrich and Cambridge Isotope Laboratories, Inc.

(Andover, MA, USA), respectively. The alkane series (C8C40) and
pyridine (silylation grade) were obtained from Fluka (Buchs,
Switzerland). N-methyl-N-trimethylsilyl-trifluoroacetamide with
1% trimethylchlorosilane was purchased from Thermo Scientific
(Rockford, IL, USA). Methanol was of chromatography grade.
Distilled water was produced with a Milli-Q Reagent Water
System (Millipore, MA, USA).

Sample preparation and GC/TOF-MS analysis

Plasma samples were thawed by incubation at 37C for 15 min.
100 mL plasma was transferred to an Eppendorf tube, and 400 mL
methanol containing [13C2]-myristic acid (2.5 mg) as an internal
standard was added, then vortex-mixed for 2 min, subsequently
kept at 4C for 1 h. The samples were centrifuged at 20 000 g for
10 min and 100 mL of the supernatant was dried in a SpeedVac
evaporator (Savant Instruments, Framingdale, NY, USA). The
dried analytes were methoximated with a 30-mL aliquot of
methoxyamine pyridine solution (10 mg/mL), vortex-mixed for
3 min, kept for 16 h at room temperature for mathoxiation. After
trimethylsilylation with 30 mL of methyl-trimethyl-silyltrifluoroacetamide (MSTFA) with 1% trimethylchlorosilane
(TMCS), 30 mL n-heptane was added into each GC vial with
methyl stearate (30 mg/mL) for quality control. The final mixture
was vortex-mixed and ready for GC/TOF-MS analysis.
Each 1.0 mL aliquot of the derivatized sample was injected into
an Agilent 6890 gas chromatography system equipped with a
fused-silica capillary column (10 m 0.18 mm i.d.) chemically
bonded with 0.18 um DB-5 stationary phase (J&W Scientific,
Folsom, CA, USA) by an Agilent 7683 Series autosampler
(Agilent Technologies, Atlanta, GA, USA) in the splitless mode.
The inlet temperature was set at 250C. Helium was used as the
carrier gas with a constant flow rate of 1.0 mL/min through the
capillary column. To achieve good resolution of the analytes, the
column temperature was initially maintained at 70C for 2 min,
and then increased at a rate of 30C/min from 70C to 310C,
where it was held for 2 min. The column effluent was introduced
into the ion source (kept at 200C) of a Pegasus III mass spectrometer (Leco Corp., St. Joseph, MI, USA) through a transfer
line, whose temperature was set at 250C.The mass fragmentation
was generated with electron beam at 70 eV with a current of
3.0 mA. Mass spectra were acquired in the full scan mode from
m/z 50 to 800 at a rate of 30 spectra/s, and following a solvent
delay of 170 s, the detector voltage was set at -1650 v.

Data processing, statistics and compound

identification
Automated peak detection and mass spectral deconvolution was
performed with the ChromaTOF software (version 3.25), and the
peak areas were obtained as reported.17 Finally, a data matrix, X,
was constructed by peak areas, with two vectors: sample names as
observations in the first column, and retention times/peaks as the
response variables in the first row. The multivariate statistical
analyses and modeling were performed with SIMCA-P 11 software (Umetrics, Ume, Sweden).
1292

Here, principal components analysis (PCA) and partial leastsquares projection to latent structures and discriminant analysis
(PLS-DA) were used to analyze the acquired GC/TOF-MS data,
which is to calculate the models with established methodologies.20
For PCA modeling, samples were not classified; while for
PLSDA modeling, samples were divided into different groups
(e.g. each stage of disease) as the qualitative dummy variable, Y.
The calculated PCA or PLS-DA model represents a K-dimensional
space (where K stands for the number of variables), and then
projected and reduced to a few principal components (PCs) that
describe the maximum variation of different groups or samples.
Thus, the comparative analysis of the GC/TOF-MS data was facilitated by reducing the dimensionality of the dataset while retaining
as much information as possible. The results of PCA or PLS-DA
are displayed as score plots that represent the scatter of the
samples, showing the similarity or difference among the samples
or groups. Close clustering of the samples/groups indicates their
compositional similarity, whereas distant clustering of the
samples/groups suggests their diverse metabolomic compositions.
The purpose of PLS-DA was to calculate models of the different
groups, and to identify the response variables that contribute most
strongly to the model.
Cross-validation21 with seven cross-validation groups was used
throughout to determine the number of principal components, and
a permutation test was performed with an iteration of 100 to
validate the model. The goodness of fit for a model is evaluated
using three quantitative parameters; that is R2X is the explained
variation in X, R2Y is the explained variation in Y, and Q2Y is the
predicted variation in Y. Statistical analysis between groups
was validated for the biochemical parameters or metabolite
abundances using one-way anova, with a significance level of
0.05 or 0.01.
The compounds were identified by the comparison of the mass
spectra and retention indices of all the detected peaks with authentic reference standards available in the National Institute of Standards and Technology library 2.0 (2005) or an in-house mass
spectra library database maintained by Umea Plant Sciences
Center, Sweden.

Results
Metabolic phenotypes of CSG, CAG, IM, DYS
and GC
Gas chromatography time-of-flight mass spectrometry analysis of
blood plasma yielded the raw data contained in the total ion
current (TIC) chromatograms, Figure 1. Visual inspection of the
chromatogram revealed obviously different metabolic phenotypes
between GC and CSG. Altogether, 223 peaks were detected and 72
were authentically identified (supplementary information,
Table S1).
For overview of the metabolic phenotypes, a PLS-DA model
was calculated with the six groups, that is CSG, CAG, IM, DYS,
GC and postoperative GC. The PLS-DA scores plot showed that
the later five gathered closely, whereas CSG clustered far away
from other them, Figure 2a. According to cross-validation, this
model (of two principal components) explained 12.0%, predicted
10.4% of the variation in Y (sample types), and explained 17.2% of
the variation in X, GC/TOF-MS response variables (R2X = 0.172,

Journal of Gastroenterology and Hepatology 26 (2011) 12901297

2011 Journal of Gastroenterology and Hepatology Foundation and Blackwell Publishing Asia Pty Ltd

L Yu et al.

Metabolomic phenotype of gastric cancer

(a)

(b)

Figure 1 The typical total ion current (TIC) chromatograms of blood plasma from patients with gastric cancer (a) and chronic superficial gastritis (b).
Part of the peaks were identified as, 1. alanine; 2. 2-hydroxybutyrate; 3. valine; 4. urea; 5. glycine; 6. serine; 7. threonine; 8. aminomalonic acid;
9 pyroglutamate; 10. aspartate; 11. creatinine; 12. phenylalanine; 13. glutamine; 14. citrate; 15. glucose; 16.tyrosine; 17. palmitic acid; 18. urate;
19. oleic acid; 20. stearic acid; 21. cholesterol.

R2Y = 0.120, Q2Y = 0.104, respectively). The relatively low explanation (R2X = 0.172) of the GC/TOF-MS data suggests that most
of the detected variables were not significantly different among the
groups; while the relatively low explicative and predictive capacities of sample types (R2Y = 0.120, Q2Y = 0.104) primarily arose
from the considerable overlap of the groups (among CAG, IM,
DYS, GC and postoperative GC), which the model tried (but
failed) to differentiate. The results were also supported by the
significantly improving performance of another validated model
(R2X = 0.228, R2Y = 0.754, Q2Y = 0.575) when the samples were
divided into only two groups (1, CSG; 2, CAG, IM, DYS, GC and

postoperative GC). In general, the later five groups (CAG, IM,

DYS, GC and postoperative GC) showed similar metabolic phenotypes that were quite different from CSG.
For further study, the differences among CAG, IM, DYS, GC
and postoperative GC, a sub-model (PLS-DA) was calculated.
Unfortunately, the PLS-DA model appeared to overlap the five
groups, Figure 2b, especially for GC and postoperative GC, suggesting that metabolic phenotype of postoperative GC was not
significantly improved after operation for a short time. Remarkably, IM clustered close to GC and postoperative GC, while CAG
and DYS clustered further away from GC.

Journal of Gastroenterology and Hepatology 26 (2011) 12901297

2011 Journal of Gastroenterology and Hepatology Foundation and Blackwell Publishing Asia Pty Ltd

1293

Metabolomic phenotype of gastric cancer

L Yu et al.

(a)

Figure 3. Moreover, after surgery operation, urate, 1monohexadecanoylglycerol, pyroglutamate, 11-eicosenoic acid,
glutamate, asparagine, 2-hydroxybutyrate, azelaic acid and threonate restored to more or less extent. The line evidence indicated
these compounds as the potential markers of GC.

Discussion

(b)

Figure 2 The projection to latent structures and discriminant analysis

(PLS-DA) score plot of the six sample groups. (a) Chronic superficial
gastritis was completely separated from others. (A) Postoperative gastric
cancer; (B) gastric cancer; (C) dysplasia; (D) intestinal metaplasia; (E)
chronic atrophic gastritis; (F) chronic superficial gastritis. (b) The PLS-DA
scores plot of chronic atrophic gastritis, intestinal metaplasia, dysplasia
and gastric cancer. (A) postoperative gastric cancer; (B) gastric cancer; (C)
dysplasia; (D) intestinal metaplasia; (E) chronic atrophic gastritis.

Discriminatory markers in plasma of

GC patients
A very good model (R2X = 0.396, R2Y = 0.982, Q2Y = 0.961; of
2 PCs) between CSG and GC completely distinguished them,
which explained 98.2%, predicted 96.1% of the variation in Y
(sample types), and explained 39.6% of the variation in X,
GC/TOF-MS response variables. Identification and statistic
analysis revealed significant elevation of urate, ornithine,
1-monohexadecanoylglycerol, pyroglutamate,11-eicosenoic acid,
glutamate, g-tocopherol, asparagine, 2-hydroxybutyrate, azelaic
acid, while revealing significant reduction of creatinine,
threonate, and two unidentified compounds, in GC, Table 2.
Interestingly, levels of the compounds more abundant in GC presented an obvious, increasing tendency as the development of
gastric carcinogenesis, while creatinine, threonate, and other
unidentified compounds presented in a decreasing tendency,
1294

Oxidative stress is associated with gastric disorders, for example

CAG, gastric ulcer and GC.22 These gastric diseases can be the
result of infection with H. pylori, which is considered to be the
major etiological factor.23,24 H. pylori infection induces an inflammatory response that is also an oxidizing reaction. H. pylori infection induced oxidative DNA damage appears in each stage of the
gastric carcinogenesis model CSGCAGIMDYSGC. It is
important that the production and accumulation of such damage in
the pathogenesis of H. pylori related diseases, and is closely
related with gastric cancer.25
Previous studies showed that lipid peroxides significantly
increased in plasma of cancer patients.26,27 Consistent with the
result, obvious elevation of azelaic acid was detected in this study.
Azelaic acid is the end product of linoleic acid when subjected to
peroxide decomposition,28 and also a marker of lipid peroxidation.29 As a response to an increasing amount of oxidative substances, antioxidant substances elevated significantly in plasma30
to maintain the balance of the body. Glutathione is an important
endogenous antioxidant and its level in plasma of patients with
gastric cancer is significantly increased.31 Although glutathione
was not detectable in GC/TOF-MS, our results showed that the
levels of three amino acids increased in patients with gastric
cancer, glutamate (fold change = 0.608, P = 0.029), cysteine (fold
change = 0.607, P = 0.108) and glycine (fold change = 0.788,
P = 0.205) necessary for glutathione synthesis. The increased
amount of glutamate provides abundant source materials for the
synthesis of glutathione. Additionally, 2-Hydroxybutyrate is either
a metabolic product of threonine or a by-product in glutathione
synthesis. Since no obvious variation of threonine was observed
between CSG and GC, the higher level of 2-hydroxybutyrate in
GC indicated the increased synthesis of glutathione.
Amino acids are the basic unit in structuring protein, and also
the precursor for tumor cell in synthesizing purine and pyrimidine
de novo. Tumor cells require large amounts of nutrient for rapid
growth and uncontrolled proliferation, in particular amino acid
using for synthesizing protein and nucleic acid of tumor cells,
which inevitably lead to perturbation in amino acid metabolism of
patients.32 In this study, four amino acids, pyroglutamate (fold
change = 0.695, P = 5.08E-03), glutamate (fold change = 0.608,
P = 0.029), asparagines (fold change = 0.466, P = 5.82E-05), and
ornithine (fold change = 0.702, P = 2.60E-02) showed statistical
significance between CSG and GC, and they presented an increasing tendency in the process from CSG to GC. They are suggested
as the potential markers to indicate a risk of GC. It is noteworthy
that most amino acids (except the above four) in plasma of GC
were not significantly different from those of CSG in terms of peak
areas; while a previous study showed that most of the amino acids
were more abundant in gastric cancer tissue.11 This inconsistent
result suggested the different metabolomic phenotypes between
topical tissue and systemic plasma.33

Journal of Gastroenterology and Hepatology 26 (2011) 12901297

2011 Journal of Gastroenterology and Hepatology Foundation and Blackwell Publishing Asia Pty Ltd

L Yu et al.

Table 2

Metabolomic phenotype of gastric cancer

Discriminatory metabolites in plasma of significant difference between chronic superficial gastritis and gastric cancer

Identified compounds

Creatinine
Threonate
Unknown compound 1
Unknown compound 2
Urate
Ornithine
1-Monohexadecanoyl-glycerol
Pyroglutamate
11-Eicosenoic acid
Glutamate
g-Tocopherol
Asparagine
2-Hydroxybutyrate
Azelaic acid

Peak areas (CSG)

Peak areas (GC)

Mean (E+05)

SD (E+05)

Mean (E+05)

SD (E+05)

24.332
9.175
38.458
99.969
115.828
48.021
0.456
54.483
5.725
10.499
0.769
4.614
68.527
1.414

18.751
6.183
19.042
2.020
25.706
22.653
0.216
16.341
3.504
5.348
0.577
2.690
65.542
0.759

5.221
4.274
18.617
7.615
154.283
68.359
0.649
78.366
9.387
17.262
1.300
9.905
195.466
46.232

5.685
2.115
12.046
1.459
32.984
17.831
0.149
24.654
4.886
10.270
0.353
2.814
119.432
70.254

Ratio CSG/GC

P-value (E-02)

4.660
2.147
2.066
13.128
0.751
0.702
0.703
0.695
0.610
0.608
0.592
0.466
0.351
0.031

2.34
3.00
0.84
0.40
0.23
2.60
2.33
0.51
3.15
2.90
1.77
< 0.01
0.14
1.43

GC, gastric cancer; CSG, chronic superficial gastritis; SD, standard deviation.

Figure 3 Normalized peak areas of the discriminatory compounds in chronic superficial gastritis (CSG), chronic atrophic gastritis (CAG), intestinal
metaplasia (IM), dysplasia (DYS), gastric cancer (GC) and postoperative gastric cancer(GC-p).

Journal of Gastroenterology and Hepatology 26 (2011) 12901297

2011 Journal of Gastroenterology and Hepatology Foundation and Blackwell Publishing Asia Pty Ltd

1295

Metabolomic phenotype of gastric cancer

L Yu et al.

Tumor cells often display enhanced synthesis of fatty acids,

which is necessary for growth and proliferation, presumably
because a large fraction of their membrane lipids are synthesized
de novo rather than scavenged from extracellular sources.34,35 In
other words, free fatty acids in plasma were not used for the
proliferation of cancer cell. Our results showed that levels of most
free fatty acids (such as palmitic acid, stearic acid, 9-(Z)hexadecenoic acid, oleic acid, linoleic acid, docosahexaenoic acid
and arachidonic acid) are equivalent in both GC and CSG, except
that plasma 11-eicosenoic acid in GC were higher than in CSG.
Urate is the end product of purine catabolism in the body in
addition to being an antioxidant.36 Consistent with a recent report
that the level of urate in plasma is significantly increased37 in GC,
we also found much higher plasma urate in GC than in CSG
(P = 2.32E-03). Notably, the level of urate in plasma of postoperative GC was significantly lower than that of GC (folder
change = 0.760, P = 2.50E-02). The result suggested that removal
of cancer tissue downregulated purine catabolism involved in the
growth and proliferation of tumor cell.

Conclusion
The plasma phenotype of CSG is significantly different from GC
and precancerous stages, while the metabolic phenotype of IM is
similar to GC. The discriminatory metabolites, such as azelaic acid,
glutamate, 2-Hydroxybutyrate, urate, creatinine and threonate characterized progressive stages from CSG to GC and might be the
potential markers to indicate a risk of GC. Additionally GC/
TOF-MS based metabonomics is a promising analytical platform to
characterize metabolic phenotype of GC and precancerous stages.

Acknowledgment
This work was supported by grants from Natural Science
Foundation of Jiangsu Educational Nature Science Foundation
(10KJB320007).

References
1 Parkin DM, Bray F, Ferlay J, Pisani P. Global cancer statistics,
2002. CA Cancer J. Clin. 2005; 55: 74108.
2 Yang L. Incidence and mortality of gastric cancer in China. World J.
Gastroenterol. 2006; 12: 1720.
3 Correa P. A human model of gastric carcinogenesis. Cancer Res.
1988; 48: 355460.
4 Sakaki N, Kozawa H, Egawa N, Tu Y, Sanaka M. Ten-year
prospective follow-up study on the relationship between
Helicobacter pylori infection and progression of atrophic gastritis,
particularly assessed by endoscopic findings. Aliment. Pharmacol.
Ther. 2002; 16 (Suppl. l2): 198203.
5 Nicholson JK, Lindon JC. Systems biology: metabonomics. Nature
2008; 455: 10546.
6 Holmes E, Wilson ID, Nicholson JK. Metabolic phenotyping in
health and disease. Cell 2008; 134: 7146.
7 Clayton TA, Lindon JC, Cloarec O et al. Pharmaco-metabonomic
phenotyping and personalized drug treatment. Nature 2006; 440:
10737.
8 Nicholson JK, Connelly J, Lindon JC, Holmes E. Metabolomics: a
platform for studying drug toxicity and gene function. Nat. Rev.
Drug Discov. 2002; 1: 15361.

1296

9 Denkert C, Budczies J, Weichert W et al. Metabolite profiling of

human colon carcinoma-deregulation of TCA cycle and amino acid
turnover. Mol. Cancer 2008; 7: 7286.
10 Chan EC, Koh PK, Mal M et al. Metabolic profiling of human
colorectal cancer using high-resolution magic angle spinning nuclear
magnetic resonance (HR-MAS NMR) spectroscopy and gas
chromatography mass spectrometry (GC/MS). J. Proteome Res.
2009; 8: 35261.
11 Hirayama A, Kami K, Sugimoto M et al. Quantitative metabolome
profiling of colon and stomach cancer microenvironment by capillary
electrophoresis time-of-flight mass spectrometry. Cancer Res. 2009;
69: 491825.
12 Sreekumar A, Poisson LM, Rajendiran TM et al. Metabolomic
profiles delineate potential role for sarcosine in prostate cancer
progression. Nature 2009; 457: 91014.
13 Yang J, Xu G, Zheng Y et al. Diagnosis of liver cancer using
HPLC-based metabonomics avoiding false-positive result from
hepatitis and atocirrhosis diseases. J. Chromatogr. B Analyt. Technol.
Biomed. Life Sci. 2004; 813: 5965.
14 Odunsi K, Wollman RM, Alnbrosone CB et al. Detection of
epithelial ovarian cancer using 1H-NMR-based metabnomies. Int. J.
Cancer 2005; 113: 7828.
15 Denkert C, Budczies J, Kind T et al. Mass spectrometry-based
metabolic profiling reveals different metabolite patterns in invasive
ovarian carcinomas and Varian borderline tumors. Cancer Res. 2006;
66: 10795804.
16 Bullinger D, Frhlich H, Klaus F et al. Bioinformatical evaluation of
modified nucleosides as biomedical markers in diagnosis of breast
cancer. Anal. Chim. Acta 2008; 618: 2934.
17 A J, Trygg J, Gullberg J et al. Extraction and GC/MS analysis of the
human blood plasma metabolome. Anal. Chem. 2005; 77: 808694.
18 Jonsson P, Johansson AI, Gullberg J et al. High-throughput data
analysis for detecting and identifying differences between samples in
GC/MS-based metabolomic analyses. Anal. Chem. 2005; 77:
563542.
19 Trygg J, Holmes E, Lundstedt T. Chemometrics in metabonomics. J.
Proteome Res. 2007; 6: 46979.
20 Eriksson L, Johansson E, Kettaneh-Wold N, Wold S. Multi- and
Megavariate Data Analysis: Principles and Applications. Sweden:
Umeatrics Academy: Umetrics AB, 2001; 1025.
21 Wold S. Cross-validatory estimation of the number of components in
factor and principal components models. Technometrics 1978; 20:
397405.
22 Janssen YM, Van Houten B, Borm PJ, Moosman BT. Cell and
tissue responses to oxidative damage. Lab. Invest. 1993; 69:
26174.
23 Atherton JC. The clinical relevance of strain types of Helicobacter
pylori. Gut 1997; 40: 7013.
24 Naito Y, Yoshikawa T. Molecular and cellular mechanisms involved
in Helicobacter pylori-induced inflammation and oxidative stress.
Free Radic. Biol. Med. 2002; 33: 32336.
25 Farinati F, Cardin R, Degan P et al. Oxidative DNA damage
accumulation in gastric carcinogenesis. Gut 1998; 42: 3516.
26 Beevi SS, Rasheed MH, Geetha A. Evidence of oxidative and
nitrosative stress in patients with cervical squamous cell carcinoma.
Clin. Chim. Acta 2007; 375: 11923.
27 Kasapovic J, Pejic S, Todorovic A, Stojiljkovic V, Pajovic SB.
Antioxidant status and lipid peroxidation in the blood of breast
cancer patients of different ages. Cell Biochem. Funct. 2008; 26:
72330.
28 Raghavamenon AC, Garelnabi M, Babu S, Aldrich A, Litvinov D,
Parthasarathy S. Alpha-tocopherol is ineffective in preventing the
decomposition of preformed lipid peroxides and may promote the
accumulation of toxic aldehydes: a potential explanation for the

Journal of Gastroenterology and Hepatology 26 (2011) 12901297

2011 Journal of Gastroenterology and Hepatology Foundation and Blackwell Publishing Asia Pty Ltd

L Yu et al.

29

30

31

32

33
34

35

failure of antioxidants to affect human atherosclerosis. Antioxid.

Redox. Signal. 2009; 11: 123748.
Schallreuter KU, Wood JM. Azelaic acid as a competitive inhibitor
of thioredoxin reductase in human melanoma cells. Cancer Lett.
1987; 36: 297305.
Lee GJ, Chung HW, Lee KH, Ahn HS. Antioxidant vitamins and
lipid peroxidation in patient with cervical intraepiththelial neoplasia.
J. Korean Med. Sci. 2005; 20: 26772.
Feng J-F, Li S-L. Correlation between oxidative stress and trace
elements in blood of patients with cancer. Chin. J. Clin. Rehabil.
2006; 10: 18790.
Cascino A, Muscaritoli M, Cangiano C et al. Plasma amino acid
imbalance in patients with lung and breast cancer. Anticancer Res.
1995; 15: 50710.
Mizushima N, Klionsky DJ. Protein turnover via autophagy:
implications for metabolism. Annu. Rev. Nutr. 2007; 27: 1940.
Kannan R, Lyon I, Baker N. Dietary control of lipogenesis in vivo in
host tissues and tumors of mice bearing Ehrlich ascites carcinoma.
Cancer Res. 1980; 40: 460611.
Ookhtens M, Kannan R, Lyon I, Baker N. Liver and adipose tissue
contributions to newly formed fatty acids in an ascites tumor. Am. J.
Physiol. 1984; 247 (1 Pt 2): R14653.

Metabolomic phenotype of gastric cancer

36 Ames BN, Cathcart R, Schwiers E, Hochstein P. Uric acid provides

an antioxidant defense in humans against oxidant- and radical-caused
aging and cancer: a hypothesis. Proc. Natl. Acad. Sci. U.S.A. 1981;
78: 685862.
37 Burgaz S, Torun M, Yardim S, Sargin H, Orman MN, Ozdamar NY.
Serum carotenoids and uric acid levels in relation to cancer. J. Clin.
Pharm. Ther. 1996; 21: 3316.

Supporting information
Additional Supporting Information may be found in the online
version of this article:
Table S1 The low-molecular-weight compounds identified in
blood plasma
Please note: Wiley-Blackwell are not responsible for the content or
functionality of any supporting materials supplied by the authors.
Any queries (other than missing material) should be directed to the
corresponding author for the article.

Journal of Gastroenterology and Hepatology 26 (2011) 12901297

2011 Journal of Gastroenterology and Hepatology Foundation and Blackwell Publishing Asia Pty Ltd

1297