Transcriptome profiling analysis reveals multiple
modulatory effects of Ginkgo biloba extract in the
liver of rats on a high-fat diet
Xiaomei Gu
1,
*, Zuoquan Xie
2,
*, Qi Wang
1,
*, Gang Liu
1
,YiQu
1
, Lu Zhang
1,2
, Jiahu Pan
3
, Guoping
Zhao
1,4
and Qinghua Zhang
1,2,4
1 National Engineering Center for Biochip at Shanghai, China
2 State Key Laboratory of Medical Genomics and Shanghai Institute of Hematology, Ruijin Hospital, Shanghai Jiaotong University School of
Medicine, China
3 School of Pharmacy, Fudan University, Shanghai, China
4 MOST-Shanghai Key Laboratory of Disease and Health Related Genomics, China
Ginkgo biloba has been used for medical purposes for
centuries in traditional Chinese medicine. The standard
extracts of G. biloba leaves [G. biloba extract (GBE)]
are now more widely used as dietary supplements or

(Received 8 September 2008, revised 29
December 2008, accepted 5 January 2009)
doi:10.1111/j.1742-4658.2009.06886.x
Leaf extract of Ginkgo biloba (GBE) is increasingly used as a herbal medi-
cine for the treatment of neurodegenerative, cardiovascular and cerebrovas-
cular diseases. Several studies have demonstrated many protective effects of
GBE in neurons, the endothelium and liver. In this study, we investigated
the molecular mechanisms underlying the effects of GBE in disorders
induced by long-term exposure to a high-fat diet (HFD). Rats were fed an
HFD with or without the GBE product GBE50 for 19 weeks. We found
that GBE50 reduced the development of fatty liver induced by an HFD
and inhibited the commonly observed elevation of serum cholesterol and
lactate dehydrogenase levels. Transcriptome profiling analysis showed that
several genes were modulated by GBE50 in liver, including those involved
in lipid metabolism, carbohydrate metabolism, vascular constriction, ion
transportation, neuronal systems and drug metabolism. Notably, a number
of genes coding for proteins involved in cholesterol metabolism were
repressed, and some were upregulated. Fatty acid biosynthesis appeared to
be repressed, whereas fatty acid metabolism appeared to be enhanced. In
conclusion, using transcriptome profiling analysis, we demonstrated the
molecular basis for the pleiotropic effects of GBE50, particularly those
involved in lipid metabolism. This study provided new clues for further
pharmacological study of GBEs.
Abbreviations
EST, expressed sequence tag; GBE, Ginkgo biloba leaf extract; HFD, high-fat diet; LDH, lactate dehydrogenase; PPAR, peroxisome
proliferator-activated receptor; SREBP, sterol regulatory element-binding protein.
1450 FEBS Journal 276 (2009) 1450–1458 ª 2009 The Authors Journal compilation ª 2009 FEBS
found to contribute to the antioxidant and ⁄ or free rad-
ical-scavenging properties, prevent stroke and transient
ischemic attack, and inhibit membrane lipid peroxida-

effects of GBE, we fed rats an HFD or an HFD plus
the GBE product GBE50 for 19 weeks. We then
removed the livers and subjected them to transcriptomic
profiling analysis using cDNA microarrays. Our results
have revealed multiple effects of GBE on rat liver as
well as the underlying molecular mechanisms mediat-
ing these effects.
Results and Discussion
Body weight and biochemical parameters
After 19 weeks of HFD (group H) or control diet
(standard chow; group C) exposure, there was no sig-
nificant difference in body weight between rats in the
two groups. However, rats fed the HFD plus GBE50
(group HG) showed a significant increase in body
weight. We also measured serum and hepatic total
cholesterol content as well as serum high-density lipo-
protein cholesterol content. We found that serum and
hepatic total cholesterol content were significantly
increased in group H over group C by 30% and 63%,
respectively. Furthermore, serum lactate dehydroge-
nase (LDH) increased three-fold in group H, reflecting
the harm caused by exposure to a long-term HFD.
The serum total cholesterol and LDH elevations found
in group H were inhibited in group HG, suggesting
that GBE50 prevented the harm caused by an HFD.
In addition, we found no significant alterations in the
levels of alanine aminotransferase, aspartate amino-
transferase or creatinine kinase in either group H or
group HG (Table 1). However, serum triglyceride
levels were slightly decreased in group H (24%) in

) 1.08 ± 0.17 0.82 ± 0.15
a
1.23 ± 0.24
ALT (UÆL
)1
) 120 ± 54 171 ± 132 134 ± 73
AST (UÆL
)1
) 187 ± 80 262 ± 96 181 ± 61
LDH (UÆL
)1
) 633 ± 513 2809 ± 1493
b
2106 ± 507
a,d
CK (UÆL
)1
) 2279 ± 1622 2723 ± 867 1813 ± 1602
Hepatic total cholesterol (mgÆg
)1
) 2.16 ± 0.17 3.54 ± 0.84
a
2.84 ± 0.25
a
a
P < 0.01,
b
P < 0.05, versus group C;
c
P < 0.01,

To verify the alterations in levels of the genes identi-
fied with microarray, eight lipid metabolism-associated
genes were selected for real-time RT-PCR validation.
As shown in Fig. 2B, the gene expression regulation
patterns found through RT-PCR were consistent with
data obtained through microarray, thereby demon-
strating the reliability of the microarray results.
Regulation of metabolism-related genes in
group H
In group H, 64 genes ⁄ ESTs were found to be upregulat-
ed and 22 genes ⁄ ESTs were found to be downregulated
in response to the HFD (Fig. S1 and Table S2). The
HFD affected carbohydrate metabolism, which could
result in insulin resistance and later development of dia-
betes [21]. Upregulation of Foxa3, which codes for a
transcription factor involved in glucose homeostasis by
binding to the proglucagon gene G2 promoter element,
and downregulation of an NADH-ubiquinone oxidore-
ductase member gene (CI-B9) and the pyruvate kinase
gene (Pklr), suggested impaired glucose metabolism in
group H rats. In addition, excess intake of dietary fat
caused deregulation of lipid metabolism. For example,
the upregulated genes for an acyl-CoA synthetase mem-
ber (Acsl1), pyruvate carboxylase (Pc), aspartoacylase 3
(Acy3) and solute carrier family 25 member 1 (Slc25a1)
are involved in lipid biosynthesis, and the downregulat-
ed genes for acetyl-CoA carboxylase beta (Acacb) and
fatty acid-binding proteins (Fabp1 and Fabp5) contrib-
ute to fatty acid metabolism. The repression of Hmgcr,
which encodes the cholesterol biosynthesis rate-limiting

staining according to the method described
by Bouma et al. [34].
GBE effects on rat liver gene expression X. Gu et al.
1452 FEBS Journal 276 (2009) 1450–1458 ª 2009 The Authors Journal compilation ª 2009 FEBS
Lipid metabolism genes
By comparison with expression in group H, the further
repression of Hmgcr and the upregulation of Insig2,
the upstream negative regulator of Hmgcr, probably
contribute to impaired cholesterol biosynthesis.
Furthermore, decreased expression of gene for the bile
acid biosynthesis rate-limiting enzyme cholesterol 7-a-
hydroxylase (Cyp7a1) and Cyp8b1 indicates that the
conversion of cholesterol into bile acid may be
repressed, which is consistent with our finding of
upregulation of the gene for bile acid synthesis
A

B
a b c d
e
f
g
h
Fig. 2. (A) Box-plot presentation of gene expression regulation. Gene expression regulation was determined with the microarray data and
compared with reference samples from group C. Detailed information is provided in Experimental procedures. A cut-off of 1.5-fold was used
to identify regulated genes filtered with
SAM. Eight groups of regulation patterns were obtained, and the number of genes is indicated above
each graph. (B) Validation of eight genes with real-time RT-PCR.
X. Gu et al. GBE effects on rat liver gene expression
FEBS Journal 276 (2009) 1450–1458 ª 2009 The Authors Journal compilation ª 2009 FEBS 1453

transferase (B4galt3) supported the hypothesis that gly-
cogenesis was impaired with GBE50 intake, although
Foxa3 remained upregulated in group HG. In addi-
tion, the decreased expression of Pc in group HG fur-
ther suggested impaired glyconeogenesis. Upregulation
of the insulin receptor substrate gene (Irs3) could con-
tribute to glucose homeostasis and utilization enhance-
ment in liver [26]. However, downregulation of
mitochondrial protein genes (CI-B9, Atp6v0e1 and
Nudfa8) indicates that the respiratory chain and oxida-
tive phosphorylation pathway was repressed in
group HG.
Regulation of cholesterol metabolism-related
gene expression by HFD and GBE50 in rat liver
To better understand cholesterol metabolism regula-
tion with the HFD and GBE50, real-time RT-PCR
was employed to validate the microarray data, includ-
ing genes that were not spotted or detected with the
microarray (Table 2). The nuclear steroid receptor
peroxisome proliferator-activated receptors (PPARs)
are transcription factors that form heterodimers with
retinoid X receptors to initiate the transcriptional regu-
lation of target genes. Three subtypes of PPARs (a,
b ⁄ d and c) and retinoid X receptors (a, b and c) have
been identified, and the PPARs are mainly regulators
of lipid metabolism [27]. In group H, genes for lipid
metabolism regulators (Ppard, Pparg
, and Rxra, Rxrb)
were upregulated in rat liver, suggesting that lipid
metabolism was activated. As the genes for low-density

that the cholesterol excretion from liver might also be
lessened by GBE50 (Table 2 and Fig. 3).
Experimental procedures
GBE
GBE50 is a standardized GBE product that matches the
standardized German product as EGb761. GBE50 is
approved by the China State Food and Drug Administra-
tion (SFDA) and is used clinically in China. GBE50 con-
tains > 44.0% total flavonoids, > 24.0% flavonoglycoside
and > 6.0% total terpen lactones. Ginkgolic acid is
GBE effects on rat liver gene expression X. Gu et al.
1454 FEBS Journal 276 (2009) 1450–1458 ª 2009 The Authors Journal compilation ª 2009 FEBS
controlled to be < 5 p.p.m. GBE50 products were kindly
provided by Shanghai Xingling Pharmaceutical (Shanghai,
China), manufactured under good manufacturing process
conditions (lot no. 20030608).
Animal diet and experiments
Thirty male Wistar rats (3 weeks old, weighing 50–55 g)
were obtained from Shanghai Laboratory Animal Center of
the Shanghai Institute of Biological Sciences, Chinese Acad-
emy of Sciences (Shanghai, China). Animals were housed in
a temperature-controlled room (23–25 °C) for 19 weeks
with a 12 h light ⁄ dark cycle, and allowed unrestricted
access to pellet food and water. The rats were randomly
divided into three groups of 10, including groups C, H and
HG. Group C was fed standard chow and acted as the nor-
mal control. Group H was fed an HFD, containing 8%
lard, 7% egg yolk powder, 0.5% sodium cholate and
84.5% standard rat chow. No pure cholesterol was added.
Group HG was fed the HFD supplemented with GBE50

0.55 ± 0.14
a
0.36 ± 0.12
a
Secretion
Abcg5 1.00 ± 0.07 2.34 ± 0.20
b
1.55 ± 0.25
a
Abcg8 1.00 ± 0.02 1.76 ± 0.16
b
1.33 ± 0.03
b
Bile acid metabolism
Cyp7a1 1.00 ± 0.67 1.00 ± 0.51 0.65 ± 0.12 1.26 ± 0.04 0.16 ± 0.09
a
0.30 ± 0.00
a
Cyp8b1 1.00 ± 0.41 1.00 ± 0.20 1.30 ± 0.30 0.76 ± 0.27 0.68 ± 0.08 0.76 ± 0.57
Regulators
Nr1h4 1.00 ± 0.18 1.00 ± 0.47 0.96 ± 0.16 1.45 ± 0.74 1.41 ± 0.21 2.05 ± 0.52
a
Ppard 1.00 ± 0.06 2.58 ± 0.46
b
6.41 ± 1.62
b
Pparg 1.00 ± 0.01 2.08 ± 0.75
a
1.41 ± 0.33
Rxra 1.00 ± 0.19 3.07 ± 0.96

room temperature to collect sera for biochemical measure-
ments. All procedures were approved by Animal Experi-
ment Committee of the School of Pharmacy Fudan
University.
Biochemical measurement and cholesterol
detection
The collected sera were assayed with a Roche-Hitachi Mod-
ular P800 Chemistry analyzer and corresponding enzymatic
reagents (Roche Diagnostics, Indianapolis, IN, USA) in the
clinical laboratory of Zhongshan Hospital, affiliated to
Fudan University. Hepatic cholesterol content was deter-
mined as previously described [28], with a modification.
Briefly, frozen stored liver tissues were weighed and homog-
enized, lipids were extracted with isopropanol, and aliquots
were used for colorimetric determination of total choles-
terol using a cholesterol oxidase assay kit (Shanghai Jiemen
Biotech Co., Shanghai, China). Liver tissue cholesterol con-
tent was normalized to the liver tissue weight.
Histopathological detection
Liver tissues were fixed in 10% formalin, dehydrated and
paraffin embedded. A series of 5-lm-thick slices were sub-
jected to standard hematoxylin ⁄ eosin staining. For Oil
Red O staining, frozen liver tissues were embedded in opti-
cal coherence tomography and cryosectioned at 8 lm thick-
ness, following standard procedures. The histology slides
were cross-inspected by two pathologists.
RNA preparation and cDNA microarray
hybridization
Rat liver total RNA was isolated using the Trizol reagent
(Invitrogen) and purified with an RNAeasy column (Qia-

(a) the signal of log ratios in the control group was signifi-
cantly variable, as detected by a one-class comparison
(setting false discovery rate = 0.01) using the software
sam 2.20 [32]; (b) the standard deviation of the log ratios in
the control group was > 1; or (c) the absolute value of the
log ratio was > 0.585 (corresponding to a 1.5-fold change)
in at least two samples in the control group. On the basis
of the remaining dataset, we identified genes that were
upregulated or downregulated in group H and group HG
relative to control. This procedure was performed using a
one-class comparison procedure (setting false discovery
rate = 0.01) with sam software 2.20. In this analysis, we
tested whether the log ratio values of each treatment group
(H or HG) were significantly different from zero. The sig-
nificant genes identified by the independent groups were
used to perform hierarchical clustering of samples by using
cluster 3.0, and the results were visualized using treeview
(cluster and treeview were developed by Eisen et al.
[33]). The regulated genes were classified into different
groups on the basis of ontology, and the online database
KEGG () was applied in gene func-
tion pathway assignment.
RT-PCR analysis
Quantitative RT-PCR was conducted with an ABI
Prism7000 sequence detection system (Applied Biosystems,
Foster City, CA, USA) using SYBR Green I (TaKaRa Bio-
tech, Dalian, China). Genes for RT-PCR were selected on
the basis of the microarray data or because they were
known to be involved in lipid metabolism but were not
included in the gene list selected with microarray. The PCR

AK & Liu L (2007) Ginkgo biloba extract prevents
ethanol induced dyslipidemia. Am J Chin Med 35,
643–652.
3 Smith U (1994) Carbohydrates, fat, and insulin action.
Am J Clin Nutr 59, 686S–689S.
4 Chen JW, Chen YH, Lin FY, Chen YL & Lin SJ
(2003) Ginkgo biloba extract inhibits tumor necrosis fac-
tor-alpha-induced reactive oxygen species generation,
transcription factor activation, and cell adhesion mole-
cule expression in human aortic endothelial cells. Arte-
rioscler Thromb Vasc Biol 23, 1559–1566.
5 Smith JV & Luo Y (2003) Elevation of oxidative free
radicals in Alzheimer’s disease models can be attenuated
by Ginkgo biloba extract EGb 761. J Alzheimers Dis 5,
287–300.
6 Kudolo GB, Wang W, Barrientos J, Elrod R & Blodg-
ett J (2004) The ingestion of Ginkgo biloba extract
(EGb 761) inhibits arachidonic acid-mediated platelet
aggregation and thromboxane B2 production in healthy
volunteers. J Herb Pharmacother 4, 13–26.
7 Dutta-Roy AK, Gordon MJ, Kelly C, Hunter K, Cros-
bie L, Knight-Carpentar T & Williams BC (1999) Inhib-
itory effect of Ginkgo biloba extract on human platelet
aggregation. Platelets 10, 298–305.
8 Koltermann A, Hartkorn A, Koch E, Furst R, Vollmar
AM & Zahler S (2007) Ginkgo biloba extract EGb 761
increases endothelial nitric oxide production in vitro
and in vivo. Cell Mol Life Sci 64, 1715–1722.
9 Chen X, Salwinski S & Lee TJ (1997) Extracts of
Ginkgo biloba and ginsenosides exert cerebral vasorelax-

tion and accumulation of C-terminal fragment of APP
in primary cortical neurons. Neurosci Lett 406, 55–59.
16 Patil S & Chan C (2005) Palmitic and stearic fatty acids
induce Alzheimer-like hyperphosphorylation of tau in
primary rat cortical neurons. Neurosci Lett 384, 288–
293.
17 Gauthier MS, Favier R & Lavoie JM (2006) Time
course of the development of non-alcoholic hepatic stea-
tosis in response to high-fat diet-induced obesity in rats.
Br J Nutr 95, 273–281.
18 Wang L, Zhou GB, Liu P, Song JH, Liang Y, Yan XJ,
Xu F, Wang BS, Mao JH, Shen ZX et al. (2008) Dis-
section of mechanisms of Chinese medicinal formula
Realgar-Indigo naturalis as an effective treatment for
promyelocytic leukemia. Proc Natl Acad Sci USA 105,
4826–4831.
19 Watanabe CM, Wolffram S, Ader P, Rimbach G,
Packer L, Maguire JJ, Schultz PG & Gohil K (2001)
X. Gu et al. GBE effects on rat liver gene expression
FEBS Journal 276 (2009) 1450–1458 ª 2009 The Authors Journal compilation ª 2009 FEBS 1457
The in vivo neuromodulatory effects of the herbal
medicine Ginkgo biloba. Proc Natl Acad Sci USA 98,
6577–6580.
20 Gohil K & Packer L (2002) Global gene expression
analysis identifies cell and tissue specific actions of
Ginkgo biloba extract, EGb 761. Cell Mol Biol 48, 625–
631.
21 Storlien LH, Pan DA, Kriketos AD & Baur LA (1993)
High fat diet-induced insulin resistance. Lessons and
implications from animal studies. Ann N Y Acad Sci

29 Li RY, Zhang QH, Liu Z, Qiao J, Zhao SX, Shao L,
Xiao HS, Chen JL, Chen MD & Song HD (2006) Effect
of short-term and long-term fasting on transcriptional
regulation of metabolic genes in rat tissues. Biochem
Biophys Res Commun 344, 562–570.
30 Zheng PZ, Wang KK, Zhang QY, Huang QH, Du YZ,
Zhang QH, Xiao DK, Shen SH, Imbeaud S, Eveno E
et al. (2005) Systems analysis of transcriptome and pro-
teome in retinoic acid ⁄ arsenic trioxide-induced cell dif-
ferentiation ⁄ apoptosis of promyelocytic leukemia. Proc
Natl Acad Sci USA 102, 7653–7658.
31 Du Y, Wang K, Fang H, Li J, Xiao D, Zheng P, Chen
Y, Fan H, Pan X, Zhao C et al. (2006) Coordination of
intrinsic, extrinsic, and endoplasmic reticulum-mediated
apoptosis by imatinib mesylate combined with arsenic
trioxide in chronic myeloid leukemia. Blood 107, 1582–
1590.
32 Tusher VG, Tibshirani R & Chu G (2001) Significance
analysis of microarrays applied to the ionizing radiation
response. Proc Natl Acad Sci USA 98, 5116–5121.
33 Eisen MB, Spellman PT, Brown PO & Botstein D
(1998) Cluster analysis and display of genome-wide
expression patterns. Proc Natl Acad Sci USA 95,
14863–14868.
34 Bouma ME, Amit N & Infante R (1979) Ultrastructural
localization of apo-b and apo-c binding to very low
density lipoproteins in rat liver. Virchows Arch B Cell
Path 30, 161–180.
Supporting information
The following supplementary material is available: