Original Paper
|
|
|
file on Synergy |
Acta Biochim Biophys
Sin 2006, 38: 900-910
doi:10.1111/j.1745-7270.2006.00235.x
Identification of Genes
Related to Nasopharyngeal Carcinoma with the Help of Pathway-based Networks
Hui LI1, Cai-Ping REN1*,
Xiao-Jun TAN2,
Xu-Yu YANG1,
Hong-Bo ZHANG1,
Wen ZHOU1,
and Kai-Tai YAO1*
1
Cancer Research Institute, Xiangya School of Medicine, Central South
University, Changsha 410078, China;
2
Reproductive and Stem Cell Engineering Institute, Xiangya School of Medicine,
Central South University, Changsha 410078, China
Received: August 30,
2006
Accepted: October
8, 2006
This work was
supported by a grant from the Hunan Science and Technology Fund
*Corresponding
authors:
Cai-Ping REN: Tel,
86-731-2355066; Fax, 86-731-4360094; E-mail, [email protected]
Kai-Tai YAO: Tel, 86-731-4805451;
Fax, 86-731-4360094; E-mail, [email protected]
Abstract cDNA microarray is a powerful tool that can be used to
simultaneously analyze the expression levels of tens of thousands of genes.
Compared with normal nasopharynx (NP) tissues, 2210 genes were highly
differentially expressed in nasopharyngeal carcinoma (NPC) tissues detected by
cDNA microarray. Since signal pathway is widely used to describe the complex
relationship between genes, a pathway-based network was constructed to
visualize the connection between the genes obtained from microarray data in
this report. We analyzed the targeted genes that may have more important
influence on this gene network with statistical methods and found that some
genes might have significant influence on this network, especially Ras-related
nuclear protein (RAN), carboxyl ester lipase (CEL), v-rel reticuloendotheliosis
viral oncogene homolog A (RELA) genes. To verify the results from
pathway-based selection, reverse transcription-polymerase chain reaction
(RT-PCR) and real-time RT-PCR were performed to detect the expression levels of
RAN, CEL and RELA genes and it was found that the RAN
and CEL genes were significantly up-regulated in more than 80% of NPC tissues.
To further elucidate the function of the RAN gene, RAN expression
was specifically suppressed in a 5-8F NPC cell line by RNA interference (RNAi).
As expected, the depletion of RAN could effectively block the
proliferation of tumor cells. Therefore, our study may open up a new way to
analyze the vast microarray data.
Key words nasopharyngeal carcinoma; microarray; real-time PCR; RNA
interference; Ras-related nuclear protein (RAN)
cDNA microarray is a useful and powerful tool to study the expression
of multiple genes simultaneously. In recent years, cDNA microarray has been
frequently applied in the field of cancer research. For example, it has been
used in the study of lung, breast, gastric and colorectal cancers and has
generated a large amount of data [1–4] leading to a thorough understanding of gene
functions, gene regulations, gene interactions and pathways. Usually microarray
data can be analyzed by clustering or classification methods, but there are
still sets of massive data remaining to be identified [5]. The current main
challenge is how to validly and rapidly extract the comprehensive overview from
this large amount of information. Biological pathways can provide detailed information about the
relationships between genes. Cancer is a disease involving changes in the
expression of a set of genes that can interact with each other, forming a large
gene network. In such a network, each pathway function as a line connecting
two genes and each gene is defined as a node. Several tools, such as DragonView
[6] and KnowDDledgeEditor [7], are used for the first time in this field to
analyze the relationship between pathway database and microarray data,
resulting in a comprehensive overview and interpretation of the gene expression
profiles. Dahlquist et al. developed a free stand-alone computer
program called GenMAPP [8] for viewing and analyzing gene expression data in
the context of biological pathways. Pandey et al. [9] have established a
pathway miner software to analyze the microarray data and visualize gene
expression profiles directly on the internet. Currently, KEGG at http://www.genome.ad.jp, BioCarta at http://www.biocarta.com and GenMAPP at http://www.genmapp.org are commonly used
in the analyses of pathway networks.Based on our previously reported results generated from the data of
four individual microarrays in our institute [10–13], a total of 2210 genes
were found to be significantly up- or down-regulated in nasopharyngeal
carcinoma (NPC) tissues, compared with normal nasopharynx (NP) tissues.
Subsequently, the 2210 genes were used to construct a pathway-based network
with ArrayxPath software. Our analysis focused on using statistical methods to
program and calculate the “degree” of every node (gene), which may
represent the action or influence of every gene on this network. Accordingly, a
set of genes that may play important roles in NPC were screened out. Among
them, the carboxyl ester lipase (CEL) gene was found to be a biomarker
candidate of NPC and the Ras-related nuclear protein (RAN) gene was of
particular interest. Finally, reverse transcription-polymerase chain reaction
(RT-PCR), real-time RT-PCR and RNA interference (RNAi) were performed on the RAN
gene to confirm the results based on a pathway network analysis and to further
explore the possible roles of RAN in NPC development.
Materials and Methods
Pathway-based network
construction and “degree” calculation
Using the ArrayXPath tool (http://147.46.184.220:8080/ArrayXPath/),
we constructed the pathway-based network of 2210 up- or down-regulated genes in
NPC tissues, which was generated automatically when the names of 2210 genes
were entered into the tool [Fig. 1(A)].
Fig. 1(B) is an example elucidating the
relationship between the RELA gene and tumor protein p53 (TP53)
gene [14]. After looking through the BioCarta database, we collected all the
pathway information concerning the 2210 genes and used the “visual
basic” tool in Excel to establish a computer program to count the
“degree” of every gene in the pathway network, according to the
principles described previously [15,16]. In graph theory, a graph is a finite
set of dots called vertices (nodes) connected by links called edges. The
degree of a vertex is the number of edges ending at that vertex. In the
pathway network we constructed, each node represents a gene and each edge
represents a pathway connecting two genes participating in the same pathway.
Using our computer program, we could calculate the “degree” of each
gene (node) and establish the pathway relationship between any two genes. For
clarity, a simple network model is shown here to elucidate how to calculate
the gene’s “degree” [Fig. 1(C)]. There are A, B, C, D, E, F
genes and 10 edges (pathways) connecting them. As there are seven edges ending
at the A gene, the “degree” of the A gene is defined as 7. Similarly,
the “degree” of B, E, F gene is 2 respectively; the
“degree” of C gene is 3; the “degree” of D is 4. In this
way, we could count the “degree” of every node (gene) in our
pathway-based network.
Selecting genes of interest
One sample t test is used to evaluate the “degree”
of every node. If the “degree” of one node (gene) is greater than the
mean+2?SD, this node (gene) will have a
significant “degree” and should have a greater influence on the
network according to the graph theory.
Tissue samples
Thirty-six poorly-differentiated NPC biopsies (T1–T36) of primary
tumors were obtained from NPC patients with consent before treatment at the
Hunan Cancer Hospital (Changsha, China). In addition, 9 normal NP tissues were
obtained from patients without NPC at the same hospital. All the specimens were
reviewed by an otorhinolaryngologic pathologist. Fresh NPC tissues or NP
tissues were snap-frozen in liquid nitrogen and stored until use.
RT-PCR
Total RNA from tissues was isolated using Trizol reagent (Gibco
BRL, Grand Island, USA). After treatment with RNase-free DNase I (Invitrogen,
Carlsbad, USA) for 15 min at 37 ?C, total RNA was reversely transcribed
according to the manufacturer’s instructions using Superscript first-strand
synthesis kit (Invitrogen). Two microliters of the first-strand cDNA mixture
was placed in a final volume of 20 ml containing 2 ml of 10?buffer, 1.7 ml of 20 mM MgCl2, 0.2 ml of 2 mM dNTPs, 1 U of Taq
polymerase, 10 pM each of the two RAN primers or GAPDH primers.
The following primers were used in RT-PCR: RAN-sense: 5‘-ATGGTGGTACTGGAAAAACGAC-3‘,
RAN-antisense: 5‘-GGGATGTTTTCACACACTCGTA-3‘; CEL-sense:
5‘-ATCGTGGTCACCTTCAACTACC-3‘, CEL-antisense: 5‘-GAAGAGTGGGTTTTTCTGGATG-3‘;
RELA-sense: 5‘-ATCCCTGAGCACCATCAACTAT-3‘, RELA-antisense:
5‘-CAGGTCTTCATCATCAAACTGC-3‘; GAPDH-sense: 5‘-CCACCCATGGCAAATTCCATGGCA-3‘,
GAPDH-antisense: 5‘-TCTAGACGGCAGGTCAGGTCCACC-3‘. PCR
conditions were as follows: 95 ?C for 5 min, followed by 35 cycles of 95 ?C
for 1 min, 58 ?C for 1 min, and 72 ?C for 1 min, then extension at 72 ?C for 5
min. Six microliters of the PCR mixture was electrophoresed on 1.5% agarose gel
stained with ethidium bromide at 0.5 mg/ml. Each RT-PCR reaction was run in
triplicate.
Real-time RT-PCR
Total RNA (1 mg) was treated with DNase I to remove residual DNA contaminants and
reversely transcribed using SuperScript II reverse transcriptase (Gibco BRL).
The synthesized cDNA was used for real-time PCR amplification with SYBR Green
I PCR kit (BioWhittaker, Walkersville, USA) as recommended by the manufacturer.
The reaction was carried out in a Smart Cycler Real-Time PCR apparatus
(Bio-Rad, Hercules, USA). The primers used for real-time PCR were as follows: RAN-sense:
5‘-ATGGTGGTACTGGAAAAACGAC-3‘, RAN-antisense: 5‘-GGGATGTTTTCACACACTCGTA-3‘; b-actin-sense: 5‘-CGCACCACTGGCATTGTCAT-3‘, b-actin-antisense: 5‘-TTCTCCTTGATGTCACGCA-3‘. PCR reaction was
initiated at 95 ?C for 90 s, followed by 40 cycles of 94 ?C for 40 s, 58 ?C for
40 s and 72 ?C for 40 s, and a final extension at 72 ?C for 5 min. Melting
curve analysis (100 cycles of 45–95 ?C for 10 s) was also performed to exclude
nonspecific PCR products. A series of diluted cDNA samples were used as
templates to produce the standard curves. Each reaction was repeated three
times.
Construction of RNAi vector
targeting on RAN gene
For cloning shRAN, a vector expressing shRNA targeting against RAN,
an oligonucleotide encoding a stem-loop structure targeting RAN with the
targeting sequence AATACTACCGTGTGACCCG or TAAGTGGTCAGTTTACTGC was designed by
the web-tools at http://www1.qiagen.com/Products/GeneSilencing/CustomSiRna/SiRnaDesigner.aspx
and http://www.dharmacon.com/sidesign/default.aspx?source=0,
and then subcloned into pSUPER.retro vector (Oligoengine, Seattle, USA)
downstream H1 promoter between the BglII and HindIII sites.
Cell culture and transfection
The 5-8F NPC cells were cultured in RPMI 1640 medium supplemented with
10% fetal bovine serum in 6-well plates. Transfections were carried out using
lipofectamine 2000 reagent (Invitrogen), in accordance with the manufacturer’s
instructions. The transfected cells were selected with 2 mg/ml puromycin
(Sigma, St. Louis, USA) for 2 weeks with the medium changed every 2 days.
Individual drug-resistant clones were picked out and amplified in the same
selective medium.
Western blotting analysis
Cells were lysed in 50 ml of lysis buffer, which is comprised of 50 mM
Tris-HCl, pH 8.0, 1 mM EDTA, pH 8.0, 5 mM dithiothreitol (DTT), 2% sodium
dodecylsulfate (SDS), on ice for 30 min and ultrasonically split on ice for 30
s, and the resultant lysates were cleared by centrifugation. Proteins were
separated by 10% sodium dodecylsulfate-polyacrylamide gel electrophoresis
(SDS-PAGE) and electroblotted onto a nitrocellulose membrane, blocked with 5%
skimmed milk, and probed with anti-RAN (Upstate, Chicago, USA) or anti-a-tubulin (Santa
Cruz, Santa Cruz, USA) antibody. After incubation with horseradish
peroxidase-conjugated goat anti-rabbit secondary antibody (Upstate),
immunoblots were visualized using enhanced chemiluminescence (Amersham
Pharmacia Biotech, Piscataway, USA).
Flow cytometric analysis
Flow cytometric analysis was performed to determine the DNA content
and the cell cycle distribution of the NPC cells. Cells were harvested and
washed with 1?PBS, and fixed in ethanol with a
concentration of 70% at 4 ?C overnight. After fixation, cells were washed in 1?BS and stained with
propidium iodide (PI) for 30 min at room temperature in the dark to exclude
dead cells. Samples were then analyzed on a BD FACSCalibur system (Becton
Dickinson, Franklin Lakes, USA).
MTT assay
The 5-8F cells were plated into a 96-well plate at a density of 1?104 cells/well. After 24 h, cells were stained
with 20 ml of 5 mg/ml 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium
bromide (MTT) (Sigma) for 4 h at 37 ?C. Then the culture media were removed and
150 ml of dimethylsulfoxide (DMSO) was added and thoroughly mixed for
10 min. The absorbance was measured at 490 nm on an EL800 microplate reader
(Bio-Tek, Winooski, USA).
Detection of the expression
levels of RAN-interacting genes by RT-PCR
RT-PCR was performed to detect the expression level of RAN-interacting
genes after RNAi, including lamin A/C (LMNA), Ets2 repressor
factor (ERF), cyclin-dependent kinase inhibitor 1A [CDKN1A (P21)],
RELA, CEL, retinoblastoma 1 (RB1), myeloproliferative
leukemia virus oncogene (MPL). The PCR primers were as follows: LMNA-sense:
5‘-ACCAAGAAGGAGGGTGACCT-3‘, LMNA-antisense: 5‘-TTGTCAATCTCCACCAGTCG-3‘;
ERF-sense: 5‘-CCTGGTGTCTTCCGAGTCTATC-3‘, ERF-antisense:
5‘-GGGCTGAGGTGGTAGTTGTAGA-3‘; CDKN1A-sense: 5‘-GAAGGGACACACAAGAAGAAG-3‘,
CDKN1A-antisense: 5‘-AGCCTCTACTGCCACCATCTTA-3‘; RB1-sense:
5‘-AAGAAGTGCTGAAGGAAGCAAC-3‘, RB1-antisense: 5‘-CCATAAACAGAACCTGGGAAAG-3‘;
MPL-sense: 5‘-CACCACCACACACAGCTAATTT-3‘, MPL-antisense:
5‘-ACACCTGTAATCCCAGCACTTT-3‘.
Results
Gene selection in pathway
networks
The microarray data were summarized from the previous work of our
lab [10–13]. In total, 2210 genes were up- or down-regulated in NPC tissues,
compared with normal NP tissues. When these 2210 genes were entered into the
pathway resources at BioCarta, which is available at http://www.biocarta.com, a pathway
network consisting of 478 genes was constructed. Since abnormal cell cycle
regulation and apoptosis is often involved in tumorigenesis, apoptosis/cell
cycle-related genes were further chosen to construct a sub-network and there
appeared 155 genes in this sub-network. Table 1 lists the names of
partial genes, the number of all pathways, and the number of apoptosis/cell
cycle related pathways they participate in. They interacted with each other
through pathways. As shown in Table 2, we counted the “degree” of
each node (gene) using our computer program. To make the “degree”
data conform to normal distribution, we transformed the “degree” data
in Table 2 into their common logarithm shown in Table 3. If the
“degree” of a gene (after data conformation) is greater than the
threshold value 2.89098 (mean+2?SD), the gene should have significant
influence on the whole network. After calculating and screening, we discovered
12 such candidate genes including RAN, CEL, RELA, A1,
ASS, UP, NP, P8, NDR, ERF, C1 and
MPL. However, the threshold value for the sub-network is 2.71208, and only
the RAN, CEL and RELA genes were found to be of
“degree” higher than 2.71208, indicating that they might play more
important roles in NPC development. As an example, the pathways that RAN
participates in are shown in Table 4. To evaluate the validity of our
new strategy, we decided to choose RAN, CEL and RELA as
promising candidates in the following experiments.
Expression levels of RAN,
CEL, RELA genes in NPC tissues
RT-PCR was used to detect the mRNA expression level of the CEL
gene in 9 NP and 24 NPC tissues. The CEL gene was weakly expressed in
only 11% (1/9) of NP tissues, while it was obviously up-regulated in 83%
(20/24) of NPC tissues (P<0.05) (Fig. 2). RELA gene expression was detected in 4
NP tissues and 12 NPC tissues, but no difference in RELA expression
level was observed between NP and NPC tissues as shown in Fig. 3.RT-PCR and real-time RT-PCR were performed to detect the expression
level of the RAN gene in 8 NP and 36 NPC tissues (Figs. 4 and 5).
In the real-time RT-PCR method, a higher Ct value indicates less RNA copy in
original samples, so more cycles were required to reach the same RNA density
point on the graph. According to RT-PCR and real-time RT-PCR results, the
expression of the RAN gene was found to be up-regulated in 83% (30/36)
and 88% (30/34) of NPC tissues respectively (both P<0.05; Table 5).
Silencing of RAN gene
expression in 5-8F cells by RNAi
The strategy of constructing a shRNA vector is illustrated in Fig.
6. To determine the effect of shRAN on RAN gene expression, we
performed the semi-quantitative RT-PCR. A 295-bp fragment of RAN and a
550-bp fragment of GAPDH were amplified by RT-PCR with specific
primers. As shown in Fig. 7, the mRNA expression level of the RAN
gene was reduced in some cell clones (lanes 4 and 7) after transfection with
shRAN vector, while the mRNA level of the GAPDH gene as control was
almost unchanged. Therefore, clone 4 and/or 7 were chosen in the following experiments.
Western blotting analysis confirmed that the protein expression level of the RAN
gene was also inhibited in clones 4 and 7 using a-tubulin protein as a
loading control (Fig. 8). Furthermore, the suppressing effect of the RAN
gene on cell proliferation was determined by MTT assay. As shown in Fig. 9,
suppression of RAN in 5-8F cells (clone 4) could effectively slow down
cell proliferation. Meanwhile, flow cytometrical analysis was carried out to
examine the profile of cell cycle distribution in 5-8F cells transfected with
shRAN (5-8F-shRAN). Using 5-8F cells untransfected and transfected with blank
vector (pSUPER.retro) as controls, we found that the percentage of 5-8F-shRAN
cells in G1
phase increased by 8% and that in G2 phase decreased by 8%, while the proportion of 5-8F-shRAN cells in
S phase almost unchanged (Fig. 10).
Detection of the expression
levels of several RAN-interacting genes after RNAi
To demonstrate how RAN-interacting genes were regulated by
the down-regulation of RAN, we carried out RT-PCR to detect the changes
in expression levels of several RAN-interacting genes, including LMNA,
ERF, CDKN1A (P21), RELA, CEL, RB1 and
MPL, as these genes participated in the same pathways as RAN.
Once RAN was down-regulated by RNAi: (1) LMNA, CDKN1A (P21)
and CEL genes were also down-regulated; (2) RELA gene was
up-regulated; (3) ERF, RB1 and MPL showed no obvious
change in their mRNA expression levels (Fig. 11).
Discussion
In this study, we explored a new approach to extract valid
biological information from microarray data. From 2210 up- or down-regulated
genes in the NPC tissues, we identified a group of genes that may play
important roles in NPC development, such as RAN, CEL, RELA,
A1, ASS, UP, NP, P8, NDR, ERF,
C1 and MPL, providing a very useful clue to the analysis of
microarray data. Some tools have been developed for analyzing pathway
relationships among genes, such as GenMAPP, Pathway Processor, KnowledgeEditor,
PathMAPA, ArrayXPath, Pathway Miner and DragonView, each with advantages and
disadvantages. In our study, we applied the widely used protein network
analysis method [15] to analyze the complex gene interaction. Our study mainly
focused on the action or influence of each node on the whole network. As cancer
is not only a disease with abnormal cell proliferation and differentiation, but
also a disease with abnormal apoptosis, we also constructed a sub-network of
apoptosis and cell cycle regulation pathways. Unexpectedly, the statistical
result of the sub-network was not so meaningful compared with that of the
whole network, probably due to fewer pathways and genes involved in the
sub-network. However, it still is in agreement with the statistical result of
the whole network in the mass. Among the genes screened out by statistical analysis, RAN, CEL
and RELA were of highest statistical significance. The CEL gene
is mainly expressed in pancreas and it encodes a glycoprotein that can be
secreted from pancreas into digestive tract and from the lactating mammary
gland into milk. CEL protein plays an important role in cholesterol and
lipid-soluble vitamin ester hydrolysis and absorption. The CEL
gene contains a variable number of tandem repeat (VNTR) polymorphisms in the
coding region that may influence the function of CEL protein [17]. Our results
showed that the CEL gene was highly expressed in 83% (20/24) of NPC
samples and was only weakly expressed in 11% (1/9) of NP samples. We postulated
that the up-regulation of CEL might be one characteristic of NPC cells after their
malignant transformation and might be a biomarker candidate of NPC, in that it
was specifically up-regulated in NPC tissues.Ras-related nuclear protein is a small GTP-binding protein
belonging to the RAS superfamily that is essential for the translocation of
RNAs and proteins through the nuclear pore complex [18]. RAN protein is also
involved in the control of DNA synthesis and cell cycle progression. It was
also found that RAN played an important function in the directionality of
nuclear transport [19]. In the nucleus, RAN is found to be the form of RanGTP,
which is important for the termination of the import reaction, which is
mediated by a complex of the cargo protein with importin-b and its adapter
protein importin-a [20]. RAN also plays a crucial role in RAN-dependent spindle
assembling process during mitosis in mammalian cells. From the pathway network
analysis, we found the RAN gene located at the most important node in
the network. Therefore, we performed RT-PCR and real-time RT-PCR to analyze its
expression level to verify our statistical approach for extracting significant
information from vast microarray data. As we expected, we found that the RAN
gene was up-regulated in more than 80% of NPC tissues.We hypothesized that the RAN gene should play an important
role in the pathogenesis of NPC. To explore the function of the RAN
gene, we carried out RNAi to specifically silence its expression in NPC cells.
As a widely used method for knocking down the expression of a specific gene,
RNAi technique has been successfully used to suppress the expression of bcl-2
gene in NPC cells [21]. We constructed a shRAN vector expressing shRNA against
the RAN gene, transfected shRAN into the 5-8F NPC cells and established
two stable 5-8F-shRAN cell lines. Subsequently, MTT assay and flow cytometric
analysis were conducted and the results demonstrated that suppression of RAN
could block the proliferation and cell cycle progression of NPC cells, just as
we expected. As few downstream genes of RAN had been reported, we
carried out RT-PCR to detect the changes in expression levels of several RAN-interacting
genes, including LMNA, ERF, CDKN1A (P21), RELA,
CEL, RB1 and MPL, which participated in the same pathways
as RAN. Once RAN was down-regulated by RNAi, the LMNA, CDKN1A
(P21) and CEL genes were also down-regulated, but the RELA
gene was up-regulated. However, why and how the down-regulation of RAN
caused these changes requires further investigation.In conclusion, our study has opened up a new view for analyzing vast
microarray data and offered new clues, thus promoting the advancement of basic
and clinical medicine.
Acknowledgements
We are grateful to the other members of Molecular Pathology
Laboratory of Cancer Research Institute in Central South University (Changsha,
China) for their technical assistance and encouragement.
References
1 Shen R, Ghosh D, Chinnaiyan AM. Prognostic
meta-signature of breast cancer developed by two-stage mixture modeling of
microarray data. BMC Genomics 2004, 5: 94–100
2 Tay ST, Leong SH, Yu K, Aggarwal A, Tan SY,
Lee CH, Wong K et al. A combined comparative genomic hybridization
and expression microarray analysis of gastric cancer reveals novel molecular
subtypes. Cancer Res 2003, 63: 3309–3316
3 Hong TM, Yang PC, Peck K, Chen JJ, Yang SC,
Chen YC, Wu CW. Profiling the downstream genes
of tumor suppressor PTEN in lung cancer cells by complementary DNA microarray.
Am J Respir Cell Mol Biol 2000, 23: 355–363
4 Hegde P, Qi R, Gaspard R, Abernathy K, Dharap
S, Earle-Hughes J, Gay C et al. Identification of tumor markers in
models of human colorectal cancer using a 19,200-element complementary DNA
microarray. Cancer Res 2001, 61: 7792–7797
5 Bickel DR. Robust cluster analysis of
microarray gene expression data with the number of clusters determined
biologically. Bioinformatics 2003, 19: 818–824
6 Bouton CM, Pevsner J. DRAGON View:
Information visualization for annotated microarray data. Bioinformatics 2002,
18: 323–324
7 Toyoda T, Konagaya A. KnowledgeEditor: A new
tool for interactive modeling and analyzing biological pathways based on
microarray data. Bioinformatics 2003, 19: 433–434
8 Dahlquist KD, Salomonis N, Vranizan K, Lawlor
SC, Conklin BR. GenMAPP, a new tool for viewing and analyzing microarray
data on biological pathways. Nat Genet 2002, 31: 19–20
9 Pandey R, Guru RK, Mount DW. Pathway
Miner: Extracting gene association networks from molecular pathways for
predicting the biological significance of gene expression microarray data.
Bioinformatics 2004, 20: 2156–2158
10 Li H, Han W, Zhang L. cDNA expression array in
the differential expression profiles of p53 regulated genes in nasopharyngeal
carcinoma and the human normal nasopharynx. Zhonghua Zhong Liu Za Zhi 2001, 23:
448–450
11 Han W, Li H, Xie L, Xu L, Zhang L, Yao K.
Plerosis of cDNA array of normal human nasopharyngeal tissue and nasopharyngeal
carcinoma. Zhonghua Zhong Liu Za Zhi 2002, 24: 114–117
12 Xie L, Xu L, He Z, Zhou W, Wang L, Zhang L,
Lan K. Identification of differentially expressed genes in nasopharyngeal
carcinoma by means of the Atlas human cancer cDNA expression array. J Cancer
Res Clin Oncol 2000, 126: 400–406
13 Fang WY, Liu TF, Xie WB, Yang XY, Wang S, Ren
CP, Deng X et al. Reexploring the possible roles of some genes
associated with nasopharyngeal carcinoma using microarray-based detection. Acta
Biochim Biophys Sin 2005, 37: 541–546
14 Chung HJ, Park CH, Han MR, Lee S, Ohn JH, Kim
J, Kim J et al. ArrayXPath II: Mapping and visualizing microarray
gene-expression data with biomedical ontologies and integrated biological
pathway resources using Scalable Vector Graphics. Nucleic Acids Res 2005, 33:
W621–W626
15 A Malcom Campbell, Laurie J. Heyer.
Discovering genomics, proteomics and bioinformatics. New York: Cold Spring
Harbor Laboratory Press and Benjamin Cummings S2003
16 Gagneur J, Krause R, Bouwmeester T, Casari G.
Modular decomposition of protein-protein interaction networks. Genome Biol
2004, 5: R57
17 Bengtsson-Ellmark SH, Nilsson J, Orho-Melander
M, Dahlenborg K, Groop L, Bjursell G. Association between a polymorphism in the
carboxyl ester lipase gene and serum cholesterol profile. Eur J Hum Genet 2004,
12: 627–632
18 Nakielny S, Dreyfuss G. Transport of proteins
and RNAs in and out of the nucleus. Cell 1999, 99: 677–690
19 Ciciarello M, Mangiacasale R, Thibier C,
Guarguaglini G, Marchetti E, Di Fiore B. Importin beta is transported to
spindle poles during mitosis and regulates Ran-dependent spindle assembly
factors in mammalian cells. J Cell Sci 2004, 117: 6511–6522
20 Rexach M, Blobel G. Protein import into
nuclei: association and dissociation reactions involving transport substrate,
transport factors, and nucleoporins. Cell 1995, 83: 683–692
21 Yin ZH, Ren CP, Li F, Yang XY, Li H, Zhao M,
Yao KT. Suppression of bcl-2 gene by RNA interference increases
chemosensitivity to cisplatin in nasopharyngeal carcinoma cell line CNE1. Acta
Biochim Biophys Sin 2004, 36: 749–753