Research
Paper
|
|
Acta Biochim Biophys
Sin 2005,37:541-546
doi:10.1111/j.1745-7270.2005.00074.x
Reexploring the Possible Roles of Some Genes Associated with
Nasopharyngeal Carcinoma Using Microarray-based Detection
Wei-Yi FANG1,2#, Teng-Fei LIU1#,
Wei-Bing XIE2, Xu-Yu YANG1, Shuang WANG2,
Cai-Ping REN1, Xin DENG2, Qiu-Zhen LIU2,
Zhong-Xi HUANG2, Xin LI2, Yan-Qing DING2, and Kai-Tai YAO1,2*
1 Cancer Research
Institute, Xiangya Medical School, Central South University, Changsha 410078,
China;
2 Cancer Research
Institute, Nanfang Medical University, Guangzhou 510515, China
Received: March 8,
2005
Accepted: May 11,
2005
This work was
supported by a grant from the Guangzhou Science and Technology Committee
(2004z2-e0111)
# These authors
contributed equally to this work
*Corresponding
author: Tel, 86-731-4805451; Fax, 86-731-4360094; E-mail, [email protected]
Abstract In gene expression
profiling, nasopharyngeal carcinoma (NPC) 5-8F cells differ from 6-10B cells in
terms of their high tumorigenicity and metastatic ability. Differentially
expressed genes from the two cell types were analyzed by combining with MILANO
(the automatic custom annotation of microarray results which is based on all
the available published work in PubMed). The results showed that five genes, including
CTSD, P63, CSE1L, BPAG1 and EGR1, have been
studied or mentioned in published work on NPC. Subsequently, we revaluated the
roles of these genes in the pathogenesis of NPC by combining the data of gene
chips from NPCs versus NPs and pooled cells from 5-8F, 6-10B and CNE2 versus
NPs. The results suggested that the roles of BPAG1 and EGR1 are
possibly different from those reported in previous NPC studies. These five
genes are likely to be involved in the proliferation, apoptosis, invasion and
metastasis of NPC. A reexploration of the genes will further define their roles
in the pathogenesis of NPC.
Key words nasopharynx (NP);
nasopharyngeal carcinoma (NPC); MILANO; differentially expressed gene
In many published reports, the genes thought to be involved in the
pathogenesis of disease are categorized according to their up-regulation or
down-regulation in differential gene expression profiling, whereas our
conjectural results are sometimes contrary to the original hypotheses. Some
experimental results are opposite to those reported in previous studies. In
these situations, we have to question if the results reported by previous
studies are correct.Recently, we encountered a similar situation when we investigated
the differentially expressed genes from nasopharyngeal carcinoma (NPC) cells
using microarray detection. From the microarray data of NPCs versus NPs, and
the pooled cells of 5-8F, 6-10B, CNE2 versus NPs and 5-8F versus 6-10B, it is
possible for us to reanalyze the roles of the genes associated with the
pathogenesis of NPC by integrating three sets of gene chip data. In this study, we reexplorated the possible functions of five genes,
including CTSD, P63, CSE1L, BPAG1 and EGR1.
A revaluation of the genes will further elucidate their roles in the
pathogenesis of NPC.
Materials and Methods
Specimen collection
Primary tumor and normal tissues were obtained with consent during
biopsies for NPC at Hunan Cancer Hospital, Hunan Province, and Jiangmen Center
Hospital, Guangdong Province, and were immediately stored in liquid nitrogen.
Then, 102 specimens of squamous cell carcinoma with poor differentiation were
strictly screened by frozen section diagnosis and Epstein-Barr virus (EBV)
detection, and 32 specimens with EBV infection that contained more than 70% of
cancer cells were finally chosen for further research.
Cells and cell culture
5-8F (high tumorigenic and metastatic ability) and 6-10B
(tumorigenic, but lacking metastatic ability) cells from colony lines of the
NPC SUNE1 cell line [2] were provided by the Cancer Center of Sun Yet-Sen
University, (Guangzhou, China). CNE2 cells were stored in our laboratory. These
three types of NPC cells all originated from poorly differentiated squamous
cell carcinoma of NP. These cell lines were maintained in RPMI 1640
supplemented with 10% NBCS at 37 ?C in a humidified atmosphere of
5% CO2 in air.
In situ NP transplantation of 5-8F and
6-10B in nude mice
Nude mice aged 6–7 weeks
with a genetic background of BALB/c nu/nu were supplied by the Animal Center,
Nan-Fang Medical University, Guangzhou, China. The mice were maintained in a
barrier facility on HEPA-filtered racks. The animals were placed on a
autoclaved laboratory rodent diet. All animal studies were conducted in
accordance with the principles and procedures outlined in the National
Institutes of Health (NIH) Guide for the Care and Use of Animals under
assurance number A3873-1. The mice were inoculated
in situ nasopharynx with a single dose of 1?106 cells (5-8F and 6-10B). Growth and metastasis of tumor cells were
monitored every other day with the help of a whole-body optical imaging system.
Preparation of fluorescence-labeled probes and hybridization to
arrays
The experiments were performed by Shenzhen Chipscreen Biosciences
Limited, Shenzhen, China. Total RNA extracted by Trizol reagent was further
purified using Qiagen RNeasy mini kit (Qiagen Inc., USA). Next, 20 mg of total RNA
from the corresponding samples (NPC versus NP, three pooled NPC cell types,
including 5-8F, 6-10B and CNE2 1:1:1 versus NP, and 5-8F versus 6-10B) were
annealed to oligo(dT)18 and reverse-transcribed in the presence
of Cy3-dCTP and Cy5-dCTP. The resulting cDNAs were treated with 2 ml of 0.5 M NaOH,
and then the pH was adjusted to neutral by 2 ml of 0.5 M HCl. The
first-strand products of synthesis were finally purified by QIAquick PCR
purification kit (Qiagen Inc., USA) and resuspended in 30 ml of
hybridization solution. Human gene chips with approximately 8000 known or
predicted genes were purchased from Shenzhen Chipscreen Biosciences Limited.
The mixture of Cy3- and Cy5-labeled probes was hybridized against the chips in
a humidified chamber with hybridization solution (7.5 ml of 4?hybridization buffer solution, 15 ml of 50% formamide and 7.5 ml of purified
water) at 42 ?C overnight, and then washed twice for 20 min each time in 0.1%
SSC at 55 ?C. The slide was dried and scanned with a Generation III array
scanner (Amersham Pharmacia, France).
Scanning and data analysis
The scanned images were converted to digital data using Arrayvision
6.0. The ratio of the Cy3 intensity to Cy5 intensity of each spot was
calculated after data normalization with LOWESS regression. This represents the
relative gene expression level of the tested sample versus the control.
Several different statistical methods were used to set the data selection
criteria and all of them were based on hybridization intensity. One common data
selection criterion was that the intensity should be more than the average
signal value plus three standard deviations of the negative controls on the
gene chip. The signal value of these negative controls were variable because of
the various experimental conditions. Based on our previous experience with
many duplication tests, we chose one signal intensity of 5E+08 for valid data
selection. Theoretically, a ratio greater than 1.0 indicates that the
expression level is higher in the tested sample than in the control and that
the corresponding gene is up-regulated. A ratio lower than 1.0 shows that the
expression level is lower in the tested sample than in the control and that the
corresponding gene is down-regulated. We therefore chose our data selection
criteria as the intensity greater than 5E+08 and a ratio greater than 2 or less
than 0.5.
Bioinformatic analysis
Genes differentially expressed in 5-8F and 6-10B cells were analyzed
by microarray literature-based annotation (MILANO). This program performs
automatic searches in PubMed or the GeneRIF collection for articles containing
co-occurrences of search terms and a list of genes (e.g. from a microarray
experiment), and is used by pasting differentially expressed genes in
the “Primary Search Term” field, and nasopharyngeal carcinoma search
terms in the “Secondary Search Term” field. The output is a table
containing the number of hits for each pair of search terms. Subsequently, the
genes selected by MILANO analysis, together with their expression levels in NPC
cells versus NP and pooled NPC cells versus NP, were further analyzed for
their roles in the pathogenesis of NPC.
Semiquantitative RT-PCR
Purified total RNA was treated with RNase-free DNase I (TaKaRa,
Japan). After removal of the DNase I, cDNA was reverse-transcribed from 1 mg of RNA using
oligo(dT)18. Randomly selected differentially expressed genes (PDGFRA, Bcl2A1,
NK4, TGFB2 in NPC versus NP and CTSD, BPAG1, P63,
MVP, APOE, AKRIC1, AdOR3 and UGT1A9 in
pooled cells versus NP tissue) and the invariant housekeeping gene control, ACGT,
were amplified in 25 cycles from 5% of synthesized cDNA. The primer pair
designed for each gene spanned at least an intron to distinguish possibly
amplified cDNA products from genomic DNA. Subsequently, 5 ml of each PCR
reaction product was analyzed on 1.5% agarose gel, and the intensity of each
band was quantitated with the Vilber gel documentation system (Vilber Limited,
France). The signal for each gene from RT-PCR was normalized by the ACGT
gene.
Results
Metastatic and tumorigenic ability of 5-8F and 6-10B cells
The results of the in situ transplantation of NP in nude mice
showed that all 18 survivors inoculated with 5-8F cells displayed obvious
metastasis, which included encephalic invasion, jugular lymphatic node
metastases and pulmonary metastases as observed in human patients. However,
most of the 20 nude mice inoculated with 6-10B cells did not show any sign of
metastasis, with metastasis found in the lungs of two nude mice only (data
will be shown in another paper). In addition, there was an obvious difference
in tumorigenic ability between the two types of cells when they were inoculated
into the subcutis of nude mice as reported previously [2].
Identification of genes associated with nasopharyngeal carcinoma in
PubMed by gene chips and bioinformatics
There were 283 differentially expressed genes in 5-8F cells compared
with 6-10B cells. These genes were subsequently analyzed and screened by
MILANO, and five genes were finally found to be associated with NPC (Table 1
and Table 2). Among these genes, CTSD, P63, CSE1L
and BPAG1 were up-regulated more than 2-fold, and EGR1 was
down-regulated less than 0.5-fold. In addition, we also determined the
differential expression status of these genes in NPC tissues versus NP tissues
and pooled NPC cells versus NP tissues.
Semiquantitative RT-PCR
Semiquantitative RT-PCR was used to confirm the differential
expression of 12 genes chosen from gene chips of NPC tissues versus NP tissues
and 5-8F versus 6-10B. We chose these genes only because they represent
differential expression patterns and can be used as examples to confirm the
reliability of the gene chips. As shown in Figs. 1 and 2, the
genes display similar expression patterns to those of the gene chips. The
results derived from the gene chips are therefore considered reliable.Fig. 1 shows the gene-specific RT-PCR
analysis of four genes detected as differentially expressed genes between NPC
and NP in gene chip hybridization. ACTG (Actin g-1 Hs.14376) was used as
the internal control standard. The gene expression patterns are compatible to
those shown in the gene chip data of PDGFRA (0.31), TGFBR2
(0.25), NK4 (6.26) and BCL2A1 (4.11).Fig. 2 shows the gene-specific RT-PCR
analysis of eight genes detected using differential hybridization of gene chips
between 5-8F and 6-10B. ACTG (Actin g-1 Hs.14376) was used as
the internal control standard. The gene expression patterns of AdOR3
(4.82) and UGT1A9 (12.45) are comparable to those shown in the gene
chip data of CTSD (12.44), BPAG1 (5.63), P63 (6.79), MVP
(6.77), APOE (7.13) and AKRIC1 (5.23).
Discussion
Tumor metastasis is the main cause of mortality in cancer patients
[3]. The transition from in situ tumor growth to metastatic disease depends
on the ability of tumors at the primary site to invade local tissues and to
cross tissue barriers. The primary tumor is composed of numerous heterogeneous
subpopulations of cells, and only a small portion of cell subpopulations with
metastatic potential in the primary tumor have the ability to invade and cause
metastasis. This means that cells with high metastatic ability and cells with
poor metastatic ability should be easily distinguished in gene expression
profiling. Therefore, microarray analysis was performed on NPC 5-8F cells and
6-10B cells to detect the possible candidate genes associated with metastasis
and other features of NPC. A total of 283 genes were discovered. In addition,
MILANO analysis was performed on these 283 genes, and only five genes, CTSD,
P63, CSE1L, BPAG1 and EGR1, were found to be
mentioned in previous NPC studies in Pubmed. We reanalyzed the roles of the
five genes in the pathogenesis of NPC by combining the other two groups of
microarray data.Hemidesmosome (HD) is a transmembrane complex that mediates
attachment of epithelial cells to the basement membrane. BPAG1 is a
major component of hemidesmosome. Lo et al. have demonstrated the
down-regulation of BPAG1 expression in NPC cells using differential gene
display, which is consistent with our microarray analysis of pooled NPC cells
from 5-8F, 6-10B and CNE2 versus NP tissues [4]. They suggested that loss of HD
expression in NPC may be associated with the undifferentiated properties of NPC
cells and may have prognostic significance. However, Herold-Mende et al.
have found a distinct expression up-regulation with the onset of invasive
growth by in situ hybridization in squamous cell carcinomas of the head
and neck [5]. Results from our two other groups also showed the up-regulation
of BPAG1 in NPC tissues and 5-8F cells. We conclude that the cause of
the low expression of BPAG1 in NPC cells is that the tumor cells no
longer interact with the tumoral extracellular microenvironment; that is, when
NPC cells interact with their microenvironment, BPAG1 expression will be
induced and quickly up-regulated. Therefore, we think that the high expression
of BPAG1 in NPC tissues and 5-8F cells with high metastatic ability
plays a role in facilitating tumor invasion and metastasis, which is contrary
to the original report [4].The protein encoded by the EGR1 gene belongs to the EGR
family of C2H2-type zinc-finger proteins. It is a nuclear protein and
functions as a transcriptional regulator. It can be rapidly induced by growth factors
to transduce the proliferative signal. The induction of Egr1 by external
stimuli is generally transient, but appears to be sustained in some prostate
tumor cell lines and tumors, suggesting that Egr1 stimulates tumor cell
growth. In contrast, in breast, lung and brain tumors, Egr1 expression
is often absent or reduced, and when re-expressed, results in growth
suppression. Re-expression of Egr1 in tumor cells also leads to
antiapoptotic activity, which would encourage tumor cell survival [6]. However,
the role of EGR1 is still unclear. EGR1 induction is markedly
augmented in cells expressing mutant p53 that contributes to enhanced
transformed properties and resistance to apoptosis [7]. Nevertheless, NPC very
rarely has p53 gene mutations in primary tumors [8], which suggests that
the induction of Egr1 by mutant p53 is nearly impossible. EGR1
expression in our studies showed down-regulation in NPC versus NP, which is
contrary to the microarray results obtained by Xie et al. [9]. The
inconsistency between the two results might be caused by the different
specimens that were selected or the different microarray systems used. In
addition, EGR1 also displayed down-regulation in NPC versus NP. These
results suggest that EGR1 might function as a cancer suppressor gene
candidate in the pathogenesis of NPC.CTSD, P63 and CSE1L showed
differential expression only in 5-8F versus 6-10B, and they could not be easily
detected because of their extremely low expression in NPC versus NP and pooled
tumor cells versus NP. The CTSD protein is a lysosomal aspartyl protease
composed of a dimer of disulfide-linked heavy and light chains. This proteinase
is a member of the peptidase C1 family that is involved in many physiological
functions because of its proteolysic activity. CTSD has been extensively
investigated for its roles in tumor invasion and metastasis [10,11], and its
overexpression has been shown in breast cancer [12] and colon cancer [13]. In
our study, higher expression of CTSD was found in 5-8F cells, which suggests
that this gene is associated with the invasion and metastasis of NPC. P63,
a member of the p53 family of transmembrane proteins, is involved in the
survival and differentiation of reserve/stem cells in different epithelia. High
frequencies of gene gain have been detected for TP63 in oral squamous
cell carcinoma (OSCC) and human esophageal squamous cell carcinoma (EC-SCC)
cell lines [14]. TP63 transactivating isoforms, such as TAp63/p73,
show TSG properties similar to p53, while isoforms lacking the N-terminal
transactivating domain, such as deltaNp63/p73, induce a functional block
against p53 as well as TAp63/p73 activities [15].
Semiquantitative RT-PCR analysis of mRNA from 25 NPC biopsies have shown that
the dominant species expressed is invariably the truncated deltaN-isotype [8].
The high expression of TP63 in 5-8F cells suggests that TP63 in
the DN-p63 isoform acts as a suppressor of
the wild-type p53 function in NPC. CSE1L/CAS, the human homolog
of the yeast gene CSE1, is believed to be an oncogene candidate. High
frequencies of gain have been shown in NPC cells [16], medulloblastomas cells
[17] and glioblastoma multiforme (GBM) [18], etc., and its
overexpression has also been shown in prostate cancer [19]. CSE1L is thought
to be involved in cell proliferation and apoptosis. CSE1L/CAS functions
in the mitotic spindle checkpoint. It is also implicated in the nuclear to
cytoplasmic reshuffling of importin alpha, which itself is necessary for the
nuclear transport (export) of several proliferation-activating proteins,
transcription factors, oncogenes and tumor suppressor gene products, such as p53
and BRCA1 [20]. The high expression of CSE1L in 5-8F cells shows
to some extent that 5-8F cells with high tumorigenic and metastatic ability
have stronger proliferative potential and resistance to apoptosis than 6-10B
cells with low tumorigenic ability and no metastatic ability.CTSD, P63 and CSE1L cannot
be easily detected in NPC versus NP and pooled tumor cells versus NP, which suggests
that the high expression of these three genes is likely to occur mainly in
highly malignant NPC cells and maintain its strongly malignant phenotype of
cells.In summary, the results of our study suggest that the roles of BPAG1
and EGR1 may be different from those reported in previous studies of
NPC. The five genes are likely to be associated with the proliferation,
apoptosis, invasion and metastasis of NPC. A reexploration of the genes will
further define their roles in the pathogenesis of NPC.
References
1 Rubinstein R, Simon I. MILANO—custom annotation
of microarray results using automatic literature searches. BMC Bioinformatics
2005, 6: 12
2 Song LB, Yan J, Jian SW, Zhang L, Li
MZ, Li D, Wang HM. Molecular mechanisms of tumorigenesis and metastasis in
nasopharyngeal carcinoma cell sublines. Ai Zheng 2002, 21: 158–162
3 Fidler IJ. The pathogenesis of
cancer metastasis: The ‘seed and soil’ hypothesis revisited. Nat Rev Cancer 2003,
3: 453–458
4 Lo AK, Yuen PW, Liu Y, Wang XH,
Cheung AL, Wong YC, Tsao SW. Downregulation of hemidesmosomal proteins in
nasopharyngeal carcinoma cells. Cancer Lett 2001, 163: 117–123
5 Herold-Mende C, Kartenbeck J,
Tomakidi P, Bosch FX. Metastatic growth of squamous cell carcinomas is
correlated with upregulation and redistribution of hemidesmosomal components.
Cell Tissue Res 2001, 306: 399–408
6 Adamson ED, Mercola D. Egr1
transcription factor: Multiple roles in prostate tumor cell growth and survival.
Tumour Biol 2002, 23: 93–102
7 Weisz L, Zalcenstein A, Stambolsky
P, Cohen Y, Goldfinger N, Oren M, Rotter V. Transactivation of the EGR1 gene
contributes to mutant p53 gain of function. Cancer Res 2004, 64: 8318–8327
8 Crook T, Nicholls JM, Brooks L,
O’Nions J, Allday MJ. High level expression of deltaN-p63: A mechanism for the
inactivation of p53 in undifferentiated nasopharyngeal carcinoma (NPC)?
Oncogene 2000, 19: 3439–3444
9 Xie L, Xu L, He Z, Zhou W, Wang L,
Zhang L, Lan K et al. 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
10 Glondu M, Liaudet-Coopman E, Derocq D,
Platet N, Rochefort H, Garcia M. Down-regulation of cathepsin-D expression by
antisense gene transfer inhibits tumor growth and experimental lung metastasis
of human breast cancer cells. Oncogene 2002, 21: 5127–5134
11 Vigneswaran N, Zhao W, Dassanayake A,
Muller S, Miller DM, Zacharias W. Variable expression of cathepsin B and D
correlates with highly invasive and metastatic phenotype of oral cancer. Hum
Pathol 2000, 31: 931–937
12 Laurent-Matha V, Maruani-Herrmann S,
Prebois C, Beaujouin M, Glondu M, Noel A, Alvarez-Gonzalez ML et al.
Catalytically inactive human cathepsin D triggers fibroblast invasive growth.
J Cell Biol 2005, 168: 489–499
13 Shaheen RM, Miseljic S, Doering DL,
Wittliff JL. Comparison of cathepsin D determinations in human carcinomas by
enzyme immunoassay and immunoradiometric assay. J Clin Lab Anal 1995, 9: 351–358
14 Chen YJ, Lin SC, Kao T, Chang CS, Hong
PS, Shieh TM, Chang KW. Genome-wide profiling of oral squamous cell carcinoma.
J Pathol 2004, 204: 326–332
15 Benard J, Douc-Rasy S, Ahomadegbe JC.
TP53 family members and human cancers. Hum Mutat 2003, 21: 182–191
16 Hui AB, Lo KW, Teo PM, Huang DP.
Genome-wide detection of oncogene amplifications in nasopharyngeal carcinoma by
array-based comparative genomic hybridization. Int J Oncol 2002, 20: 467–473
17 Tong CY, Hui AB, Yin XL, Pang JC, Zhu XL,
Poon WS, Ng HK. Detection of oncogene amplifications in medulloblastomas by
comparative genomic hybridization and array-based comparative genomic
hybridization. J Neurosurg 2004, 100: 187–193
18 Hui AB, Lo KW, Yin XL, Poon WS, Ng HK. Detection
of multiple gene amplifications in glioblastoma multiforme using array-based
comparative genomic hybridization. Lab Invest 2001, 81: 717–723
19 Bar-Shira A, Pinthus JH, Rozovsky U,
Goldstein M, Sellers WR, Yaron Y, Eshhar Z et al. Multiple genes in
human 20q13 chromosomal region are involved in an advanced prostate cancer
xenograft. Cancer Res 2002, 62: 6803–6807
20 Behrens P, Brinkmann U, Wellmann A.
CSE1L/CAS: Its role in proliferation and apoptosis. Apoptosis 2003, 8: 39–44