Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
Vol. 4, 3045-3050, December 1998 Clinical Cancer Research 3045
3 The abbreviations used are: HPV, human papillomavirus: RT, reverse
transcription.
Profiling of Differentially Expressed Genes in Human Primary
Cervical Cancer by Complementary DNA Expression Array’
Chanseob Shim, Wei Zhang, Chang Hun Rhee,
and Je-Ho Lee2
Center for Clinical Research, Samsung Biomedical Research Institute[C. S.], and Department of Obstetrics and Gynecology, College of
Medicine, Sung Kyun Kwan University [J-H. L.J, Samsung Medical
Center, Seoul I 35-7 10, Korea: and Department of Neuro-Oncology,
The University of Texas M. D. Anderson Cancer Center. Houston,Texas 77030 [W. Z., C. H. R.l
ABSTRACT
The profiling of differentially expressed genes fromprimary tumor samples using cDNA expression array can
reveal new tumor markers as well as target genes for ther-
apeutic intervention. Using cDNA expression array technol-
ogy, we produced an expression profile of genes that are
associated with human cervical cancer. Hybridization of thecDNA blotting membrane (588 genes on a single membrane)
was performed with 32P-labeled eDNA probes synthesized
from RNA isolated from either normal cervix or cervicalcancer. Parallel analyses of the hybridized signals enabled usto profile genes that were differentially expressed in cervicalcancer. In each experiment, the extent of hybridization ofeach gene was evaluated by comparison with the most abun-dant mRNAs in the human cervix. These include myc proto-
oncogene, 405 ribosomal protein 519, heat shock proteins,
leukosialin S (CD43), integrin otL (CD1 1A), calgranulin (A),and CDK4 inhibitor (p16�4). No detectable changes were
observed in the expression levels of these genes. Several
mRNAs, such as those encoding guanine nucleotide-binding
protein Gs (a subunit), leukocyte adhesion protein (LFA1-
1�), nuclear factor NF45, homeobox protein Hox-Al, and�-catenin were detected in increased levels in cervical can-
cer. Genes that showed decreased expression in cervicalcancer tissue were a group of apoptosis-related proteins, cell
adhesion molecules, nuclear transcription factors, and a
homeobox protein (Hox7). For example, the expression 1ev-els of Smadi and Hox7 were consistently decreased in all
Received 7/7/98: revised 9/14/98; accepted 9/21/98.
The costs of publication of this article were defrayed in part by the
payment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely toindicate this fact.
I This work was supported by G7 Strategic Fund Grant 08-01-17,
Samsung Biomedical Research Institute Fund Grant 1-95-016, KoreaScience and Engineering Foundation Grant CRC97-2-l through theCancer Research Center at Seoul National University. and Ministry ofScience and Technology of Korea Grant NG1O1O.2 To whom requests for reprints should be addressed, at Department ofObstetrics and Gynecology, College of Medicine, Sung Kyun Kwan
University, Samsung Medical Center, 50 Ilwon-Dong, Kangnam-Ku,
Seoul 135-710, Korea. Phone: 82-2-3410-3510: Fax: 82-2-3410-0044;E-mail: [email protected].
tumor tissues tested. Northern analysis of Smadi and Hox7RNA in primary cervical tumor tissues and cervical carci-
noma cell lines indicated that, in general, the mRNA levels ofthese genes were decreased in human cervical cancer. The
precise relationship between the altered expression of these
genes and cervical tumorigenesis is a matter of further
investigation.
INTRODUCTION
It is well understood that HPVs,3 especially the high-risk
HPV types 16 and 18 (1), are potent etiological entities in the
development of cervical cancer. Viral proteins (E6 and E7) from
HPV bind to tumor suppressor genes (p53 and the retinoblas-
toma genes), leading to the transformation and immortalization
of infected cells (2, 3). Inactivation of tumor suppressor genes,
by either genomic mutations or virally induced down-regulation
of gene expression, results in the deregulation of cell cycle
regulatory proteins and activation of various proto-oncogenes
(4). However, HPV infection alone is insufficient to trigger the
tumorigenic cascade in human cervical cancer cells (4). Only a
small fraction of HPV-infected women developed cervical can-
cer, indicating that additional factors and cellular events are
required for cervical carcinogenesis. These events may include
the altered expression of specific genes, such as those encoding
proto-oncogenes, tumor suppressors, cell cycle regulatory pro-
teins, intracellular signal transducers, apoptosis-associated pro-
teins, DNA synthesis/repair/recombination proteins, transcrip-
tion factors, cellular adhesion proteins, molecular chaperon
proteins, cell surface receptor proteins, and so on. General
profiles for the altered expression of such genes in cervical
cancer cells are not available.
Differential expression cloning has been used as a means of
deciphering changes in gene expression in specific cell popula-
tions. A unique mRNA subtraction technique that uses PCR has
been developed. The so-called cDNA representational differen-
tial analysis enables the detection of unique difference in gene
expression between related cells (5). Another widely used PCR-
based method for detection of changes in gene expression is
differential display (6). Differential display allows the study of
expression patterns of multiple genes by comparison of PCR-
amplified products in adjacent lanes on sequencing gels. In
addition, quantitative global analysis of expressed gene se-
quence tags by the recently developed serial analysis of gene
expression technique (7) may a provide complete catalogue of
specific gene expression patterns in the future. The large-scale
sequence data generated by the human genome project will then
be added to this global profile of expression patterns (8, 9).
However, identification of the biological functions of the many
Association for Cancer Research. by guest on August 28, 2020. Copyright 1998 Americanhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from
3046 Differential Gene Expression in Human Cervical Cancer
identified genes falls far behind the rapid expansion of the
expressed gene lists.
To begin to understand how the altered expression of a
specific gene plays an integral role in a selected biological
process, one can use the recently developed, high-throughput
technique called cDNA expression array analysis (10, 11). This
technique allows the large-scale comparison of multiple nucleic
acid sequences in a single hybridization. In this study, we used
cDNA expression array technology to analyze genes that are
differentially expressed in human cervical cancer and to identify
complex genetic alterations involved in tumorigenesis and tu-
mor progression.
MATERIALS AND METHODS
Total RNA Isolation. Cervical cancer tissues were ob-
tamed from patients in the Department of Obstetrics and Gyn-
ecology. Samsung Medical Center (Seoul, Korea). The disease
stage of tissue sample was assigned according to the classifica-
tion proposed by the clinical staging criteria of the International
Federation of Gynecology and Obstetrics. Five pairs of normal
and cervical cancer tissues were used in this study. The stages of
cancer were lb to ha. Histologically. all the biopsies were
squamous cell carcinoma. The tissues were frozen in liquid
nitrogen and then stored at -70#{176}C. Before RNA extraction, a
part of each tissue sample was sliced by paraffin section and
examined by H&E staining. Tissue samples containing >50%
tumor cells were used in this study.
Total RNAs were obtained by extracting tissues in Trizol
(Life Technologies Inc., Gaithersburg, MD) according to the
manufacturer’s instructions. Normal cervical tissues and cervi-
cal cancer tissues (- 100 mg of each) were homogenized in
Thzol solution ( 1 ml) using a Polytron homogenizer (Brinkman,
Littau, Switzerland). Homogenates were incubated for 10 mm
on ice, and then a 0.2 volume of chloroform was added to the
homogenates. After vigorous agitation for 5 mm, the inorganic
phase was separated by centrifugation at 12,000 X g for 20 mm
at 4#{176}C.RNAs were then precipitated in the presence of 1 volume
of isopropanol. RNA pellets were washed with 70% ice-cold
ethanol and then dissolved in RNase-free water. Total RNA
concentration was assessed by absorbency at 260 nm using an
UV spectrophotometer (Pharmacia, Uppsala, Sweden).
cDNA Synthesis and Hybridization. 32P-labeled
cDNAs were synthesized with the use of total RNA from both
normal and tumor tissues by RT in the presence of [a-32P]dCTP.
Briefly, total RNAs (20 p.g each) were denatured at 75#{176}Cfor 10
mm in the presence of 8 pmol of dTISVN (V = A, G, and C;
N = A, G, C, and T) mix. After the denaturation step, cDNAs
were synthesized by incubation at 37#{176}Cfor 1 h in a master mix
(total reaction volume = 40 p.1) containing 3 p.1 of dNTP (500
p.M. without dCTP). 5 p.1 of [a-32P]dCTP (3000 Ci/mmol;
Amersham Life Science, Cleveland, OH), and 1600 units of
Moloney murine leukemia virus reverse transcriptase (Promega,
Madison, WI) in 1 X RT buffer (supplied by Promega). The
reaction was terminated by heating for 10 mm at 75#{176}C, and
unincorporated nucleotides were removed by spin column pu-
rification (Chroma Spin-200: Clontech Laboratories Inc., Palo
Alto, CA). For each reaction, -2 X l0� cpm were incorporated
in the final product. After purification, labeled cDNAs were
denatured by boiling for 5 mm and then hybridized to Atlas
human cDNA array blots in hybridization solution (-2 X 106
cpm/ml; ExpressHyb hybridization solution, Clontech). Mem-
branes were prehybridized at 68#{176}Cfor at least 2 h prior to probe
addition. Hybridization was performed at 68#{176}Cin a rolling
bottle overnight. After the first two washes with 2X SSC [1 X
SSC: 0. 15 M NaCI- 15 msi sodium citrate (pH 7.0)] and 0.1%
SDS at 68#{176}Cfor 20 mm, the membranes were subjected to a
stringent wash with 0.1 X SSC, 0.5% SDS, and 0. 1 ms� EDTA
at 68#{176}C.Membranes were then exposed to X-ray film (Hyper-
film, Amersham) for 1 or 3 days at -70#{176}C. The density of each
signal was determined by Fuji MCID-Bas image analysis sys-
tem (Fuji Film, Tokyo, Japan).
Northern Hybridization. For Northern blot hybridiza-
tion, total RNA (10 p.g) was denatured in the presence of 50%
formamide, 2.2 M formaldehyde, 20 m�i 3-[N-morpholino]pro-
panesulfonic acid, 4 mM sodium acetate, and 0.5 mrvi EDTA at
65#{176}Cfor 10 mm. After electrophoresis in a 1 .2% agarose gel
containing 2.2 M formaldehyde, RNA was transferred onto a
nylon membrane (Nytran, 0.45-p.m pore size; Schleicher and
Schuell, Dassel, Germany) by capillary action under lOX SSPE
[1 X SSPE: 0.18 M NaC1, 10 mrvi Na,HPO4 (pH 7.7), and 1 msi
EDTA]. RNA transfer and loading efficiency was estimated by
staining a separate membrane with 0. 1 % methylene blue. RNA
intactness was estimated by comparing the intensities of the 285
and 185 ribosomal RNA bands. For hybridization, the mem-
brane was washed in 6X SSPE for 5 mm and air dried, and the
RNA was permanently attached to the membrane by UV illu-
mination for 1 mm. Hybridization was performed overnight in a
heat-sealable polyethylene bag containing 40 ml of hybridiza-
tion buffer [5X SSPE (pH 7.4), 5X Denhardt’s solution, 0.5%
SDS, 0.2 mg/ml heat-denatured salmon sperm DNA, and 50%
formamide] and the hybridization probe. The Smadi and Hox7
cDNAs, containing the entire coding sequences, were obtained
by PCR amplification and used for hybridization probes. 32P-
labeled cDNA probes were synthesized using a Rediprime
cDNA synthesis kit (Amersham).
RESULTS
The 32P-labeled cDNAs were synthesized by RT of total
RNAs from normal cervical and cervical tumor tissues and used
to hybridize to the cDNA array blots. After stringent washes and
exposure to X-ray film, general expression profiles of 588 genes
in normal cervical and cervical tumor tissues were obtained
(Fig. 1). By comparing hybridized blots for normal versus tumor
tissues, we identified several genes with tumor-specific changes
in expression pattern. Representative genes are marked (Fig. 1,
arrows). As also shown in Fig. 1, no signals are visible in blank
spots and negative control spots (M13 DNA, X phage DNA, and
pUC18 DNA), indicating that hybridization was highly specific.
For low-copy genes that were not detectable with 2 days of
exposure to X-ray film, exposure time was extended (data not
shown). Expression levels of specific genes, either increased or
decreased, were judged by densitometric scanning of hybridized
signals. The hybridization results obtained from five cervical
cancer patients are summarized in Table 1 . Only genes with
expression levels that were altered by > I .5-fold between nor-
Association for Cancer Research. by guest on August 28, 2020. Copyright 1998 Americanhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from
p. � � *#{149}�� � .1�.
,* . �
� . #{149}7#{149}� .9
� #{149}4S.-.�#{149}4S � .4 �s #{149}�.�S5
6
� .. ..
Clinical Cancer Research 3047
A.
B.
�1’
�,. .-�#{149}�‘ .�_.u
Fig. 1 Parallel analysis of gene expression in human cervix and cer-
vical cancer using cDNA expression array. 32P-labeled cDNAs were
synthesized by RT of total RNAs isolated from normal cervix (A) or
cervical cancer tissues (B) and hybridized to the Atlas cDNA array blots.
After stringent washes, the blots were exposed to X-ray film for 2 days.
Arrows, representative cDNA spots that show significant differences in
gene expression. Constitutively expressed genes in human cervix: 1,
Mvc: 2, RPS/9; 3, Cyclin H; 4, RPLO; 5, HSP27: 6, HSP7O; 7, Leuko-
sialin S: 8, Integrin ctL: 9, Mch4; /0, Calgranulin (A); II, Smadi; 12,
Hox7.
mal and cancer samples in more than two of the samples are
included.
Constitutive expression of the most abundant mRNAs in
the human cervix was observed in all experiments and did not
show noticeable changes between normal and tumor tissues.
These genes include, among others, myc proto-oncogene, a
series of ribosomal proteins, heat shock proteins, leukosialin S
(CD43), integrin aL (CD1 lA), and calgranulin (A) (Fig. 1,
marked by numbers). Similar expression patterns for these genes
were obtained in independent hybridizations, thus indicating
that our analysis system is reproducible.
Genes that showed increased expression in cervical cancer
tissues are guanine nucleotide binding protein Gs (a subunit),
leukocyte adhesion protein (LFAl-�3), nuclear factor NF45,
homeobox protein Hox-Al, and f3-catenin (Table 1). In four
cervical cancer samples, Gsa expression was consistently higher
than that in normal cervical tissues.
Genes with decreased expression levels in cervical cancer
tissues include a group of apoptosis-related genes, cell cycle
regulators, cell adhesion molecules, nuclear transcription fac-
tors, and homeobox proteins. Particularly interesting is Smad 1,
known as a transforming growth factor-�3 signaling protein-I
(bsp- 1 ), which showed consistently decreased expression levels
in all cancer samples (Fig. 1, 11). Northern analysis of Smadl
mRNA in primary cervical tumors also showed a decreased
expression level, suggesting that a decrease in expression of
Smadl is a general event observed in cervical cancers (Fig. 2).
In contrast, Smad 1 mRNA was detectable in four cervical car-
cinoma cell lines, HeLa, SiHa, Caski, and HT-3. Hox7 (Msxl;
Fig. 1 , 12) expression was decreased in cervical cancer but
highly expressed in human reproductive organs, including en-
dometrium, myometrium, and cervix (Fig. 2). Human ovary and
placenta did not express Hox7 mRNA (Fig. 2). In primary
cervical tumors and cervical carcinoma cells, no Hox7 mRNA
transcripts were detectable by Northern hybridization.
DISCUSSION
In this study, we profiled differentially expressed genes in
human primary cervical cancer by comparing cervical cancer
tissue mRNA with that of normal cervix using cDNA array
technology. Human cDNA array technology is used to examine
simultaneously the expression of specific genes in an entire
cDNA population on a single blot. Although human genome
projects have generated large-scale sequence data for thousands
of genes, the biological functions of such genes remain to be
deciphered. Prior to understanding the functional significance of
specific gene products, it is necessary to define differential gene
expression profiles by comparing the expression patterns of
different tissue types, developmental stages, and disease states.
Although expression analysis techniques such as RT-PCR and
Northern blot analysis have been widely used in the past, these
studies are time consuming and are only applicable to a re-
stricted number of genes. Thus, a systematic approach for ex-
amining large numbers of genes simultaneously is required.
Recently, tools for parallel analysis of gene expression in the
format of ordered gene arrays have been developed.
The Atlas human cDNA expression array system (Clon-
tech) provides a convenient method for profiling the expression
of 588 human genes in one simple experiment. The analytical
principle of human cDNA expression array is based on reverse
Northern blot hybridization. DNA fragments representing 588
human genes are immobilized in duplicate onto a nylon mem-
brane. Each cDNA fragment is 200-500 bp long and is selected
as a unique sequence without a poly(A) tail, repetitive elements,
or highly homologous sequences to minimize cross-hybridiza-
tion and nonspecific binding of a cDNA probe. 32P-labeled
cDNA probes generated by RT of total RNA samples are then
hybridized to an array blot. After a high-stringency wash, the
hybridization pattern can be analyzed by autoradiography.
Among the genes showing differential expression in cervi-
cal cancer, as compared with the normal cervix, were some
previously implicated in tumorigenesis. For example, it is well
known that abnormal G protein signaling, resulting either from
altered gene expression or mutations in the protein coding
region, plays a pathogenetic role in several human diseases (12).
In our analysis, Gsa showed increased expression in four of five
Association for Cancer Research. by guest on August 28, 2020. Copyright 1998 Americanhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from
Position#{176} Name of proteinlgene
d3k
d#{244}g
d6i
e6n
e7m
a4m
b5n
c2e
c4g
c4j
c5c
c7b
c7g
d5i
e7f
e7h
Oh
f5gf 7)
(I A gene list is available at Clontech’s world wide web site (http://www.clontech.comJclontechIAPR97UPD/Atlaslist.html).S �,_, I .5-fold or more increased/decreased: ND, not determined (because of high background or not expressed): NC. not changed.
3048 Differential Gene Expression in Human Cervical Cancer
Table I Summary of differently expressed genes in human cervical cancer
Case”
Guanine nucleotide-binding protein S Gsa
Nuclear factor NF45
Homeobox protein Hox-A I
Leukocyte adhesion protein LFA-I�3
�3-CateninMothers against DPP protein/hsp/, Smadl
Extracellular signal-related protein 2/Erk-2
ApopainJCPP32 isoform a
Ich-2 protease/lCErel-lI
Apoptosis inhibitor IAP/IAP2
WSL-LR, WSL-SI, and WSL-52 proteins
Superoxide dismutase/SOD/ [Cu/Zn]Growth arrest and DNA damage-inducible protein/GADDI53
Homeobox protein/Hox7Integrin 31
lntegrin 134Pleiotrophin
Insulin-like growth factor IA
Retinoic acid-binding protein II
I 2 3 4 5
NC + + + +
NC NC + + +
+ ND ND NC +
+ NC + + ND
+ ND ND + ND
- - ND - ND
- ND ND - -
- ND ND - ND
- ND ND - ND
- ND ND - ND
- ND ND - ND- ND ND - ND
- - ND NC ND
- - ND NC ND
- ND ND NC ND
- - ND NC ND
- ND ND - -
- - NC - NC
Normal cervix Cervical cancer
::�i#{216}�
28S� #{149}�
I. P18S ;,
.-.�
E E Cervical cancer.c .� cell lines
�C-G) � � � � � .� c� :� ����V>�O)Q)I c�l�-�
hSmadi � �
LHox7
‘I.
�by*�*#{149}i�i �GAPDH�[��, ,o �Fig. 2 Northern analysis of Smadl and Hox7 genes. Five cases of normal cervical and cervical cancer tissues were used to determine tumor-
specificity of Smad I and Hox7 expression (left). To examine tissue- and cervical carcinoma cell type-specific expression patterns, RNA was extracted
from the placenta, breast, ovary, endometrium, myometrium, and cervix, as well as human cervical carcinoma cell lines, including HeLa, SiHa, Caski,
and HT-3 (right). RNA (10 p.g) was analyzed on a 1.2% denaturing agarose gel and transferred onto a nylon membrane. 32P-labeled cDNA probes
for Smadl and Hox7 were hybridized to the RNA-blotted membranes. After stringent washes, membranes were exposed to X-ray film overnight at
-70#{176}C. For an RNA loading control, the same membranes were rehybridized with human glyceraldehyde-3-phosphate dehydrogenase (GAPDH)cDNA.
cases of cervical cancer samples (Table I ), indicating that acti- that lack Smad2 and Smad4 acquire resistance to TGF-13-medi-
vation of Gsa gene expression may be involved in the devel- ated growth inhibition, resulting in cancer progression (14).
opment of cervical cancer. In addition, the Smad proteins, These results indicate that abnormal Smad gene function may
known as downstream effectors of TGF-13-induced signaling, account for the widespread resistance to the growth inhibitory
have been suggested as putative tumor suppressors ( 13). Cells cytokine TGF-13 in human cancers ( 15 ). Smad 1 is a cellular
Association for Cancer Research. by guest on August 28, 2020. Copyright 1998 Americanhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from
Clinical Cancer Research 3049
mediator of bone morphogenic protein-induced signal transduc-
tion (16). Although there is no direct evidence for a tumor
suppressive activity of Smad 1 , our results demonstrate that
Smadl may play a role in cervical carcinogenesis.
The Hox7 gene, also known as Msxl, a homeobox-contain-
ing gene in humans ( 17), was also shown to be decreased in
cervical cancer tissues. Although the biological significance of
this finding needs further study, our result shows that altered
expression of the Hox7 gene may be related to cervical cancer
development. Recently, Marazzi et a!. (18) observed that Msx2,
a closely related homologue of Msx 1 , mediates the programmed
cell death pathway induced by bone morphogenic protein-4 in
P19 embryonal carcinoma cells. In addition, the close spatial
correlation of Msxl and Msx2 gene expression with the site of
programmed cell death has provided indirect evidence for Msx
gene function in cellular proliferation and differentiation (19).
Our Northern analyses showed that Hox7 niRNA was expressed
predominantly in reproductive organs that show cyclic changes
in growth and differentiation according to the reproductive
stages (that is, endometrium, myometrium, and cervix; Fig. 2).
This result is consistent with a report showing that, in adult
mouse, Hox7 mRNA was only detectable in the MOllerian
epithelium of the uterus, cervix, and vagina (20). In primary
cervical cancers and cervical carcinoma cell lines (HeLa, SiHa,
Caski, and HT-3), no Hox7 mRNA was detectable by Northern
hybridization, indicating that a close correlation exists between
Hox7 expression and tumorigenesis. Deregulation of homeobox
proteins Hox and Pax has been reported in various human
cancers (21-25). However, the mechanisms by which ho-
meobox gene products regulate tumeric growth are largely un-
known.
Considering that cell-surface adhesion molecules are in-
volved in the regulation of cell growth and movement, their
abnormal functions may lead to cancer development (26). Inte-
grins are cell surface heterodimeric proteins that mediate cell-
cell and cell-extracellular matrix interactions, and deregulation
of integrin gene expression is commonly observed in various
tumors (27, 28). Our detection of decreased levels ofthe integrin
subunits 131 and 134 in cervical cancer (Table 1 ) is consistent
with previous results (29-3 1). Disruption of the programmed
cell death cascade and inappropriate cell cycle regulation are
likely involved in abnormal cellular growth and tumorigenesis
(32, 33). Thus, it is not strange that cell death-related genes
involved in the apoptotic signal cascade are differentially ex-
pressed in cervical cancer.
In conclusion, our study demonstrates the use of the cDNA
array technique in monitoring the overall profile of gene expres-
sion in cervical tumorigenesis. The profiling of additional pri-
mary tumor samples will help to reveal genes that are commonly
expressed in this type of tumor. Information from these studies
may be useful in the development of therapeutic drugs for the
treatment of cervical cancer.
REFERENCES
1. Lorincz, A. T., Temple, G. F., Kurman, R. J., Jenson, A. B., and
Lancaster, W. D. Oncogenic association of specific human papillomavirus types with cervical neoplasia. J. Int. Cancer Inst., 79: 67 1-677,
1987.
2. Scheffner, M., Werness, B. A., Huibregtse, J. M., Levine, A. J., andHowley. P. M. The E6 oncoprotein encoded by human papillomavirustypes 16 and 18 promotes the degradation of p53. Cell, 63: 1 129-1 136,
1990.
3. Scheffner, M., Munger, K., Huibregtse, J. M., and Howley, P. M.
Targeted degradation of the retinoblastoma protein by human papillo-
mavirus E7-E6 fusion proteins. EMBO J., II: 2425-2431, 1992.
4. Park, T-W., Fujiwara, H., and Wright. T. C. Molecular biology of
cervical cancer and its precursors. Cancer (Phila.). 76 (Suppl.): 1902-
1913, 1995.
5. Hubank, M., and Schatz, D. G. Identifying differences in mRNA
expression by representational difference analysis. Nucleic Acid Res..
22: 5640-5648, 1994.
6. Liang, P., and Pardee, A. B. Differential display of eukaryoticmessenger RNA by means of the polymerase chain reaction. Science
(Washington DC), 257: 967-97 1 , 1992.
7. Velculescu, V. E., Zhang, L., Vogelstein, B., and Kinzler, K. W.
Serial analysis of gene expression. Science (Washington DC), 270:
484-487, 1995.
8. Hiller, L. D., Lennon, G., Becker, M., Bonaldo. M. F., and Chiapelli,
B. Generation and analysis of 280,000 human expressed sequence tags.Genome Res., 6: 807-828, 1996.
9. Aaronson, J. S., Eckman, B., and Belvins, R. A. Toward the devel-opment of a gene index to the human genome: an assessment of the
nature of high-throughput EST sequence data. Genome Res., 6: 829-845. 1996.
10. DeRisi, J., Penland. L., Brown, P. 0., Bittner, M. L., Meltzer, P. S.,
Ray, M., Chen, Y., Su, Y. A., and Trent, J. M. Use of a cDNAmicroarray to analyse gene expression patterns in human cancer. Nat.
Genet., 14: 457-460, 1996.
I 1. Heller, R. A.. Schena, M., Chai, A., Shalon, D., Bedilion, T.,
Gilmor, J., Wooley, D. E., and Davis, R. W. Discovery and analysis of
inflammatory disease-related genes using cDNA microarrays. Proc.
NatI. Acad. Sci. USA, 94: 2150-2155, 1997.
12. Weinstein, L. S., and Shenker, A. G protein mutations in humandisease. Clin. Biochem.. 26: 333-338, 1993.
13. Lechleider. R. J., de Caestecker, M. P.. Dehejia, A., Polymeropou-
los, M. H., and Roberts, A. B. Serine phosphorylation, chromosomal
localization, and transforming growth factor-13 signal transduction by
human bsp-l. J. Biol. Chem., 271: 17617-17620, 1996.
14. Nakao, A.. Roijer, E., Imamura, T., Souchelnytskyi. S.. Stenman.
G., Heldin, C. H., and Dijke, P. Identification of Smad2, a human
Mad-related protein in the transforming growth factor 13 signaling path-
way. J. Biol. Chem., 272: 2896-2900, 1997.
15. Riggins, G. J., Kinzler, K. W., Vogelstein, B., and Thiagalingam, S.Frequency of Smad gene mutations in human cancers. Cancer Res., 57:
2578-2580. 1997.
16. Liu, F., Hata, A.. Baker. J. C., Doody. J., Carcamo, J.. Harland.R. M., and Massagu#{233},J. Human Smadl, a mediator of BMP signals, is
a transcription activator. Nature (Lond.), 381: 620-623, 1996.
17. Hewitt, J. E., Clark, L. N., Ivens, A., and Williamson, R. Structure
and sequence of the human homeobox gene HOX7. Genomics, II:
670-678, 1991.
18. Marazzi, G., Wang, Y., and Sassoon, D. Msx2 is a transcriptional
regulator in the BMP4-mediated programmed cell death pathway. Dev.
Biol., 186: 127-138, 1997.
19. Davidson, D. The function and evolution of MSX genes: pointers
and paradoxes. Trends Genet.. 11: 405-41 1, 1995.
20. Pavlova, A., Boutin, E., Cunha, G., and Sassoon, S. Msxl (Hox-7.l)
in the adult mouse uterus: cellular interactions underlying regulation of
expression. Development (Camb.), 120: 335-346, 1994.
21. Barba, P., Magli. M. C., Tiberio, C., and Cillo, C. HOX geneexpression in human cancers. Adv. Exp. Med. Biol., 348: 45-57. 1993.
22. Cillo. C. HOX genes in human cancers. Invasion Metastasis, 14:
38-49, 1994.
Association for Cancer Research. by guest on August 28, 2020. Copyright 1998 Americanhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from
3050 Differential Gene Expression in Human Cervical Cancer
23. Stuart, E. T., and Gruss, P. PAX genes: what’s new in developmen-
tal biology and cancer! Hum. Mol. Genet., 4: 1717-1720, 1995.
24. Stuart, E. T., Yokota, Y., and Gruss, P. PAX and HOX in neoplasia.
Adv. Genet., 33: 255-274, 1995.
25. Anbazhagan, R., and Raman, V. Homeobox genes: molecular linkbetween congenital anomalies and cancer. Eur. J. Cancer, 33: 635-637,
1997.
26. Giancotti, F. G. Integrin signaling: specificity and control of cell
survival and cell cycle progression. Curr. Opin. Cell Biol., 9: 691-700,
1997.
27. Juliano, R. L., and Varner, J. A. Adhesion molecules in cancer: therole of integrins. Curr. Opin. Cell Biol., 5: 812-818, 1993.
28. Varner, J. A., and Cheresh, D. A. Integrins and cancer. Curr. Opin.Cell Biol., 8: 724-730, 1996.
29. Aplin, J. D., Dawson, S., and Seif, M. W. Abnormal expression of
integnn a6134 in cervical intraepithelial neoplasia. Br. J. Cancer, 74:
240-245 1996.
30. Hodivala, K. J., Pci, X. F., Liu, Q. Y., Jones, P. H., Rytina, E. R.,Gilbert, C., Singer, A., and Wart, F. M. Integrin expression and functionin HPV 16-immortalised human keratinocytes in the presence or ab-
sence of v-Ha-ras. Comparison with cervical intraepithelial neoplasia.Oncogene. 9: 943-948, 1994.
31. Carico, E., French, D., Bucci. B., Falcioni, R., Vecchione, A., and
Mariani-Costantini, R. Integrin 134 expression in the neoplastic progres-sion of cervical epithelium. Gynecol. Oncol., 49: 61-66, 1993.
32. Metcalfe, A., and Streuli, C. Epithelial apoptosis. Bioessays, 19:
711-720, 1997.
33. Wyllie, A. H. Apoptosis and carcinogenesis. Eur. J. Cell Biol., 73:
189-197, 1997.
Association for Cancer Research. by guest on August 28, 2020. Copyright 1998 Americanhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from