Transcript Emerging issues of the expression profiling technologies for the

Emerging issues of the expression
profiling technologies for the study
of gynecologic cancer
American Journal of Obstetrics and Gynecology
(2005) 193, 908-18
R4 박영미
1
The technology of complementary DNA
microarrays and its impact in cancer biology
2

The novel experimental approaches
: the technology of complementary DNA or
oligonucleotide microarrays

The analysis of the levels of expression of
thousands of cellular genes

The establishment of distinct patterns in different
kinds of tumors
3

Genome




생물이 살아가기 위해서 필요한 최소한의 유전자군을 가지고 있
는 염색체의 1 세트
유전자(gene)와 염색체(chromosome)의 두 단어를 합성한 말
단상성(n)의 염색체, 또는 거기에 포함되는 유전 정보 전체
complementary DNA


mRNA를 Reverse transcripase란 효소를 사용하여 만든 mRNA
의 상보적인 DNA
게놈에서 전사된 1차 산물인 RNA는 단일나선으로 불안정하며
또한 수명이 짧기 때문에 연구자가 시험관내에서 취급하기가 쉽
지 않아, 이러한 RNA를 인위적으로 이중나선인 DNA로 전환시켜
연구에 사용
4
5
The principle of DNA microarrays

The differential gene expression assay between 2
tissue samples using the DNA microarray technology
: normal or reference RNA vs cancer or test RNA

Total cellular mRNA is reversed transcribed to cDNA
in the presence of 2 differentially fluor-labelled
nucleotides (UTP)
: cDNA molecules with either a green or a red
fluoresence tag
6

These tagged cDNAs are then combined and
hybridized to the microarray plate
: containing tens of thousands of immobilized
short DNA fragments of specified sequences

After hybridization and washing to remove
nonspecific binding, the DNA microarray plate is
subjected to excitation with lasers of different
wavelengths
: leading to distinct emissions from the Cy3 and Cy5
labelled probes
7

After detection by microscope scanning, the separate
image files are exported to analysis software
: they are converted to pseudo-colored images and
a visual interpretation of expression changes is
provided by merging the 2 images

The data are subjected to analysis with specific
bioinformatic tools
8
9

These gene expression data

Referred to as signatures
: because the expression patterns are distinct
and unique for each type of tumor

Can be used to histologically classify similar
tumors into specific subtypes

Providing clinically novel and relevant information
10

Reflecting the origin and function of different cell
types
 cf. Histochemical approaches
: only discriminate between malignant and
nonmalignant cells on the basis of
morphologic appearance

Anticipated to provide in the immediate future a
more accurate prognosis and prediction of
response to individual therapy
11
The molecular basis of cervical and
endometrial carcinomas
12
The prognosis of cervical & endometrial
cancer

The major conventional parameters





Clinical and/or surgical stage
Size and grade of tumor
Histologic type
Lymphatic spread
Vascular invasion
13

Additional parameters


Cytogenetic abnormalities
Acquired mutations of several proto-oncogenes,
tumor suppressor genes, other cell cycle
regulatory molecules
-> Considered as candidate molecular events in the
pathogenesis and the eventual evolution of the
tumor to metastasis
14
Human papilloma virus (HPV)

The role of HPV infection
: the development of preinvasive or invasive
carcinoma of the cervix

Recent studies that use microarrays

HPV 16 E6 oncoprotein
: Regulate differentiation-associated genes
15

HPV 16 E7 protein
: regulate several key modulators
 Signaling factors
 Cell cycle regulators
 Chaperones
: Escape immune surveillance

E2 viral protein
: Delay mitotic progression in HPV-mediated
tumorigenesis
16

The recently established microarray DNA
technology

The detection and typing of HPV infection

Sensitive high-throughput screening test for the
detection of latent HPV

Essential insights on the mechanisms of multiple
infections and the various genotypes of HPV
17
Single gene-based approaches for the
investigation of endometrial carcinomas
18

No common molecular parameters have been
formulated so far in endometrial cancer

K-ras proto-oncogene mutation

No prognostic value
: Because, no correlation to
 The stage
 The histologic type
 The grade of the tumor
19


Early endometrial cancer
: Specific for detecting submicroscopic
myometrial strips infiltrated with tumor cells
COX-2 expression


Lymph node metastasis
Parametrial invasion
20
Functional genomics of cervical and
endometrial cancer
21

Early studies : a very limited research

The identification of genes related to
radiosensitiveity of cervical squamous cell
carcinomas

The identification of genes involved in the
development of cervical carcinomas

The identification of lovostatin-induced
apoptosis-specific genes
22

The characterization of genes that are
transactivated by the PTEN tumor suppressor
gene in endometrial cells
# Limited number of samples
# Small number of genes contained in the cDNA
microarrays
→ not provided any conclusive pattern of
expression or gene signature for the
neoplasia
→ not clarified the precise genes and the
various steps of carcinogenesis
23

New studies : used cDNA microarrays
: Systematic analysis of the pattern of expression
during the various steps of cervical tumorigensis

Specific patterns of gene expression assigned to
several cellular processes
① establish early enough during carcinogenesis
② can discriminate normal cervical tissue and
LSIL from HSIL and cervical cancers
24

Many of these genes are also expressed in stroma
adjacent to the cancer tissue

The extent of gene overexpression is increased
during the progression from LSIL to HSIL and
finally to cancer
→ The identification of reliable biomarkers
associated with every stage of tumor progression
→ Eventually, leading to the improvement of early
detection
25
Gene array expression profiles of ovarian
cancer
26

4 major histologic types of ovarian carcinoma




Clear cell
Endometrioid
Mucinous
Serous
→ Directly analyzed using oligonucleotide
microarrays
27

Expression profiling patterns

Mucinous and clear cell types can be readily
distinguished from serous type, regardless of
tumor stage and type

Clear cell carcinomas seem to be more similar to a
subset of mucinous and endometrioid carcinomas
than to serous carcinomas

Endometrioid carcinmas exhibit extensive overlap
with the other histologic types
28

Clear cell ovarian carcinomas exhibit an
expression profile distinct from other poor
prognosis types
: the 73 genes upregulated in clear cell
→ consistent with the capacity of being
chemotherapy-resistant
→ hence of poor prognosis
29
Feasible pattern distinction between
normal and ovarian cancer
30

Alternative ovarian cancer model

The clear distinction between normal human
ovarian surface epithelial (HOSE) cell and
epithelial ovarian carcinoma (EOC)

Based on primary cells expanded in vitro either
from normal ovaries or from EOC

The elucidation of the molecular events occurring
at specific cell types and stages of ovarian
carcinogenesis
31

The set of differentially expressed genes in the
epithelial ovarian carcinoma cell from this study
: previously associated with ovarian tumorigenesis
: novel genes -> in embryonic pattern formation

Mxil (a single tumor suppressor gene)
: tissue-specific expression and upregulation in
normal ovarian cell
: downregulation in the EOC cell
32
Which is the “right” normal sample?
33

The continuous passage and the
nonreproducible culture conditions
: tumor cell lines may not reflect the actual
biologic events of the primary tumor

The loss of tumor markers during culture

The possibility of propagating selected
subpopulation from the original primary tissue

The putative modification of gene expression by
the in vitro expansion conditions
34

The choice of normal ovarian control
: A relevant critical issue in correctly identifying the
differentially expressed genes in tumors

The selection of the type of the normal control cell
type can influence the set of differentially
expressed genes with a tumor sample

The lack of a universal control tissue sample each
time does not permit meaningful comparisons
among similar studies

The type of normal tissue should be clearly
specified
35
Early events of ovarian carcinogenesis
36

Lack of available early malignant ovarian
tissues
: A specific problem for a systematic analysis of the
biologic phenomena of the early ovarian
carcinogenesis

Most of the ovarian tumors are discovered when
they have already proceeded beyond stage I

The required tissue size for the extraction of 10 to
50ug of total RNA for gene expression analysis
exceeds the available amount of premalignant
ovarian tissue
37

To address these difficulties

The short-term in vitro expansion of normal and
malignant ovarian epithelial cells before RNA
harvesting

The purification of ovarian epithelium

The amplification of the RNA
38
Stage-specific expression patterns
39

Genome-wide examination of chromosome
: The expression patterns of the different stages of
tumor progression of ovarian cancer

The patterns of 21 early stage : 17 late stage
-> compared by using cDNA microarray analysis,
comparative genomic hybridization
Early stage : stage I/II, endometrial or serous
carcinomas
 Late stage : stage III/IV, serous carcinomas

40

A more recent study
: Identification of gene expression patterns of the 2
survival group

Advanced stage (III or IV), serous ovarian cancer
-> short (< 2yr) : long (> 7yr) survival

Significant number of differentially expressed
genes
: IL2 receptor, chemokine ligands (CCL4, CCL5),
several interferon pathway activities
: immune system functions
: upregulation -> a favorable outcome
41
Metastasis mechanisms revisited
42

The pattern of aberrant expression in the early
stages of ovarian carcinogenesis
: The potential for metastasis is an inherent feature
of early stage cancers

Important implication on future strategies for
rational treatment and screening
 Aggressive treatment of poorly differentiated
stage I ovarian cancers

The rationale for screening strategies
43

These tumors seem
 to include subpopulations of cells exhibiting
tissue-specific profiles predicting the site of
metastasis
 to argue against the current model that
metastasis is derived from rare cells residing
within the tumor

These new data alter radically our current model of
tumor progression and metastasis
: Sequential accumulation of numerous mutations
occurring on a rare subpopulation of tumor cells
44
Therapy-related predictions
45

The prediction of response to chemotherapy
: Recent studies have addressed the feasibility of
generating novel reliable molecular markers using
gene expression profiling

The paclitaxel-induced apoptosis pathways
 The p53-independent mitochondrial pathway
 The stress reaction-induced pathway
-> Suppression of these pathways can contribute
to the acquisition of resistance to paclitaxel
46

Epothilone B

Nontaxane agent being active in paclitaxedresistant cells

The pattern of expression of ovarian cancer cell
to Epothilone B
: triggering of stress-related signal
transduction pathways associated to
TNF-a

This finding provides the impetus for further
studies to delineate the mechanisms of drug
resistance
47
Future perspectives
48

Microarray analysis
-> to characterize more efficiently human
tumors at the gene expression level

Revealing significantly and highly altered genes
between tumor and normal tissue
-> to begin deciphering the pathways mostly
affected in disease process
49

The diversification of the microarray platform
expanding beyond DNA such as




proteins
carbohydrates
peptides
nanotube precursors
-> to add a significant dimension to our
understanding of the complexity of
cellular machinery
50

The evaluation of the biologically and clinically
relevant genes
-> should be mediated by the diverse bioinformatic
protocols
① novel classifiers with a capacity for accurate and
robust cancer diagnosis
② novel statistical methods such as comparative
metaprofiling and high- throughput tissue
microarrays
51

Bioinformatic analysis
: Incorporating the clinical and pathologic
information as well as the patient history
-> We could begin to see trends important in
better tumor classification, patient
diagnosis, and prognosis
52