Transcript Gene Expression and Cancer - CS

Gene Expression and Cancer
Presentation: Inna Weiner
Cancer
• Cellular level: over–proliferation of the cell
• Tissue level: cells deviate from their natural
place in the tissue and spread
• 3 main principles:
– Tumors are mono-clonal
– DNA mutations (6-7 usually)
– Selection (from bad to worse)
Cellular mechanisms in cancer
•
•
•
•
•
•
•
•
Signaling pathways damage
Tumor cells uncontrolled proliferation
Growth factors constitutive activity
Constitutive up/down regulation
DNA repair problem
Apoptosis mechanism not active
Cells acquire metastatic potential
…
Primary Tumor
Cancer – metastatic pathway
Articles
• A molecular signature of metastasis in primary solid
tumors.
S. Ramaswamy et al. Nature Genetics, 2002
• Robustness, scalability, and integration of a woundresponse gene expression signature in predicting
breast cancer survival.
H. Y. Chang et al. PNAS, 2005
• An oncogenic
identified
by
analysis.
KRAS2 expression signature
cross-species
gene-expression
A. Sweet-Cordero et al. Nature Genetics, 2004
A molecular signature of metastasis
in primary solid tumors
Sridhar Ramaswamy, Ken N. Ross, Eric S. Lander
& Todd R. Golub
Nature Genetics, December 2002
Motivation for Predicting Metastasis
• Metastasis (Greek: change of the state): spread of
cancer from its primary site to other places in the
body (e.g., brain, liver)
• Metastasis is the principal event leading to
death in individuals with cancer
Model of Metastasis
• Most primary tumor cells have low metastatic
potential
• Rare cells (estimated at less than 1 in
10,000,000) within large primary tumors acquire
metastatic capacity through somatic mutation
Metastatic Phenotype
• Has the ability to
– migrate from the primary tumor
– survive in blood or lymphatic circulation
– invade distant tissues
– establish distant metastatic nodules
• Supported by animal models
Setup
• 12 metastatic adenocarcinoma nodules of
diverse origin (lung, breast, prostate,
colorectal, uterus, ovary)
• 64 primary adenocarcinomas representing
the same spectrum of tumor types
Hypothesis:
a gene-expression program of metastasis
may already be present in the bulk of some
primary tumors at the time of diagnosis
Hypothesis testing
• 62 stage I/II primary
lung
adenocarcinomas
• Hierarchical clustering
in the space 128
metastases-derived
genes
Clinical Outcome Prediction
128 pre-defined
genes
17 unique
genes nearest
the centroids of
the two lung
cancer clusters
all genes
Generality of metastatic signature
17-gene metastatic signature
Upregulation: Protein translation apparatus
17-gene metastatic signature
Upregulation: Non-epithelial components of the tumor
17-gene metastatic signature
Downregulation: Antigene presenting cell
17-gene metastatic signature
Downregulation: Tumor suppressor
Novel Model of Metastasis
• Prevailing Model:
incidence of
metastasis is related
to the number of cells
susceptible to
metastasis-promoting
mutations, and hence
to tumor size
• selection process favoring
the metastatic phenotype
• rare metastatic phenotype
• New Model:
the propensity to
metastasize reflects
the predominant
genetic state of a
primary tumor
• consequence of particular
mechanisms of
transformation
• metastasis-potential tumor
Critical View
• The authors did not prove that there is a single
cell with all metastatic functions
• Maybe a small fraction of primary tumors (the
biggest?) did acquire metastatic-potential cells
Robustness, scalability, and integration
of a wound-response gene expression
signature in predicting breast cancer
survival
H. Y. Chang, D. S. A. Nuyten, J. B. Sneddon, T. Hastie, R.
Tibshirani, T. Sørlie, H. Dai, Y. D. He, L. J. van’t Veer, H.
Bartelink, M. van de Rijn, P. O. Brown, and M. J. van de
Vijver
PNAS, March 8, 2005
Chang et al (2004), PLoS
• Hypothesis:
Molecular program of normal wound
healing might play an important role in
cancer metastasis
• Procedure:
Measured gene expression of serum response
of cultured fibroblasts from 10 anatomic sites in
vitro
• Result:
Identified a set of “core serum response” genes
and their canonical expression profile in
fibroblasts activated with serum
512 core serum
response genes
were identified
and were
considered
representative
of a ‘‘wound’’
signature
Chang et al (2004):
Identified Annotations of Genes
•
•
•
•
•
Matrix remodeling
Cytoskeletal rearrangement
Cell–cell signaling
Angiogenesis
Cell motility
Likely to contribute to cancer invasion and metastasis
Robustness, scalability, and integration
of a wound-response gene expression
signature in predicting breast cancer
survival
H. Y. Chang, D. S. A. Nuyten, J. B. Sneddon, T. Hastie, R.
Tibshirani, T. Sørlie, H. Dai, Y. D. He, L. J. van’t Veer, H.
Bartelink, M. van de Rijn, P. O. Brown, and M. J. van de
Vijver
PNAS, March 8, 2005
Performance of “wound-response”
signature
295 breast cancer samples using 442 available core serum response genes
Chang et al (2004):
Clinical Outcome Prediction
Scalable Prognostic Score
• Problem: Hierarchical clustering
provides biologically arbitrary threshold
• Solution: Create the centroid of the
differential expression in response to
serum in cultured fibroblasts from 10
anatomic sites (Chang, 2004)
• Score = corr (centroid, new example)
Improving Clinical Decision Making
• 30% of women with early
breast cancer develop
metastasis
• For young women
chemotherapy increases 10
year survival at ~10%
• Chemotherapy does not benefit
for 89-93% of all breast cancer
patients
Summary
• Mechanism-driven approach to prognostic
biomarker discovery on a genome scale
• Uncovered the catalog of genes involved in a
potentially new cellular process that defines the
clinical biology of breast cancer
– pathogenic mechanisms
– potential targets for treatment
• New findings applicable for clinical decision
making
Cancer course, I. Ben-Neria
The MAP-K cascade :
Protein-Protein interactions bridging the
plasma membrane and the nucleus
Cancer course, I. Ben-Neria
RAS Activation
RAS is oncogenic due to constitutive
activation in the GTP-bound form
An oncogenic KRAS2 expression
signature identified by crossspecies gene-expression analysis.
A. Sweet-Cordero, S. Mukherjee, A Subramanian,
H. You, J.J. Roix, C. Ladd-Acosta, T. R. Golub
and T.Jacks
Nature Genetics, December 2004
Why use animal models?
• Initiated by single wellcharacterized event
• Discover novel
pathways obscure in
human data
• Endogenous activation
of oncogenes in vivo is
distinct from
overexpression in vitro
Experimental Setup
• Goal: build animal model for human lung
adenocarcinoma
• Create KrasLA mouse model: Mice with
sporadically activated Kras2 through spontaneous
homologous recombination
• Mice develop lung adenoma
• Through time acquire characteristics similar to
human tumor: nuclear atypia and high
mitotic index
Gene Set Enrichment Analysis
(GSEA)
Is Rank-Ordered Gene List (from human analysis)
enriched in independent a priori defined Gene set (from
mouse model)?
Gene Set Enrichment Analysis
(GSEA)
Comparison of Gene Expression in
mouse and human lung cancer
• Using GSEA was found
– Differentially expressed genes in KrasLA mouse
model were significantly enriched in Human Lung
Adenocarcinoma but not in other lung subtypes
– NNK mouse model (induced by chemical mutogen)
adenoma and carcinoma did not provide enriched
Differentially Expressed Gene Set
• Mouse tumor from KrasLA and NNK model
were not distinguishable histologically
Oncogenic KRAS2 signature
• 89 differentialy expressed genes (upregulated)
in KrasLA mouse model that contributed
maximally to the GSEA score in human data
set
KRAS2 signature verification (1)
• KRAS2 signature is enriched in pancreatic adenocarcinoma
• KRAS2 mutation occurs in >90% of pancreatic
adenocarcinomas
 Link between KRAS2 signature and mutation of KRAS2
KRAS2 signature verification (2)
• Real-time PCR analysis of expression of selected
KRAS2 signature genes (in human cell lines)
KRAS2 signature verification (3)
• Knock-down of KRAS2 in human lung cancer cell
line
Summary
• Integrative analysis of mouse model and human
cancer can
– Validate the animal model
– Extract an evidence of oncogene-specific program
– Compare several models against human cancer types
• In this research were identified many potential
effectors of KRAS2
– New directions for anti-Ras pathway therapeutic
strategies