C. Marcelo Aldaz, M.D. Anderson Cancer Center
"Gene Expression Signature of Early Breast Cancer"
Abstract: The primary objective of this Program Project is to
generate a highly comprehensive gene expression profile of a critical
early stage in human breast cancer evolution. We will specifically focus
our studies to compare sets of lymph node negative (LNN) breast cancer
that, although having identical morphological and histopathological characteristics,
will follow significantly different clinical courses. It is of critical
importance to identify prognostic factors that would assist in the decision
of whether to subject LNN breast cancer patients to postsurgical systemic
adjuvant treatment. Thus there is much interest in the identification
and development of such prognostic tools that would allow us to better
estimate the risk of disease recurrence.
Our overall hypothesis is that clues to the prognosis of these lesions
are reflected at the time of surgical removal in the pattern of gene expression
in the primary tumor. Our ultimate goal is to identify specific "gene
expression signature profiles" that define subsets of tumors and that
ultimately will allow us to predict the clinical course of lymph node
negative breast cancers.
For this Program Project, we have assembled a unique team of oncologist,
pathologist, molecular biologists, statisticians, computer programmers
and we count with the unique patient and sample resources available at
the M.D. Anderson Cancer Center.
A major highlight of our studies is that two technologies - one comprehensive
and one targeted - are integrated to give us a higher probability of finding
key components of the molecular signatures. We feel that these state-of
the-art tools are uniquely suited to the task at hand: the discovery of
molecular signatures. Different tools may be chosen ultimately for the
analysis of gene expression signature profiles in a rapid fashion in a
clinical pathology laboratory setting.
Three important intermediate goals will also be achieved in the course
of this Program. First, we will cross validate two complementary gene
expression technologies to obtain the most comprehensive, accurate and
robust picture of the gene expression profile of lymph node negative breast
cancer. Second, we will develop new, more useful statistical models, methods
and software for the analysis and interpretation of global gene expression
profiles. Third, we will develop an understanding of the contribution
of different cell types to the overall expression profiles of breast tumors,
in order to address problems of tumor heterogeneity.
Leonard H. Augenlicht, Montefiore Medical Center
"Molecular Classification of Colon Tumors"
Abstract: The Albert Einstein Cancer Center has bean in the forefront
of development of technologies for the analysis of gene expression, beginning
with the first large scale gene array and image analysis systems developed
over a decade ago, and continuing to high throughput genomics facilities
for microarray analysis, mapping and sequencing, and novel technologies
for quantitative in situ analysis of multiple mRNA molecules at single
molecule resolution and sensitivity.
An integrated program utilizing these technologies is presented that
focuses on identifying subsets of Dukes' B2 and C colon cancer patients,
an area of basic and clinical research in which we have bean active throughout
this decade. Approximately 50,000 patients per year will have adjuvant
chemotherapy recommended following resection, and there is a critical
need to distinguish subsets who will benefit from this treatment from
subsets who either do not require further treatment, or who will not benefit,
and should be spared the expense and toxicity of treatment. The latter
patients with poor prognosis would also be candidates for more aggressive
treatment, including gene therapy. Patients will be identified and tissue
obtained through a collaborative program among the Departments of Surgery,
Pathology and Oncology for capturing and following patients on clinical
trials.
A unique feature of our application is our development and analysis of
extensive microarray databases on lineage specific differentiation of
colonic epithelial calls both accompanied by, or independent of, apoptotic
pathways, and our microarray analysis of isogenic colonic cell lines that
over-express c-myc mRNA and protein to differing degrees. These in vitro
data provide biological rationales which will help guide analyses of the
in vivo data.
In addition to fully operational facilities, other unique features of
our program include the use of "Real Time" PCR for microarray calibration
and quality control, and analysis of tissue obtained by laser capture
microscopy; linkage to major programs in structural genomics and development
and analysis of novel mouse models of colon cancer; an in place bioinformatical
program which services the genomics and genetics programs of the Cancer
Center; the collaboration of a biostatistics group with extensive experience
in basic and translational research for the planning and analysis of the
experiments; and a resource of banked normal and tumor DNA from over 700
patients entered into 2 phase III studies of adjuvant therapy for colon
cancer, including the last multi-institutional study for which there is
a control arm for analysis of the natural history of the disease, which
will permit rigorous analysis of structural gene alterations that may
underlie alterations in expression.
Jeffrey A. Boyd, Sloan-Kettering Institute for Cancer Research
"Molecular Classification of Ovarian Cancers"
Abstract: Epithelial ovarian carcinoma is the leading cause of
mortality among all gynecologic cancers. Poor survival rates associated
with this malignancy are attributable to a frequently advanced stage at
diagnosis and the inability to successively treat most cases of advanced
stage disease using current standards of surgery and chemotherapy. While
there are a number of clinical, surgical outcome, and histopathological
variables that correlate with prognosis for ovarian cancer, there exists
wide variation in the length of recurrence-free interval and survival
among typical cases, i.e., post-menopausal women with moderately to poorly
differentiated serous tumors of advanced surgical stage. This heterogeneity
in clinical outcome presumably reflects differences in the underlying
molecular genetic characteristics of individual cancers, but the molecular
basis of ovarian cancer remains largely obscure. Thus, the long-term goal
of this project is to address the hypothesis that a comprehensive molecular
genetic classification of ovarian cancer will improve our ability to predict
clinical outcome, specifically with regard to duration of recurrence-free
interval following chemotherapy, and survival. Furthermore, elucidation
of the molecular determinants of these clinical outcome parameters should
provide substantial insights into the biological basis of ovarian tumorigenesis,
as well as suggest potential new targets or strategies for ovarian cancer
therapy. We propose that this goal may be accomplished through the systematic
application of new, comprehensive molecular screening methodologies, specifically,
use of cDNA microarrays for gene expression analysis and comparative genomic
hybridization for genetic analysis, together with a large ovarian cancer
tissue resource linked to extensive surgical, histopathological, and clinical
data. The specific aims of this project are to: 1) obtain RNA expression
profiles of a large number of ovarian cancers using cDNA microarrays to
identify subgroups assembled by expression characteristics; 2) obtain
comprehensive information concerning chromosomal imbalances in ovarian
cancers by performing comparative genomic hybridizations on the same samples;
3) define distinct molecular and genetic subsets of ovarian cancer that
will predict time to recurrence following chemotherapy, and survival,
by correlating these molecular and genetic observations with clinical
and pathological outcomes; and 4) confirm and refine observations from
specific aims one and two using targeted assays and the development of
tissue-based approaches. These aims will be accomplished by a National
Cooperative Tumor Signature Group with expertise in molecular biology,
genetics, cDNA and tissue array technology, bioinformatics, gynecologic
oncology and pathology.
Patrick O. Brown, Stanford University School of Medicine
"A Cancer Taxonomy Based on Gene Expression Patterns"
Abstract: Current systems for classification of cancer group together
tumors with important differences in clinical behavior. As might have
been expected from the manifest diversity in clinical behavior, we have
found that there is enormous variation in gene expression patterns in
tumors that would classically be grouped together. The variation among
tumors in global gene expression patterns is, however, orderly and systematic
and it provides a distinctive and reproducible signature for each patient
s tumor, and even paints a picture of their biological differences. Moreover,
we have found that variation in expression profiles can highlight unrecognized
similarities and differences among tumors, and can provide a basis for
systematic clustering of subsets of tumors. We therefore believe that
underlying the apparent heterogeneity among cancers that we currently
call by the same name, there may be a systematic taxonomy that is not
readily apparent from histology or the small set of markers usually used
to define subgroups of tumors. We propose to characterize the molecular
variations among cancers of the breast, prostate, brain, and liver, by
systematically and quantitatively measuring variation in transcript abundance
for at least 20,000 different genes, in several hundred independent tumor
samples from each of these tumor types. We will use multivariate clustering
methods to search for ways to group tumors into clusters that are internally
coherent in their expression patterns and thus, we hope, in their clinical
behavior. Most of the tumor samples that are now available for the large
retrospective studies that will be required to test the clinical utility
of the new taxonomic groups we define are not suitable for analysis of
gene expression at the RNA level. They are, however, well suited to immunohistochemical
characterization. To make the transition from exploration and discovery
of the molecular variation in cancer, to testing its connection to clinical
behavior, we therefore propose to identify a large set of genes whose
expression pattern varies most, and most independently, among the tumors
we study, and raise antibodies against the predicted protein products.
These antibodies will be used for immunohistochemistry, to characterize
the variation in expression of the corresponding proteins among a diverse
set of normal tissues, tumor samples and cultured cell lines. These antibody
reagents will then be used for retrospective studies aimed at classifying
tumors for which the natural history and treatment response is already
known, to determine whether a new cancer taxonomy based on gene expression
patterns can successfully order these cancers into groups with distinctive
and consistent natural histories and patterns of response to treatment.
These antibodies will aid investigations of the molecular pathogenesis
of cancer. Some of them may provide a basis for non-invasive screens for
early detection of cancers, and others could eventually even be used therapeutically.
Thesaurus Terms: gene expression, neoplasm /cancer classification /staging,
neoplasm /cancer genetics, statistics /biometry brain neoplasm, breast
neoplasm, cooperative study, genetic registry /resource /referral center,
liver neoplasm, molecular biology information system, nucleic acid quantitation
/detection, prostate neoplasm clinical research, human data, human genetic
material tag, human subject, human tissue, immunocytochemistry, nucleic
acid hybridization
Wing C. Chan, University of Nebraska Medical Center
"Molecular Classification of B-cell Lymphoma"
Abstract: Tumors derived from the same cell type and having similar
morphology may nevertheless have a distinctly different clinical behavior
and response to therapy. Differences in the genetic lesions in these tumors,
as reflected by their gene expression profiles, will provide insight into
the mechanisms underlying the divergent clinical spectrum that is observed.
Comparative genomic hybridization (CGH) and spectral karyotyping (SKY)
are highly complementary novel techniques that examine the entire genome
for genetic abnormalities and can supplement and extend conventional cytogenetic
studies. In addition, the recently - developed high-density cDNA microarray
technology is a very promising method for displaying the pattern of gene
expression in tumor tissues. These powerful technologies with their associated
informatic systems are now available for translational research. In order
to evaluate the information generated by these technologies, an adequate
number of well-characterized tumors with detailed clinical data must be
available. We propose a multi-institutional, comprehensive molecular analysis
of a large series of B-cell non-Hodgkin's lymphoma (NHL). The molecular
data obtained will be correlated with the clinical and pathologic information
in the extensive databases kept at our institutions to identify clinically
and biologically distinct subsets of B- NHL. When unique molecular profiles
of clinical and biological significance are identified, we will then define
which components within each profile are essential determinants of the
clinical features and outcome. Specific confirmatory assays for the expression
of key genes, and the cytogenetic abnormalities involving these genes,
will be performed. Our longer term goal is to use this information to
design a simpler and less expensive microarray for diagnostic use. This
"diagnostic chip" could provide rapid molecular characterization of every
B-NHL at presentation for optimal treatment decisions and prognostication.
We also anticipate the identification of new and significant genetic alterations
that will contribute to our understanding of the key events in neoplastic
transformation and tumor progression. The insights gained from this project
may also identify novel targets for preventive and therapeutic interventions.
William L. Gerald, Sloan-Kettering Institute for Cancer Research
"Molecular Classification of Prostate Cancer"
Abstract: Prostate cancer is the most commonly diagnosed cancer
in men and affects millions of people. In recent years both the detection
and incidence have increased dramatically. The natural history of prostate
cancer is enigmatic leading to significant controversies concerning screening
tests and proper therapeutic management. The only established systemic
therapy for this disease focuses on androgen ablation. Androgen deprivation
induces programmed cell death in hormone-dependent prostatic cancer. However,
androgen-independent cells are present early in the evolution of prostate
cancer and virtually all patients eventually develop androgen-independent
tumor, a serious clinical problem for which no effective therapy exists.
The mechanisms of development and biochemical pathways that contribute
to androgen-independent growth are unknown. Existing classifications of
prostate cancer offer little information regarding prognosis or predicted
response to therapy for individual patients. Molecular profiles that distinguish
androgen-dependent from androgen-independent prostate cancer will provide
insight into the critical pathways that regulate tumor growth and response
to therapy, and can be used for classification with regard to hormone
sensitivity. The overall goal of our research program is to characterize
the molecular events underlying the clinical features of prostate cancer
and provide a basis for molecular classification and targeted therapy.
Comprehensive analyses of a human prostate cancer xenograft have demonstrated
that expression levels in large numbers of genes change dramatically in
the course of selection for androgen-independent growth. Androgen-independent
tumors have a distinct molecular profile resulting from altered expression
of many genes, some of which are known to play a role in the androgen
signal transduction pathway, but many are of unknown function. Based on
this encouraging result we expect that androgen-independent primary human
prostate cancer will have a distinct molecular profile. We propose 1)
to define molecular profiles that are representative of androgen-independent
prostate carcinoma by comprehensive, microarray-based gene expression
analysis and comparison of androgen-dependent and androgen independent
tumors, 2) to develop robust molecular histopathologic methods to identify
and quantitate androgen-dependent and -independent tumor cell subclones
within primary human prostate carcinomas for molecular classification
of individual tumors, and 3) to analyze the relationship between molecular
classification and clinical course of disease in statistically sound retrospective
and prospective studies. These studies will provide the basis for an in-depth
understanding of the role of androgens in prostate cancer, mechanisms
for development of hormone refractory disease and the clinical utility
of a molecular classification based on this information.
Samir M. Hanash, The University of Michigan Medical Center
"Toward a Molecular Classification of Tumors"
Abstract: The University of Michigan has responded to the Director's
Challenge with a proposal to utilize an integrated genomic and proteomic
approach for DNA and protein analysis, for the molecular classification
of tumors. A multi-disciplinary team with expertise in fields including
Oncology, Pathology, Molecular Biology, Cancer Prevention, Clinical Trials,
of national and international participating groups in place. The targeted
tumors consist of a judicious choice of specific groups of colon, lung
and ovarian cancers for which current classification schemes are uninformative
with respect to the clinical behavior of the tumors. A detailed rationale
for inclusion of different tumor types is provided which is based in part
on common molecular features including the occurrence of mutations in
the same pathogenetically relevant genes in the three tumor types. A special
feature of this program is integration of gene expression analysis at
both the RNA and protein levels. The program is organized into projects
and cores. Each core will serve all projects thus substantially facilitating
comparable processing of samples and integration of information across
projects. Project one is aimed at devising a molecular classification
on stages II and II colon tumors. Another project is aimed at devising
a molecular based classification of serious carcinoma of the ovary. The
last project is aimed at devising a molecular based classification for
tumors that currently belong to stage I adenocarcinoma and squamous cell
carcinoma of the ring. The first Core provides administrative support
for the program. Another Core provides support for tissue procurement,
pathologic evaluation of tissue specimens and tissue microdissection.
The next Core is concerned with proteomic analysis of tumors. It is expected
that approximately 1000 proteins will be identified in each tumor type
and their abundance in individual tumors determined by quantitative two-dimensional
electrophoresis. Another Core is concerned with tumor DNA microarray analysis.
Initially, DNA microarrays will be utilized for expression profiling and
for the analysis of mutant genes. In subsequent years DNA microarrays
designed for genomic investigations, including analysis of deletions,
amplifications, and methylation status, currently under development in
a separate project and at no cost to this program will be utilized. Thus,
for a large subset of genes that are expressed in the tumor types investigated,
expression will be determined at both the RNA and protein levels. The
last Core provides support with bioinformatics and statistical analysis.
Internal and external advisory committees will assist in the operation
of this program of research. Plans have been put into place for dissemination
of data and for addressing issues of intellectual property.
Steven W. Johnson, University of Pennsylvania
"Molecular Classification of Ovarian Tumors"
Abstract: Ovarian cancer is the fourth leading cause of cancer
deaths among women in the United States and is the most fatal gynecologic
malignancy. There is a need to be able to accurately predict outcome based
on the molecular characteristics of ovarian tumors such as subtype, grade
and degree of malignancy. In addition, understanding the molecular basis
for the response of tumors to chemotherapy is also important for the design
of new treatment regimens and for the identification of new drug targets.
Therefore, defining the gene expression or molecular profiles of ovarian
tumors will be an important step towards improving diagnosis and treatment.
Recently, significant advances have been made in methodologies directed
at the identification and quantitation of differentially expressed genes.
The technique of cDNA array screening is capable of establishing gene
expression profiles for the majority of the genes expressed in the human
genome and new developments in "real time" quantitative RT-PCR enable
the accurate determination of gene expression levels in tumor specimens
and microdissected tissue. Using this technology, we propose to: Specific
Aim number 1. Establish comprehensive molecular profiles for ovarian tumors
with respect to grade and degree of malignancy using cDNA array screening.
This will be a retrospective study in which primary ovarian tumors will
be analyzed by standard histologic methods. Messenger RNA will be isolated
from the same specimens and used as probes for cDNA array screening. The
molecular profiles that are generated will be used to define a set of
genes whose expression denotes a specific ovarian tumor grade and degree
of malignancy. Specific Aim number 2. Establish comprehensive molecular
profiles for ovarian tumors from patients that are either responsive or
non-responsive to platinum-based combination chemotherapy. The availability
of a molecular profile that predicts patient response to chemotherapy
will enable physicians to individualize treatment for ovarian cancer.
Also, the identification of genes that are associated with poor prognosis
may define targets that will lead to the design of new drugs or treatment
regimens. Specific Aim number 3. Validate the molecular profiles established
in Specific Aims number 1 and number 2 by applying quantitative assays
of gene expression to a statistically significant number of tumor samples
representing the same histologic types, outcome and response to chemotherapy.
The set of expressed cDNAs that define the type of ovarian tumor and response
to chemotherapy will be measured in a statistically significant number
of tumor samples representing each phenotype using "real time" quantitative
PCR. This high-throughput technology will facilitate the measurement of
gene expression in large numbers of tumor specimens and will establish
a standard pathologic assay for routine analyses.
Ching C. Lau, Baylor College of Medicine
"Molecular Classification of Osteosarcoma"
Abstract: Osteosarcoma is the most common malignant bone tumor
in children. Approximately 80 percent of patients present with non-metastatic
disease. After the diagnosis is made by a biopsy, treatment involves 3-4
courses of neoadjuvant chemotherapy before definitive surgery, followed
by post-operative chemotherapy. With currently available treatment, approximately
30-40 percent of patients with non-metastatic disease relapse after therapy.
There is no prognostic factor that can be used at the time of diagnosis
to predict which patients will have a high risk of relapse. The only significant
prognostic factor predicting the outcome in a patient with non-metastatic
osteosarcoma is the histopathologic response of the primary tumor resected
at the time of definitive surgery. The degree of necrosis in the primary
tumor is a reflection of the tumor response to neoadjuvant chemotherapy.
Higher degree of necrosis is associated with lower risk of relapse and
therefore better outcome. Patients with lower degree of necrosis have
a much higher risk of relapse and poor outcome even after complete resection
of the primary tumor. Unfortunately this poor outcome cannot be altered
despite modification of post-operative chemotherapy to account for the
resistance of the primary tumor to neoadjuvant chemotherapy. Thus there
is an urgent need to identify prognostic factors that can be used at the
time of diagnosis to recognize the subtypes of osteosarcomas patients
that have high risk of relapse so that more appropriate chemotherapy can
be used at the outset to improve the outcome. We propose to establish
a molecular classification system to distinguish such subsets of osteosarcoma
based on their gene expression profiles. This project will be a collaboration
among several institutions including the Texas Children's Cancer Center,
Baylor College of Medicine, Pediatric Oncology Branch, NCI, Cancer Genetics
Branch, NHGRI, Biometric Research Branch, NCI and Incyte Pharmaceuticals,
Inc. We plan to recruit 100 osteosarcoma patients who are receiving the
same therapy through a treatment protocol. Using cDNA microarrays, we
will investigate the gene expression profiles of the tumor tissues at
the time of biopsy and definitive surgery. These profiles will be correlated
with clinical outcome. In addition, we also plan to compare the gene expression
profiles of the primary tumor and those of the metastatic lesions. The
specific aims are: 1. To validate and optimize cDNA microarray technology
for gene expression profiling of clinical specimens. 2. To establish the
relevant gene expression profiles for molecular classification of osteosarcoma
by correlating these profiles with clinical outcome, chemosensitivity,
and metastatic potential.
David M. Livingston, Dana-Farber Cancer Institute
"Gene Expression Analysis of Cancers of the Prostate and Lung"
Abstract: The successful treatment of cancer is dependent upon
an accurate diagnosis of the tumor. It has become clear that while many
tumors appear indistinguishable at the morphological level, they are in
fact molecularly distinct, and such molecular distinctions can be predictive
of clinical outcome. The present research proposal lays out a strategy
for developing a molecular classification system for two of the most common
human tumors: adenocarcinoma of the lung and prostate. The classification
system will be based upon gene expression profiles obtained using DNA
microarray technologies. There are three phases to the proposed project:
1) gene expression data collection for 42,000 genes and ESTs using oligonucleotide
arrays for a series of lung and prostate adenocarcinoma patients with
known clinical outcome, 2) classification model building using both supervised
and unsupervised learning techniques, and 3) testing of the validity of
these models on an independent set of lung and prostate adenocarcinoma
samples. It is hoped that the development of a molecular classification
system for these common tumors will help to optimize the use of existing
anti-cancer therapies, and may also lay the groundwork for the development
of new therapeutic strategies targeted to patients with particular subsets
of these diseases.
Sanford D. Markowitz, Ireland Cancer Center, Case Western Reserve
University
"Gene Expression of Colon Cancers that Metastasize"
Abstract: This proposal will test the hypothesis that differences
in patterns of gene expression determine the differing biologic behaviors
between colon cancers that are curable with primary surgical therapy and
those that ultimately metastasize to the liver and kill. Additionally,
we hypothesize that in colon cancer primary tumors, only a minority of
cells will be "prometastatic", that is competent to give rise to liver
metastases, and that assays of whole primary tumor lysates may thus fail
to distinguish the crucial presence or absence of this "prometastatic"
subpopulation. To test these hypotheses, Dr. Sanford Markowitz and his
colleagues in the cancer genetics program at the Case Western Reserve
University-NCI designated Comprehensive Cancer Center have forged a collaboration
with Eos Biotechnology, Inc., a leader in gene expression array technology.
The collaboration aims to establish an accurate molecular classification
of colon cancer by focusing on a unique collection of surgically resected
colon cancer liver metastases, all of whose cells have in vivo demonstrated
metastatic ability. Using the Affymetrix human 40K GeneChip expression
array technology, these investigators will generate a comprehensive description
of global gene expression of these liver metastases. Comparing these liver
metastases versus control nonmetastatic colon cancers, that were all cured
by surgical excision, will specify a set of metastases specific genes
whose expression defines a "metastatic signature." The goal of identifying
those colon cancer primary tumors that can metastasize will be achieved
by showing that they bear "prometastatic" cells recognizable by in situ
hybridization assay of "metastatic signature" genes. Project aims are:
i) To elucidate the "metastatic signature" by comparing on Affymetrix
arrays colon cancer liver metastases versus non-metastatic colon cancer
primary tumors. ii) To identify among metastases signature genes those
specifying early metastatic events detectable by array analyses of colon
cancer primary tumors that did metastasize. iii) To use in situ hybridization
to confirm the metastatic signature of liver metastases arises from colon
epithelial cells. iv) To use in situ hybridization to detect expression
of metastatic signature genes by prometastatic cells in colon cancer primary
tumors that are simultaneous with lever metastases or are precursor of
liver metastases relapse. v) To use in situ hybridization to determine
the areas of maximum concentration of "prometastatic" cells in colon cancer
primary tumors. vi) To validate the metastatic signature and show it has
prognostic power in an independent validation archive of 350 colon cancers.
vii) To develop immunohistochemical assays for detection of the metastatic
signature.
Dan Mercola, Sidney Kimmel Cancer Center
"Molecular Characterization of Early Stage Prostate Cancer"
Abstract: This proposal will define a molecular classification
schema for organ-confined prostate cancer. The schema will be based on
patterns of gene expression and DNA methylation found within prostate
tumors and the surrounding stroma. The aim of the project is to create
a method for classifying such tumors based on tumor biology rather than
simply on histology. The development of a classification schema based
on molecular profiles would improve our ability to treat and manage organ-confined
prostate cancer. Four complementary approaches to the generation of molecular
profiles will be used. Gene expression molecular profiles will be generated
using 1) Affymetrix Gene Chip technology and 2) RNA arbitrarily primed
(RAP-PCR) based cDNA array analysis. DNA methylation molecular profiles
of will be generated using 1) methylation-sensitive array analysis and
2) restriction landmark genomic scanning (RLGS). The utility and complementarity
of each approach in relation to the others will be evaluated. Prospective
clinical data will be collected to allow for the correlation of each molecular
profile with important tumor characteristics. With the completion of this
project we expect that we will have: Defined a molecular profile for organ-confined
prostate cancer that will supplant the standard pathologic diagnosis and
staging of these tumors. Defined additional profiles that correlate with
important clinical characteristics in these patients. Such characteristics
should include both important pathologic parameters (i.e. Gleason score,
PSA levels, stage), and clinical outcome (likelihood of metastatic disease,
likelihood of hormonal, radiation or chemotherapy response, time to progression
and survival). Discovering molecular profiles within prostate tumors that
prognosticate for patient outcome will greatly improve the quality of
life for these patients.
George K. Michalopoulos, University of Pittsburgh
"Molecular Reclassification of Prostatic Cancer"
Abstract: The main aim of this proposal is to analyze gene expression
patterns in cancer of the prostate and to establish correlations with
distinct groups of cancer behavior. These cancer subgroups are currently
covered under histopathologic diagnoses that do not allow prediction of
behavior from morphologic criteria. The studies will allow us to establish
a molecular reclassification of prostate cancer based on coordinated expression
of groups of specific genes. Complete prostatectomy specimens available
in our Western Pennsylvania Prostate Tissue Bank (run by our department
of Pathology) will be processed by microdissection and used to extract
RNA. This will in turn be processed for analysis through the Affymetrix
gene chip set, based on existing active strong and long term commitment
of collaboration with the Molecular Oncology team of Hoffman LaRoche,
Inc., at Nutley, New Jersey and our department of Pathology at the University
of Pittsburgh. Our tissue bank contains complete and well stratified information
that will be used by the bioinformatics teams of HLR and Pitt to provide
correlation between coordinated expression of specific gene sets and distinct
tumor behavior. We will be processing prostate cancer samples from the
following groups: 1. Normal prostate. 2. Prostatic cancer without capsular
invasion. 3. Prostatic cancer with capsular invasion that did not progress
to systemic disease. 3. Prostatic cancer with capsular invasion that did
progress to widespread systemic disease. 4. Metastatic foci. The data
from the gene expression analysis will be processed by both the Pitt and
the HLR bioinformatics team to provide cohesive and complete correlation
from gene expression to clinical behavior, in order to establish new diagnostic
groups of prostate cancer based on molecular sub- classification. Subsequent
studies will also use the Differential Subtraction Chain technique and
Fluorescence In Site Hybridization (FISH) to conduct complete genomic
screening of the new sub-classification groups in order to detect genomic
abnormalities that correlate with the gene expression patterns in the
groups established from the above studies. While altered expression patterns
are undoubtedly to become the basis for future tumor diagnostic methodology,
repeated paradigms with all types of cancer suggest that the basis for
altered gene expression patterns in tumors is the accumulation of genomic
alterations linked to tumor progression. The integrated approach of this
proposal will allow not only molecular sub-classification of prostate
cancer but also establishment of easy to perform diagnostic tools (selective
gene expression analysis by Real Time PCR Matrix, detection of genomic
abnormality markers, etc.) that can be easily applied as predictors for
tumor behavior. Preliminary results already provide strong evidence of
correlation between invasive behavior and altered expression of specific
genes. These include altered expression of membrane bound proteases and
matrix bound growth factors, as well as increase in groups of G-protein
linked receptors and the ligands, and decrease in enzymes responsible
for their degradation.
Stanley F. Nelson, UCLA Medical Center
"Gene Expression Based Classification of Glial Tumors"
Abstract: Astrocytic brain tumors are among the most lethal and
morbid tumors of adults, often occurring during the prime of life. The
current system of diagnosis and classification of brain tumors is partially
predictive of outcomes, and remains based primarily upon morphologic criteria.
Although recent work has shown a number of genetic differences which are
critical in the oncogenesis and progression of astrocytic tumors, there
is insufficient data to develop a molecular classification system. The
availability of cDNA clones, large amounts of sequence, data and the technology
for cDNA arrays provides a platform for the large scale analysis of gene
expression in astrocytoma. We propose to identify a set of genes that
will allow the molecular characterization of brain tumors by using cDNA
microarray technology. Using a flexible microarray format will enable
us to easily alter the arrayed genes whose expression patterns are most
informative allowing us to create cost-effective glial tumor-related reagents.
It is our central hypothesis that a much more detailed analysis of the
genes that are expressed in astrocytomas will provide a more precise prognostic
ability, subgroup patients for optimal treatment, and help identify appropriate
therapeutic targets, subgroups patients for optimal treatment 1) To determine
the optimal means of sampling low grade astrocytomas, anaplastic astrocytomas,
and glioblastoma multiformes, to determine the degree of molecular heterogeneity
within astrocytic tumors, to determine whether the heterogeneity is greater
between tumors than within an individual tumor at each gene, and to determine
the level of variance of each gene on the microarray. 2) To determine
the gene expression profiles of 120 excisional glioma and meningioma brain
tumor biopsies to develop a reclassification of the tumors based on gene
expression profiles. 3) To develop a set of genes with prognostic importance
in low grade astrocytomas. 4) To validate the importance of the genes
from specific aims 2 and 3 in the prognosis of low grade astrocytomas.
Elizabeth J. Perlman, Children's Memorial Hospital
"Categorization of Wilms Tumor by Genetic Expression"
Abstract: Wilms tumor represents the most common renal neoplasm
of childhood. Remarkable success has been achieved in the therapy of WT
through the National Wilms Tumor Study Group (NWTS). Despite the responsiveness
of most WT to adjuvant chemotherapy, some tumors are unresponsive and
result in tumor progression and death. Most tumors lack distinctive histologic
or clinical features to enable targeting with more or less aggressive
chemotherapy. Recognizing biologically distinctive subsets of WT may be
useful for predicting clinical behavior and targeting therapy. The goal
of this project is to identify molecular categories of WT that have predictable
clinical properties, including propensity to metastasize and response
to therapy. We hypothesize that gene expression profiles will aid in the
recognition of these categories.
AIM ONE: To identify candidate marker genes that are differentially expressed:
Using commercially available cDNA macro-arrays, we will comprehensively
analyze gene expression in a small group of Wilms tumors. A subset of
genes whose expression varies throughout these tumors will be identified.
AIM TWO: To identify molecular categories of Wilms tumor based upon the
expression of candidate marker genes identified in Aim One: Utilizing
custom cDNA microarrays approximately 200-300 pathologically, clinically
and genetically characterized WT will be analyzed. Using computationally
assisted methods, profiles of gene expression that define molecular categories
of Wilms tumors will be identified.
AIM THREE: To verify, test, and model the new molecular categories of
Wilms tumor and to examine these new categories in their clinical, pathologic,
and genetic context. Each molecular category will be analyzed and validated
for clinical, pathologic and genetic features using the extensive resources
of the NWTS. Genes predictive of molecular categories will be verified
using in situ hybridization or immunohistochemistry. A model categorization
will be proposed and tested on 200 additional Wilms tumors.
Jerry Radich, Fred Hutchinson Cancer Research Center
"Gene Expression Profile of Progression and Response in CML"
Abstract: Chronic myeloid leukemia (CML) is a hematopoetic stem
cell disease with distinct biological and clinical features, presenting
as a relatively clinically benign state ( chronic phase ), which invariably
evolves to an incurable aggressive disease ( blast crisis ). Treatment
can range from low intensity chemotherapy to the curative yet potentially
lethal therapy of bone marrow transplantation (BMT). Unfortunately little
is known about the molecular events that trigger the evolution of chronic
phase to blast crisis. Thus, tailoring therapy to individual patient's
risk is impossible. This proposal aims to identify changes in gene expression
that occur in the evolution of the chronic phase to blast crisis, as well
as discovering gene expression patterns that are associated with good
outcomes to conventional interferon-based therapy. Specifically, we will:
1) optimize and validate the expression array technology, then 2) use
mRNA expression arrays to identify genes involved in the progression of
chronic phase to blast phase CML; and 3) identify genes associated with
good or poor outcome following conventional interferon-based therapy.
These studies will allow us to begin to study the biology of CML transformation,
and understand at a genetic level why some patients respond to conventional
therapy, while other patients are refractory to therapy, and quickly transform
to highly aggressive disease. The identification of low v high risk patients
will allow therapy to be appropriately tailored to each individual s disease.
In addition, the application of large-scale expression analysis in this
model system will be ideal to iron out unforeseen technical problems,
and thus the experience gained may be very valuable in future investigations
other more complex tumor systems.
Greg J. Riggins, Duke University Medical Center
"A Molecular Classification of Brain Tumors"
Abstract: There is a predicted 0.44 percent lifetime risk of dying
from a malignant brain tumor in the US. An overall five-year survival
of less than 30 percent attests to our lack of ability to effectively
treat these cancers. Malignant brain tumors are a very heterogeneous group
of tumors, and a logical area to apply a rational molecular-based classification
system. We plan to address classification and response to therapy as the
two main goals of this grant.
First, we plan to develop and test a gene expression based classification
of malignant gliomas and embryonal CNS tumors. We will look within large
histologically similar groups such as glioblastomas and medulloblastomas,
to sub-classify these tumors. Second, we will develop for the malignant
gliomas; a gene expression based test that predicts response to therapy.
Here the emphasis will be to make it possible to select the chemotherapy
that has the greatest chances of success, prior to starting treatment.
Our approach for identifying RNA levels that predict class or response
will be to first generate candidate using Serial Analysis of Gene Expression
(SAGE). Our experience with SAGE as part of the Cancer Genome Anatomy
Project (CGAP) indicates that this is a powerful way to initially assess
all the expressed genes. By performing a limited sequence SAGE analysis
spread out over 68 brain tumors, we can create a cost and labor effective
comprehensive profile of the candidate genes most likely to predict class
or response. The CGAP infrastructure for this analysis is in place and
running, including low-cost high throughput sequencing, ideal for this
pursuit.
It will be necessary to test or verify each candidate gene in a large
independent set of tumors. For this purpose we have chosen real-time quantitative
PCR and tissue microarrays. Real-time PCR has the advantage of being able
to produce accurate transcript levels, rapidly and economically, from
multiple small samples. Tissue microarrays have the advantage of being
able to assay protein, RNA or DNA levels, determine the location of the
expressing cell in the tissue, and utilize fixed archived samples from
hundreds of tumors simultaneously. We will be collaborating with the recognized
leaders for this technology to produce comprehensive brain tumor tissue
microarrays.
We have led the way in public release of gene expression data, in part
because the absolute and digital transcript levels from SAGE adapt well
to data sharing. All SAGE data will be immediately posted with CGAP using
our web site, SAGEmap (http://www.ncbi.nlm.nih.gov/SAGE/),
with a similar site for tissue microarray and real-time PCR data.
Louise C. Showe, The Wistar Institute
"CTCL Staging Using Gene Expression Profiles"
Abstract: The cutaneous T-cell lymphomas (CTCL) including Mycosis
fungoides (MF) and Sezary syndrome (SS) are indolent lymphomas that progress
in stages, starting with skin lesions, sometimes proceeding through a
leukemic phase with circulating tumor cells and eventually spreading to
the visceral organs. Treatments for CTCL vary in efficacy even for patients
with what appears to be similar level of disease, emphasizing the likely
existence of undetectable heterogeneity. These characteristics added to
the availability of a large archive of patient samples make it a good
candidate for tumor staging by molecular profiles. SS, the leukemia form
of CTCL will be the initial focus of these studies as it provides easy
access to large numbers of purified malignant cells. RNA from 10 patients,
with diverse patterns of disease presentation and progression, will be
analyzed during the first year against arrays of cDNA probes for 20,000
sequence verified Unigene clusters in order to determine the global gene
expression patterns of these cells. Samples will be selected from newly
diagnosed SS patients and from an archive of viably frozen SS cells including
samples collected at progressive stages of disease over a period of greater
than 10 years. Since CTCL cells represent Th-2 T-cells, RNA from healthy
donor PBL, stimulated to develop a TH-2 phenotype will be used as controls.
Genes that are over or under-expressed in patient RNAs, compared to controls,
will be candidate tumor markers for a reduced panel of genes that will
be used to screen a larger group of patient samples. In the second phase
of the study, 100 patients will be selected for gene expression studies
with a reduced panel of 1000-2000 genes. These expression profiles will
be analyzed, using statistical techniques, to identify groups of genes
that behave in a similar fashion in subsets of patients. The results of
these analyses will be a putative diagnostic panel of genes whose expression
levels describe classes of tumors. The correlation between expression
levels and tumor groups will be confirmed using alternative methods for
measuring gene expression. Finally, clinical information from patient
histories will be compared with tumors clustered by gene expression levels
to determine whether important clinical outcomes, e.g., responsiveness
to treatment, can be predicted from the specific gene expression patterns.
Concurrent with the above studies, samples from patients with MF, the
skin-associated early form of CTCL, will be queried with the panel of
genes identified as being diagnostic for SS to determine whether the same
genes are also sufficient to characterize different classes of MF. If
novel gene clusters are found, they will be added to the data base of
candidate markers. If not, up to 10 MF patients will be analyzed on 20,000-gene
filters for genes whose expression pattern distinguishes MF from SS patients.
If found, these will be added to the panel of candidate SS genes. Finally,
techniques will be developed to assay expression profiles in a clinical
setting.
Timothy J. Triche, Children's Hospital of Los Angeles
"Gene Expression in Sarcomas of Childhood and Adolescence"
Abstract: Sarcomas in children, adolescents, and young adults
account for about 10 percent of cancer in this age range, with survivals
ranging from about 95 percent in favorable rhabdomyosarcoma to nearly
10 percent in most patients with metastatic disease. Currently, there
is no known biologic reason for these vastly different behaviors. For
the majority of these sarcomas, we lack reliable methods to predictively
segregate histologically similar tumors with very different outcomes,
and we do not know the molecular basis for their cellular or disease phenotypes,
including different drug responsiveness. To address these problems, we
propose a scalable, high throughput functional genomics approach centered
on generating and analyzing large scale gene expression profiles. A primary
goal is to obtain the most comprehensive gene expression state measurements
possible for approximately 500 tumors per year. Samples will be from the
three pediatric cooperative groups that, in aggregate, account for nearly
95 percent of children in North America with cancer. To maximize gene
representation, productivity, and economics we will use a mix of two different
kinds of array measurements already established in our labs. Selected
results from array measurements will be subjected to confirmatory experimental
analyses (Northerns, quantitative PCR, tissue In Situ hybridization, immunohistochemistry,
etc.). We believe, however, that the greatest challenge in this work is
in data management and analysis. To meet this challenge all data enter
an integrated object database that is web accessible object database.
To meet this challenge, arrays are made and expression data is acquired,
and stored in web accessible object database. It is linked to MIMIR, an
evolving suite of both novel and standard clustering algorithms and statistical
methods that will be used to analyze expression data and other types of
pertinent data. Gene expression "signatures" derived from initial clustering
analyses will then be mined for correlations with clinical data and those
correlations evaluated for significance. Within this project we are also
developing ways to measure the robustness of gene expression clusters,
the strength of membership of a gene in one or more clusters, and the
relatedness of clusters with each other and with other data types. Proposed
work also includes ongoing development of user friendly interfaces for
viewing data and its annotations to help biologists use the results to
generate new hypotheses about drug targets, the biological basis for metastasis,
drug sensitivities, and tumor classification.
James C. Willey, Medical College of Ohio
"Gene Expression Indices in Chemoresistant Lung Cancer"
Abstract: Non-small cell lung cancer (NSCLC) is one of the most
common causes of cancer death in this country and it is poorly responsive
to current chemotherapeutic regimens with an overall regression rate of
only 30-50 percent. Histological categorization provides extremely limited
information regarding biological behavior of a particular NSCLC tissue.
Progress in the genome project and advances in high throughput measurement
of gene expression are providing the opportunity to re-define diagnosis
of NSCLC tissues on the basis of important phenotypes, such as chemoresistance,
rather than on the basis of histology. The primary long-term objectives
of the proposed investigation are to improve mechanistic understanding
of NSCLC chemoresistance and to develop a method for predicting which
NSCLC tumors will respond. The mechanisms of resistance likely to involve
multiple gene products. For example, in other studies it was determined
that indices comprising multiple independent gene expressions values measured
in bronchial epithelial cells correlated better than individual gene expression
values when phenotypes for malignancy (c-myc x E2F-1/p21) and risk for
lung cancer (GSTP1 x mGST x GSHPx). In preliminary studies, the H1435
non-small cell lung cancer (NSCLC) cell line is 50-fold more resistant
to carboplatin than H460. Evaluation of 20 genes putatively associated
with carboplatin chemoresistance using standardized mixtures of competitive
templates in quantitative RT-PCR revealed that glutathione transferase
(GST) p1, Bax alpha, GADD45, ERCC3, glutathione peroxidase and mGST genes
are expressed at 100, 20, 10, 6, 5, and 4-fold higher levels respectively
in H1435. These genes and other putative chemoresistance genes may be
effectively combined into gene expression indices to produce a better
marker for the chemoresistant phenotype. The over-all hypothesis of this
proposal is that patterns of individual gene expression and/or indices
comprising the expression values of multiple individual genes will provide
an effective marker for chemoresistant NSCLC tumors. A National Cooperative
Tumor Signature Group has assembled to test the hypothesis through completion
of the following specific aims. AIM 1) Measure expression of putative
chemoresistant genes in primary NSCLC tumor tissues then identify which,
if any, correlate with resistant phenotype. AIM 2) Identify gene expression
indices that correlate with NSCLC tumor chemoresistance AIM 3) Develop
a standardized mixture of competitive templates that will allow inter-laboratory
comparison of gene expression data. AIM 4) Automate the quantitative RT-PCR
method. AIM 5) Develop an internet based databank for storage of the data
acquired during this study and for storage of data acquired by other laboratories.
Cheryl L. Willman, University of New Mexico Health Science Center
"Molecular Taxonomy of Pediatric and Adult Acute Leukemia"
Abstract: Although remarkable advances have been made in the treatment
of the acute leukemias, particularly resistant forms of leukemia remain.
In 1999, 28,000 children and adults in the U.S. will be diagnosed with
leukemia and 21,000 will die of their disease. This variability in clinical
response is due in part to the tremendous heterogeneity of the disease
itself. Traditionally classified solely on the basis of morphology and
cytochemistry, the acute lymphoid or lymphoblastic leukemias (ALL) and
the acute myeloid leukemias (AML) are characterized by highly variable
clinical and biologic behavior, immunophenotypes, and chromosomal abnormalities.
Striking differences in outcome may be seen in cases with the same cytogenetic
profile, implying that more subtle genetic abnormalities also impact disease
biology and response. We hypothesize that cDNA microarray technology will
yield quantitative, orderly, and systematic gene expression profiles that
can be used to design more clinically relevant classification schemes
and to predict therapeutic response. By conducting correlative science
studies accompanying NCI-sponsored clinical trials in children and adults
affected by acute leukemia for the Pediatric Oncology Group, Children's
Cancer Study Group, and Southwest Oncology Group, and by maintaining the
largest leukemia tissue repositories in the world, we are poised to propose
the following specific aims: 1. To Further Optimize cDNA Microarray Technology
for Studies in Primary Human Leukemia Samples. 2. To Characterize the
Molecular Variations Among Highly Selected Acute Leukemia Cases Using
at Least 30,000 Genes. Cases have been selected using two approaches:
1) therapeutic response/resistance and 2) the presence of specific cytogenetic
abnormalities. Study sets in AML include: 1) patients with "primary resistant"
disease; 2) patients in long-term remission; 3) paired pre-treatment and
relapse samples; 4) patients responding or failing specific treatment
regimens; and 4) cases selected by genotype [t(8.21), inv(16), t(15;17),
t(4;11), t(9;11), and complex]. In ALL, cases are being selected prospectively
using two approaches: 1) the presence of residual disease vs. complete
molecular response during the treatment course using automated quantitative
molecular monitoring methods; and 2) by genotype [hyperdiploid, t(12;21),
t(9;22), t(1;19), and t(4;11)]. 3. To Apply Multivariate Clustering Methods
to Group Acute Leukemias That are Coherent in their Expression Patterns.
4. To Use Automated Quantitative "Real-Time" PCR Technologies to Validate
cDNA Microarray Analyses. 5. To Use High Performance Computing and Informatics
Technologies to Link Large Genomic Data Sets with Clinical Databases.
All leukemia samples have associated clinical databases containing detailed
patient information, laboratory data (cytogenetics, correlative scientific
studies), and therapeutic response data. Our experienced clinical trials
biostatisticians will work with the UNM High Performance Computing Center
(a National Supercomputing Facility) and Sandia National Laboratory (both
world leaders in massively scalable parallel computing, statistics, informatics,
and visualization tools) to meet this aim.
Timothy J. Yeatman, H. Lee Moffitt Cancer Center
"Decoding Fingerprints Portending Colon Cancer Metastasis"
Abstract: While surgical extirpation of colorectal cancer remains
the primary modality for cure, patients who have metastasized to distant
sites at the time of surgical intervention frequently die from their disease.
Unfortunately, there is no accurate means of identifying the patients
who are at risk for metastasis. Current staging systems, based only on
clinicopathologic factors, are not very precise. Moreover, attempts at
improving these staging systems, using molecular techniques to assay the
expression of single or a small number of genes, have been relatively
unsuccessful. This is likely because the process of metastasis is complex
and linked to the expression of numerous gene families and pathways. Recently,
methods have been developed which allow the analysis of gene expression
for thousands of genes in a single experiment. We hypothesize that, by
conducting comprehensive analyses of both RNA (microarray analysis) and
protein (proteomic analysis) on the same tumor specimens, molecular fingerprints
can be identified in primary tumors that portend metastasis. A three-party
consortium between H. Lee Moffitt Cancer Center, The Institute for Genomic
Research, and Large Scale Biology Corp. has been constructed to address
this hypothesis. Based on the ubiquitous presence of endogenous ribonucleases,
we predict that the success of this study will be based on our capacity
to obtain fresh tissue specimens without significant ischemic effect.
Tumors will be rationally selected to address the biological questions
we have posed. And finally, statistical analyses designed for microarray
analysis, which apply probability statistics to each and every expressed
gene, will permit us to decipher specific fingerprints from complex datasets.
|