Customizing chemotherapy for colon cancer: the potential of gene expression profiling

Drug Resistance Updates 7 (2004) 209–218
Customizing chemotherapy for colon cancer: the potential
of gene expression profiling
John M. Mariadason a,∗ , Diego Arango b , Leonard H. Augenlicht a
Department of Oncology, Montefiore Medical Center, Albert Einstein Cancer Center, 111, East 210th Street, Bronx, NY 10467, USA
b Department of Medical Genetics, Biomedicum Helsinki, Haartmaninkatu 8, FIN-00014 University of Helsinki, Finland
Received 7 May 2004; accepted 19 May 2004
The value of gene expression profiling, or microarray analysis, for the classification and prognosis of multiple forms of cancer is now
clearly established. For colon cancer, expression profiling can readily discriminate between normal and tumor tissue, and to some extent
between tumors of different histopathological stage and prognosis. While a definitive in vivo study demonstrating the potential of this
methodology for predicting response to chemotherapy is presently lacking, the ability of microarrays to distinguish other subtleties of
colon cancer phenotype, as well as recent in vitro proof-of-principle experiments utilizing colon cancer cell lines, illustrate the potential of
this methodology for predicting the probability of response to specific chemotherapeutic agents. This review discusses some of the recent
advances in the use of microarray analysis for understanding and distinguishing colon cancer subtypes, and attempts to identify challenges
that need to be overcome in order to achieve the goal of using gene expression profiling for customizing chemotherapy in colon cancer.
© 2004 Elsevier Ltd. All rights reserved.
Keywords: Microarray; Gene expression profiling; Colon cancer; 5-FU
1. Introduction
Colorectal cancer is a leading cause of cancer related
death in the western world. However, current treatment
strategies for this disease are far from optimal, due in part
to the inability to accurately distinguish subgroups of patients that differ in their prognosis and the probability of
response to treatment.
Based upon their histopathology, colorectal cancers are
classified utilizing either the traditional Dukes staging system or the tumor node mucosa (TNM) staging system
(Compton, 2002). While use of the TNM staging system
is now becoming standard practice, we will use the Dukes
staging system throughout this review as it is the predominant staging system used in the studies discussed herein.
Colorectal cancers are classified as Dukes A when the
tumor is confined to the mucosa, Dukes B when locally
advanced, Dukes C if positive for lymph node metastasis,
and as Dukes D when distant metastases are detected at the
time of diagnosis (Compton, 2002). While the Dukes (and
TNM) staging system identifies broad patient groups that
vary in their long-term prognosis, considerable heterogene∗
Corresponding author. Tel.: +1 718 920 2025; fax: +1 718 882 4464.
E-mail address: [email protected] (J.M. Mariadason).
1368-7646/$ – see front matter © 2004 Elsevier Ltd. All rights reserved.
ity exists within each of these stages with regards to both
prognosis and response to chemotherapy.
For treatment of colorectal cancer, 5-fluorouracil (5-FU)
remains the agent with the highest clinical efficacy (Moertel
et al., 1995a,b), and (often in combination with other agents),
has become the standard treatment for patients with locally
advanced and metastatic disease. However, approximately
80% of these patients do not benefit from this treatment,
either because they have been surgically cured and need
no further treatment, or because their tumor is refractory to
5-FU-based chemotherapy. Therefore, there is clear need to
develop biomarkers capable of distinguishing between these
patient sub-groups.
First, patients unlikely to respond to 5-FU can be spared
the toxicity, time, and expense associated with this treatment
regimen, and more important, can be placed on alternate
therapies. To this end, irinotecan and oxaliplatin are now
effective alternative treatment options (Conti et al., 1996;
Cvitkovic and Bekradda, 1999; de Gramont et al., 1997;
Machover et al., 1996), while the anti-VEGF antibody and
inhibitor of tumor angiogenesis, bevacizumab, was recently
approved by the US FDA as a first-line therapy for metastatic
colon cancer (Ferrara et al., 2004). Many chemotherapeutic
agents also may promote the acquisition of multidrug resistance (Morin, 2003; Vasilevskaya and O’Dwyer, 2003).
J.M. Mariadason et al. / Drug Resistance Updates 7 (2004) 209–218
Administration of the agent likely to induce maximal response in the first course of treatment is therefore critical to
enhance overall treatment success.
Second, identification of markers that predict response
may provide important biological insights into the mechanisms that determine response to therapy. Particularly, such
studies may identify avenues by which tumor cells can be
manipulated in order to maximize response to the primary
agent. A good example was the identification of thymidylate synthase (TS) as the enzyme targeted by 5-FU, and the
subsequent discovery that an excess of intracellular reduced
folates are necessary for maximal inhibition of TS activity
(Kinsella et al., 1997). In turn, clinical trials demonstrated
that administration of 5-FU + folinic acid (leucovorin) resulted in improved response rates compared to administration of 5-FU alone (Dannenberg, 2002; Kinsella et al., 1997).
For these reasons, the identification of biomarkers capable of predicting 5-FU response in colon cancer has been
a topic of intense investigation (Longley et al., 2003). A
number of studies have examined the predictive value of
the 5-FU target enzyme, TS, and related enzymes that affect 5-FU metabolism including thymidine phosphorylase
(TP) and dipyrimidine dehydrogenase (DPD) (Metzger et
al., 1998; Salonga et al., 2000). However, while several reports have linked low TS expression with improved response
to 5-FU in vivo, others have shown no relationship between
these parameters (Allegra et al., 2003; Berglund et al., 2002;
Findlay et al., 1997; Johnston et al., 2003). Likewise, in vitro,
while TS is often overexpressed in cell lines selected for
resistance to 5-FU by continuous exposure (Johnston et al.,
1992), studies of unselected panels of cell lines have failed
to consistently show a correlation between intrinsic cellular TS levels and 5-FU response (Grem et al., 2001;
Mirjolet et al., 1998), suggesting alterations in TS levels
may be more closely linked to acquired rather than intrinsic
resistance to 5-FU. The predictive efficacy of TP is also
unclear, with both high and low levels of TP linked to 5-FU
response depending on whether the studies were performed
in vitro or in vivo, respectively (Metzger et al., 1998; Saito
et al., 1999; Schwartz et al., 1995).
Several studies also suggest that factors involved in regulating cell growth and apoptosis, including p53, myc, and the
ratio of anti-apoptotic to pro-apoptotic bcl-2 family members (Violette et al., 2002), can predict 5-FU response. For
example, improved response to 5-FU and prolonged survival
has been observed in patients with tumors wild type for
p53 (Ahnen et al., 1998; Benhattar et al., 1996; Goh et al.,
1995), although contrasting findings have also been observed
(Allegra et al., 2003). In our own investigations, we have
established that low level amplification of c-myc was associated with longer overall survival in response to 5-FU-based
adjuvant therapy (Augenlicht et al., 1997). Furthermore,
these findings were extended to demonstrate that tumors
with amplification of c-myc, that also retained wild-type p53
function, had significantly improved response to 5-FU both
in vitro and in vivo (Arango et al., 2001). The mechanistic
explanation for this interaction was the demonstration that
high levels of c-myc represses p53 induction of p21WAF1
in response to drug treatment, promoting the induction of
apoptosis over cell cycle arrest (Arango et al., 2003; Seoane
et al., 2002).
It has also been reported that various allelic deletions, and
mismatch repair status, may identify tumor subsets with differential 5-FU sensitivity. For example, tumors that retain
heterozygosity at either 17p or 18q show improved response
to 5-FU-based adjuvant therapy (Barratt et al., 2002). Likewise, tumors that are mismatch repair deficient have been
reported to show improved response to 5-FU (Elsaleh et al.,
2000, 2001; Elsaleh and Iacopetta, 2001), although studies
reporting no difference, and the converse, have also been
published (Barratt et al., 2002; Ribic et al., 2003). While
plausible explanations for many of these discrepancies have
been offered, including, for example, differences in the median age (which in turn may effect DNA methylation status)
of the patient cohorts in the studies linking MMR status with
response to 5-FU (Iacopetta et al., 2003; Ribic et al., 2003),
the predictive efficacy of these markers remains insufficient
to permit their use in routine clinical practice. A further limitation of the use of these predictors is that they are often
designed to predict response to a specific agent (5-FU), and
thus generally fail to identify alternative treatment options.
A robust assay, capable of predicting the probability of response of a given tumor to the multiple chemotherapeutic
regimens that are increasingly becoming available, would
therefore, have significant clinical utility.
2. Gene expression profiling for the classification and
prognosis of colorectal cancer
The development of colorectal tumorigenesis requires mutation of multiple key genes (Fodde et al., 2001). In turn,
these mutations, through both direct and indirect mechanisms and clonal selection, result in alterations in expression of hundreds if not thousands of genes (Notterman et al.,
2001; Zou et al., 2002). Therefore, it is not surprising that
colorectal tumors vary significantly in terms of prognosis
and response to therapy. The complexity of the genetic abnormalities that define a given colorectal tumor argue that
an assay capable of collectively considering all of this variability may be more informative for the classification and
determination of prognosis and response to therapy. The sequencing of the human genome, combined with the development of high throughput screening technologies such as
microarray analysis, now makes such an approach possible.
The utility of gene expression profiling for the classification and prognosis of most cancer types, including leukemia
(Bullinger et al., 2004; Golub et al., 1999), lung (Beer et al.,
2002; Bhattacharjee et al., 2001), breast (Sorlie et al., 2001;
Van de Vijver et al., 2002; Van ’t Veer et al., 2002), brain
(Pomeroy et al., 2002), gastric (Boussioutas et al., 2003; Tay
et al., 2003), prostate (Dhanasekaran et al., 2001) and ovarian
J.M. Mariadason et al. / Drug Resistance Updates 7 (2004) 209–218
cancer (Schaner et al., 2003) is now clearly established. With
regards to colorectal cancer, almost two decades ago Augenlicht et al. demonstrated that gene expression profiling could
successfully distinguish between normal colonic mucosa,
benign adenoma and malignant carcinoma (Augenlicht et al.,
1987), and furthermore, that profiling the normal-appearing
colonic mucosa could distinguish patients who were at elevated risk for tumor development (FAP and HNPCC family members), from low risk individuals (Augenlicht et al.,
1991). More recently, and since the advent of modern microarray technologies, several studies have demonstrated the
ability of gene expression profiling to distinguish among
other clinically important colorectal cancer subgroups.
Multiple studies have demonstrated the ability of microarrays to successfully distinguish between colonic tumor
tissue and adjacent normal mucosa. Importantly, this can
be done in an unsupervised manner, with no prior selection of differentially expressed genes (Alon et al., 1999;
Bertucci et al., 2004; Notterman et al., 2001; Zou et al.,
2002). Illustrating an additional layer of sensitivity of this
methodology, Notterman et al. demonstrated that gene expression profiling could separate normal colonic tissue from
adenoma, as well as adenoma from adenocarcinoma in an
unsupervised manner (Notterman et al., 2001). Extending
this further still, Bertucci et al. recently demonstrated that
metastatic (Dukes D) versus non-metastatic colon tumors
(Dukes A–C) could be distinguished using microarrays in
an unsupervised manner (Bertucci et al., 2004). Finally, two
studies from Torben Orntoft’s group demonstrated the capability of gene expression profiling to distinguish normal
tissue from colorectal tumors of different Dukes staging,
in an unsupervised manner (Birkenkamp-Demtroder et al.,
2002; Frederiksen et al., 2003). However, while these results are encouraging, not all tumor stages were classified
correctly, indicating the need for further methodological
improvements, and possibly the use of supervised data
analysis strategies, to improve classification.
While unsupervised analyses have been successful at
discriminating colon tumor from normal tissue, and to a
lesser extent among colon tumors of different histopathological stage, the identification of gene expression profiles
or signatures associated with more subtle subgroups of
colorectal cancer have required the use of more complex,
supervised, analytical approaches. Selaru et al. used a training set of samples to develop and train an artificial neural
network (ANN) capable of discriminating sporadic colorectal adenomas and cancers, from inflammatory bowel
disease-associated dysplasias and cancers. Importantly,
when applied to a validation set of samples, this ANN was
able to accurately discriminate between these two tumor
types (Selaru et al., 2002). Gene subsets differentially expressed between tumors from the left and right colon, or
from tumors with or without microsatellite instability, have
also been described (Mori et al., 2003, 2004). However,
while the gene lists generated in these studies are certainly
interesting biologically, for the most part their predictive
powers have not been tested either internally (using a cross
validation strategy such as a “leave one out” analysis), or
externally, using an independent sample set.
Collectively, these studies have significant clinical implications for improving the management of colorectal cancer.
The ability of expression profiling to predict prognosis or
the likelihood of metastasis (Bertucci et al., 2004), identifies patients that require more rigorous follow up and/or
more aggressive treatment. Gene expression profiling has
also shed considerable light on the molecular mechanisms
and pathways that characterize transformation of the colonic
epithelium, and have identified a number of potential targets
for treatment. Examples of the latter include the secreted
integrin-binding protein osteopontin, which was identified
as progressively upregulated in parallel with the progression of colorectal carcinoma from adenoma, to B1 through
D stage cancers, and finally liver metastases (Agrawal et al.,
2002, 2003). Likewise, using the SAGE methodology, Saha
et al. identified the protein tyrosine phosphatase, PRL-3, as
a gene consistently upregulated in metastatic colon cancers
(Saha et al., 2001). Similarly, and consistent with previous
reports (Sarris and Lee, 2001), nucleoside diphosphate kinase (NM23) was identified using a combination of gene expression profiling and tissue microarrays, to be upregulated
in colon tumors compared to the normal colonic mucosa,
and further, to have higher expression levels in tumors with
more favorable outcome (Bertucci et al., 2004).
3. Gene expression profiling for the prediction of
response of colon cancer cells to chemotherapy
While the ability of gene expression profiling to distinguish between tumor tissue and normal colonic mucosa, as
well as among colon tumors of different histopathological
grade has been demonstrated, a definitive study demonstrating the utility of this approach for prediction of response of
colon cancer to chemotherapy is presently lacking. Nevertheless, a number of encouraging studies, both in other cancer types, and in colon cancer cell lines in vitro, highlight
the potential of gene expression profiling for this purpose.
For example, Kihara et al. demonstrated the feasibility of
microarray-based expression profiling to predict survival in
esophageal cancer patients receiving 5-FU based adjuvant
chemotherapy (Kihara et al., 2001). The authors developed a
“drug resistance score” based upon 52 genes each of whose
level of expression was correlated with survival, and thus
possibly response to 5-FU. This drug resistance score was
shown to accurately predict survival in six independent patient samples (Kihara et al., 2001).
Focusing specifically on colon cancer, we recently utilized
a panel of 30 colon carcinoma cell lines to identify genes
correlated with response to 5-FU in vitro (Mariadason et al.,
2003) (Fig. 1). Furthermore, using a “leave one out” cross
validation strategy, we formally demonstrated the ability of
these genes to predict response to 5-FU-induced apoptosis.
J.M. Mariadason et al. / Drug Resistance Updates 7 (2004) 209–218
agents, or specific combinations of therapies, would be most
appropriate for treatment of a specific tumor.
However, while this study illustrates the potential of this
methodology for predicting response of colon cancer to specific chemotherapeutic agents, it is limited by the fact that it
was performed in vitro. Therefore, the immediate challenge
remains the demonstration of the utility of gene expression
profiling for the prediction of response in patients. Collection of such gene expression and supporting clinical data is
ongoing at our and other institutions.
4. Gene expression profiling for understanding
mechanisms of 5-FU action
An additional utility of gene expression profiling studies
is that they have provided a number of interesting biological
insights into both the factors that determine 5-FU response,
and the pathways induced upon 5-FU exposure.
4.1. Determinants of intrinsic 5-FU resistance and
likelihood of response
Fig. 1. Genes correlated with 5-FU-induced apoptosis. Response of panel
of 30 colon cancer cell lines to 5 ␮M 5-FU-induced apoptosis was determined by propidium iodide staining and FACS analysis following 72 h.
Basal gene expression ratios of the same 30 cell lines were correlated with
5-FU-induced apoptosis, and 420 significantly correlated genes identified.
Genes more highly expressed in 5-FU-sensitive cells are shown in cluster
A, and those more highly expressed in 5-FU resistant cells are shown in
cluster B. The predictive efficacy of this gene subset was subsequently
validated using a leave one out cross validation analysis.
Importantly, this study demonstrated that measurement of
multiple, rather than single markers, results in more accurate
prediction of drug response when compared to four previously reported determinants of 5-FU response—TS and TP
activity, and p53 and MMR status. The study also illustrated
the potential of gene expression profiling to predict response
to each of many chemotherapeutic agents. This was shown
by the fact that reanalysis of the same microarray data was
able to identify a signature capable of predicting response
to camptothecin (CPT). We have subsequently demonstrated
similar efficacy for prediction of response to oxaliplatin, the
non-steroidal anti-inflammatory drug sulindac, and the histone deacetylase-inhibitor butyrate, using the same gene expression data (unpublished findings). The ability to predict
probability of response to different agents from a single assay has the added potential of determining whether single
To identify genes that determine the probability of response to 5-FU, several studies have correlated basal levels of gene expression with the magnitude of 5-FU-induced
apoptosis or growth arrest in panels of cell lines or human
tumor xenografts. Consistent with published reports, some
of these studies identified previously established determinants of 5-FU response. In the landmark study of Scherf et
al, which generated a large matrix of data linking the basal
gene expression profiles of the NCI panel of 60 cell lines
with response to the multitude of drugs tested in the NCI
screen, dipyrimidine dehydrogenase (DPD) was identified as
highly negatively correlated with 5-FU response. This finding is consistent with DPD being the rate limiting enzyme
in 5-FU metabolism, and supports the idea that high DPD
levels may confer resistance to 5-FU by reducing exposure
of cells to the active forms of 5-FU (Scherf et al., 2000).
Similarly, Zembutsu et al. determined the basal gene expression profile of 85 human xenografts from a variety of cancers, and identified TS as negatively correlated with 5-FU
response (Zembutsu et al., 2002).
These studies also identified a number of novel links between basal levels of gene expression and 5-FU response.
Using a panel of 39 cell lines from various tumor types,
Dan et al. identified the anti-apoptotic gene survivin as more
highly expressed in 5-FU resistant cells, while members of
the aldo-keto reductase and aldehyde dehydrogenase families, and galectin 4, were more highly expressed in 5-FU
responsive cells (Dan et al., 2002). Establishing a previously
unrecognized link, Zembutsu et al. reported a significant
correlation between expression levels of mdr3 and mdr4,
and 5-FU response. In our own study using a panel of 30
colon cancer cell lines, we identified a positive correlation
between basal expression levels of the pro-apoptotic Bak
J.M. Mariadason et al. / Drug Resistance Updates 7 (2004) 209–218
protein and sensitivity to 5-FU. We validated this finding by
demonstrating that Bak was localized to the mitochondria
following 5-FU treatment, which in turn was linked to release of cytochrome c (Mariadason et al., 2003). Consistent
with this finding, induction of Bak protein in response to
5-FU in colon cancer cell lines has previously been demonstrated (Nita et al., 1998). In addition to Bak, 5-FU response
has also been linked to the basal levels of expression or mutation status of other bcl-2 family members including Bcl-2,
Bcl-xL , Bax and Bid (Sax et al., 2002; Violette et al., 2002;
Zhang et al., 2000). Our study also identified a subset of
genes with a role in protein processing and folding, including a number of chaperones, as more highly expressed in
5-FU resistant cell lines. Chaperones can play a role in protecting cells from environmental stress by binding denatured
proteins and dissociating protein aggregates (Leppa and
Sistonen, 1997), raising the possibility that higher basal expression of such proteins may enhance the ability of these
cells to survive following 5-FU-induced genotoxic stress.
Consistent with this, an upregulation of stress response
genes, including chaperonin 10 (Maxwell et al., 2003) and
vitamin D3-upregulated protein 1 (Takahashi et al., 2002)
have recently been reported in microarray studies examining
genes induced in response to 5-FU treatment. A limitation
of these studies, however, is that in many cases, it remains
to be determined whether the genes identified confer sensitivity or resistance by directly modulating 5-FU action,
or whether they represent surrogate markers of a broader
response phenotype.
nism of 5-FU-induced cell killing (Hwang et al., 2001). Finally, using cDNA microarrays, Maxwell et al. identified
a number of genes altered in expression by 5-FU in vitro
(Maxwell et al., 2003). Examples of the genes upregulated
by 5-FU included those with a role in signal transduction
(K-ras), apoptosis (COP9 homolog, FLIP protein), cell cycle (cdk2, cdc2, cyclin G) and cell structure (gelsolin), as
well as genes such as MAT-8, annexin II and IV and FGF
receptor 2 which are expressed at the cell surface (Maxwell
et al., 2003).
4.3. Determinants of acquired 5-FU resistance
A few investigators have also applied expression profiling to gain insight into the mechanisms of acquired 5-FU
resistance. Using microarrays in which a panel of cell lines
selected for resistance to 5-FU were compared to parental
cells, Wang et al. identified an upregulation of the 5-FU target gene, TS (and the closely linked Yes-1 gene), in 5-FU
resistant cells (Wang et al., 2001). Importantly, using Southern blotting and fluorescence in situ hybridization, the increase in TS mRNA levels was linked to amplification of the
TS gene, which is consistent with in vivo findings (Wang
et al., 2004). Also upregulated in expression in 5-FU resistant cells were CAK1 antigen and alpha catenin, while
checkpoint suppressor 1 and c-rel were downregulated in
expression (Wang et al., 2001).
5. Challenges and recent technological advances
4.2. Genes induced in response to 5-FU
Gene expression profiling studies have also been used to
characterize genes induced in response to 5-FU treatment,
and have identified a number of novel pathways modulated
by this agent. Clarke et al. determined the effect of treatment of late stage colon tumors with 5-FU and mitomycin on
gene expression in vivo, by comparing RNA extracted from
biopsies isolated pre-treatment or following 6 weeks of continuous drug treatment (Clarke et al., 2003). They demonstrated a coordinate downregulation in expression of genes
required for RNA and protein synthesis, including a number
of HnRNP’s and snRNP’s, ribosomal proteins, translation
initiation factors, tRNA synthases, and genes involved in
protein folding, following 5-FU/mitomycin treatment. The
authors interpret these changes as likely reflecting the inhibition of cell proliferation induced by 5-FU/mitomycin. A
number of these genes have been shown to be regulated by
c-myc (Boon et al., 2001; Coller et al., 2000), which was
also downregulated by drug treatment, suggesting a possible
mechanism for their coordinate regulation.
Using SAGE analysis, Hwang et al. demonstrated that
ferredoxin reductase (FR) was induced by 5-FU in vitro in a
p53-dependent manner. Follow up studies demonstrated that
FR may contribute to apoptosis via the excessive generation
of reactive oxygen species, a previously unknown mecha-
A major factor limiting the advancement of gene expression profiling as a tool for the prediction of response to
chemotherapy, is access to fresh frozen patient tumor samples with documented clinical follow up. A strong emphasis
needs to be placed in Surgical and Pathology departments
on the importance of developing tumor banks of fresh frozen
tissue. Without these samples, and follow-up information,
the potential of the currently available technologies cannot
be fully realized.
Once samples are acquired, a further challenge is dealing with the heterogeneity of tumor tissue. Maximizing
the percentage of tumor cells in a given sample is certainly critical for gene expression analyses that involve,
for example, discrimination of tumor cells from normal
tissue, or classification of tumor grade (Gillespie et al.,
2001). To maximize the percentage of tumor cells in a
heterogeneous sample most investigators presently use
“macrodissection” and apply criteria such as the requirement that samples comprise at least 50% tumor cells prior
to inclusion for further analysis (Bertucci et al., 2004). An
alternative approach has been to mathematically subtract
non-tumor-cell derived signatures from the gene expression data “in silico”, including stromal and muscle-cell
specific genes, and genes related to immune function
expressed in lymphocytes (Stuart et al., 2004). The util-
J.M. Mariadason et al. / Drug Resistance Updates 7 (2004) 209–218
ity of this approach was demonstrated by the improved
clustering of colorectal samples of different Dukes staging subsequent to removal of the non-tumor cell profiles
(Frederiksen et al., 2003). While these approaches have
produced encouraging results, improvements in tissue processing may enhance these distinctions even further. To
this end, laser capture-based microdissection of tumor cells
currently represents the gold standard (Crnogorac-Jurcevic
et al., 2002; Ma et al., 2003). However, for analyses that
aim to determine prognosis or response to chemotherapy,
it is unclear whether analysis of pure populations of tumor cells or whether consideration of the tumor cells in
the context of its host environment, may be more informative (Perou et al., 1999). For example, consideration of
the gene expression signatures of non-tumor cells may be
essential for prediction of response to specific therapies.
Illustrating this, St. Croix et al. carefully isolated endothelial cells from colorectal tumors or normal mucosa and
compared gene expression profiles using SAGE analysis.
This study identified a number of genes differentially expressed between tumor and normal endothelium, which
the authors conclude may have significant implications
for the development of anti-angiogenic therapies (St Croix
et al., 2000). Different approaches to sample preparation
are therefore necessary depending upon the question being addressed. With regards to which approach may be
more informative in terms of prognosis and prediction of
response to chemotherapy, a definitive study that compares the use of whole versus microdissected tumor is
An exciting novel methodology termed transcriptional
imaging, which has the potential to circumvent the issue of
tissue heterogeneity, is being developed by Singer and colleagues at the Albert Einstein College of Medicine (Femino
et al., 1998; Wilson et al., 2002). Here, probes directed
against different target genes are generated by labeling with
different fluorochromes. Addition of this cocktail of probes
to a heterogeneous tissue sample results in hybridization of
the individual probes to their respective target mRNA sequence. This hybridization is most evident at the specific site
at which the target gene is transcribed (transcription site),
where transcript levels are most concentrated. Using this approach, transcription sites for specific loci can be identified
and characterized in situ in individual cells (Femino et al.,
1998; Levsky et al., 2002). Levsky et al. demonstrated the
potential of this technology for identifying the profiles of
activation of 11 different loci in individual cells, while Wilson et al. used a variation of this methodology to demonstrate that the HDAC-inhibitor, butyrate, induced repression
of c-myc in colon cancer cells by inducing a pause in c-myc
transcription (Levsky et al., 2002; Wilson et al., 2002). Importantly, and because spectral analysis of each transcription
site is possible, this methodology further enables the identification and quantification of multiple fluorescent wavelengths at a single transcription site. Therefore, by designing
probes with combinations of multiple fluorochromes, it is
possible to simultaneously detect the expression of a large
number of loci using a small number of spectrally distinct
probes. Thus, transcriptional profiling on a single cell basis can be achieved for each of a large number of cells in a
manner that preserves the architecture of the tissue. Theoretically, this approach may be able to identify sub-populations
of cells with important phenotypes (e.g. metastatic potential,
drug resistance) that may represent only a minor percentage
of the overall cell population. Such sensitivity may not be
achievable through microdissection of individual cells coupled with gene expression profiling, because such cells cannot be identified a priori, and therefore the number of individual cells to be microdissected and analyzed would need
to be very large.
While the amounts of RNA required for microarray analysis was previously a limiting factor, significant progress
has been made in RNA amplification protocols that has
significantly reduced the amount of starting RNA required
(1–5 ␮g) for expression profiling. There is also considerable
interest in protocols that enable extraction of good quality RNA from archived formalin fixed paraffin embedded
(FFPE) sections (Coudry et al., 2004; Ding et al., 2004;
Erlander et al., 2003). Should this develop into a truly viable and routine methodology, and the RNA extracted be of
sufficient quality for use in microarray analyses, the opportunities for performing retrospective studies using archived
samples would be significant.
Given that tissue availability remains a major limitation
in establishing the potential of gene expression profiling for
predicting probability of response to chemotherapy, the establishment of systems that improve the ability to integrate
microarray datasets from multiple investigators would be of
significant benefit. Currently, microarrays are performed on
many different platforms, and a variety of strategies for experimental design and data analysis are utilized to process
and interpret the data, which makes integration of different
datasets a major challenge. To deal with this issue, a number of consortia have been formed and several positive systems implemented. Examples include the recommended use
of a universal reference RNA pool for platforms that utilize
a two-color hybridization system such as cDNA microarrays (Novoradovskaya et al., 2004), and compliance with
the MIAME standards, which are designed to enhance interpretation of gene expression data through standardizing the
documentation of each step of the microarray experiment
including methods of array design, tissue preparation, hybridization, image acquisition and analysis, and data normalization (Brazma et al., 2001). Furthermore, the constantly
improving annotation of the human genome, development
of more advanced and user-friendly bioinformatics tools and
resources, and the ability to incorporate progressively more
comprehensive gene sets on array platforms, will undoubtedly contribute towards improving data integration. Nevertheless there is a continued need for individual investigators
to increase compliance with the aforementioned recommendations of experimental design and documentation, to utilize
J.M. Mariadason et al. / Drug Resistance Updates 7 (2004) 209–218
the most widely available microarray platforms when possible, and to deposit databases in public repositories, in order
to maximize the potential of this methodology.
In addition to advances in gene expression profiling methods, such as the gradual movement towards profiling using whole genome arrays, advances have also been made in
other high throughput profiling technologies. For example,
array-based platforms such as BAC arrays are now available
for screening genomic DNA for genomic imbalances such
as insertions and deletions (comparative genome hybridization or CGH arrays), a characteristic event in the development and progression of a number of human cancers (Cai
et al., 2002; Ishkanian et al., 2004). It is also possible to hybridize DNA to the same cDNA or oligonucleotide microarrays used for mRNA profiling, and thus obtain information
on DNA copy number. Using such a strategy, Pollack et al.
conducted an elegant study in which changes in DNA copy
number in a cohort of breast cancers were correlated with
parallel microarray measurements of mRNA levels (Pollack
et al., 2002). This study concluded that approximately 12%
of the variation in gene expression could be attributed to
underlying variation in gene copy number (Pollack et al.,
It is also becoming increasingly possible to use
array-based approaches to screen DNA for epigenetic modification such as methylation, using CpG island arrays
(Gitan et al., 2002; Yan et al., 2000), or for single nucleotide
polymorphisms (SNPs), using high density oligonucleotide
arrays (Dong et al., 2001; Lindblad-Toh et al., 2000). Finally, high throughput proteomic technologies are also
becoming increasing accessible to investigators (Hanash,
2003). These platforms may provide important information not afforded by gene expression profiling. For example, expression profiling may fail to provide information
regarding the functionality or likelihood of induction of
critical drug response-determining genes. For example, the
pro-apoptotic genes p53 and Bax may be inactivated by
mutation (Hollstein et al., 1991; Rampino et al., 1997),
while inducibility of p16 or mdr1 can vary according to
the extent of promoter methylation (Herman et al., 1995;
Shannon and Iacopetta, 2001), neither of which may be
evident by gene expression profiling. It is conceivable
therefore that “systems biology” approaches involving the
integration of gene expression profiling data with complementary genomic, epigenetic, and proteomic data, could
ultimately provide the most accurate method for determining both prognosis and response to chemotherapy of colon
In conclusion, the power of gene expression profiling as
a tool for class discovery and prediction of prognosis has
now been firmly established for a number of cancer types.
A number of in vivo and in vitro experiments have also
demonstrated the potential of this methodology for distinction among more subtle cancer phenotypes, suggesting its
utility as a tool for customizing chemotherapy in colon cancer is an attainable goal in the foreseeable future.
Agrawal, D., Chen, T., Irby, R., Quackenbush, J., Chambers, A.F.,
Szabo, M., Cantor, A., Coppola, D., Yeatman, T.J., 2002. Osteopontin
identified as lead marker of colon cancer progression, using pooled
sample expression profiling. J. Natl. Cancer Inst. 94, 513–521.
Agrawal, D., Chen, T., Irby, R., Quackenbush, J., Chambers, A.F.,
Szabo, M., Cantor, A., Coppola, D., Yeatman, T.J., 2003. Osteopontin
identified as colon cancer tumor progression marker. C. R. Biol. 326,
Ahnen, D.J., Feigl, P., Quan, G., Fenoglio-Preiser, C., Lovato, L.C., Bunn,
Jr., P.A., et al., 1998. Ki-ras mutation and p53 overexpression predict
the clinical behavior of colorectal cancer: a Southwest Oncology Group
study. Cancer Res. 58, 1149–1158.
Allegra, C.J., Paik, S., Colangelo, L.H., Parr, A.L., Kirsch, I., Kim, G.,
Klein, P., Johnston, P.G., Wolmark, N., Wieand, H.S., 2003. Prognostic
value of thymidylate synthase, Ki-67, and p53 in patients with Dukes’
B and C colon cancer: a National Cancer Institute-National Surgical
Adjuvant Breast and Bowel Project collaborative study. J. Clin. Oncol.
21, 241–250.
Alon, U., Barkai, N., Notterman, D.A., Gish, K., Ybarra, S., Mack,
D., Levine, A.J., 1999. Broad patterns of gene expression revealed
by clustering analysis of tumor and normal colon tissues probed by
oligonucleotide arrays. Proc. Natl. Acad. Sci. USA 96, 6745–6750.
Arango, D., Corner, G.A., Wadler, S., Catalano, P.J., Augenlicht, L.H.,
2001. c-myc/p53 interaction determines sensitivity of human colon
carcinoma cells to 5-fluorouracil in vitro and in vivo. Cancer Res. 61,
Arango, D., Mariadason, J.M., Wilson, A.J., Yang, W., Corner,
G.A., Nicholas, C., Aranes, M.J., Augenlicht, L.H., 2003. c-Myc
overexpression sensitises colon cancer cells to camptothecin-induced
apoptosis. Br. J. Cancer. 89, 1757–1765.
Augenlicht, L.H., Taylor, J., Anderson, L., Lipkin, M., 1991. Patterns
of gene expression that characterize the colonic mucosa in patients
at genetic risk for colonic cancer. Proc. Natl. Acad. Sci. USA 88,
Augenlicht, L.H., Wadler, S., Corner, G., Richards, C., Ryan, L., Multani,
A.S., Pathak, S., Benson, A., Haller, D., Heerdt, B.G., 1997. Lowlevel c-myc amplification in human colonic carcinoma cell lines and
tumors: a frequent, p53-independent mutation associated with improved
outcome in a randomized multi-institutional trial. Cancer Res. 57,
Augenlicht, L.H., Wahrman, M.Z., Halsey, H., Anderson, L., Taylor, J.,
Lipkin, M., 1987. Expression of cloned sequences in biopsies of human
colonic tissue and in colonic carcinoma cells induced to differentiate
in vitro. Cancer Res. 47, 6017–6021.
Barratt, P.L., Seymour, M.T., Stenning, S.P., Georgiades, I., Walker, C.,
Birbeck, K., Quirke, P., 2002. DNA markers predicting benefit from
adjuvant fluorouracil in patients with colon cancer: a molecular study.
Lancet 360, 1381–1391.
Beer, D.G., Kardia, S.L., Huang, C.C., Giordano, T.J., Levin, A.M., Misek,
D.E., et al., 2002. Gene-expression profiles predict survival of patients
with lung adenocarcinoma. Nat. Med. 8, 816–824.
Benhattar, J., Cerottini, J.P., Saraga, E., Metthez, G., Givel, J.C., 1996.
p53 mutations as a possible predictor of response to chemotherapy in
metastatic colorectal carcinomas. Int. J. Cancer. 69, 190–192.
Berglund, A., Edler, D., Molin, D., Nordlinder, H., Graf, W., Glimelius, B.,
2002. Thymidylate synthase and p53 expression in primary tumor do
not predict chemotherapy outcome in metastatic colorectal carcinoma.
Anticancer Res. 22, 3653–3659.
Bertucci, F., Salas, S., Eysteries, S., Nasser, V., Finetti, P., Ginestier,
C., et al., 2004. Gene expression profiling of colon cancer by DNA
microarrays and correlation with histoclinical parameters. Oncogene
23, 1377–1391.
Bhattacharjee, A., Richards, W.G., Staunton, J., Li, C., Monti, S., Vasa,
P., et al., 2001. Classification of human lung carcinomas by mRNA
J.M. Mariadason et al. / Drug Resistance Updates 7 (2004) 209–218
expression profiling reveals distinct adenocarcinoma subclasses. Proc.
Natl. Acad. Sci. USA 98, 13790–13795.
Birkenkamp-Demtroder, K., Christensen, L.L., Olesen, S.H., Frederiksen,
C.M., Laiho, P., Aaltonen, L.A., et al., 2002. Gene expression in
colorectal cancer. Cancer Res. 62, 4352–4363.
Boon, K., Caron, H.N., van Asperen, R., Valentijn, L., Hermus, M.C.,
van Sluis, P., et al., 2001. N-myc enhances the expression of a large
set of genes functioning in ribosome biogenesis and protein synthesis.
EMBO J. 20, 1383–1393.
Boussioutas, A., Li, H., Liu, J., Waring, P., Lade, S., Holloway, A.J.,
Taupin, D., Gorringe, K., Haviv, I., Desmond, P.V., Bowtell, D.D.,
2003. Distinctive patterns of gene expression in premalignant gastric
mucosa and gastric cancer. Cancer Res. 63, 2569–2577.
Brazma, A., Hingamp, P., Quackenbush, J., Sherlock, G., Spellman, P.,
Stoeckert, C., et al., 2001. Minimum information about a microarray
experiment (MIAME)-toward standards for microarray data. Nat.
Genet. 29, 365–371.
Bullinger, L., Dohner, K., Bair, E., Frohling, S., Schlenk, R.F., Tibshirani,
R., Dohner, H., Pollack, J.R., 2004. Use of gene-expression profiling
to identify prognostic subclasses in adult acute myeloid leukemia. N.
Engl. J. Med. 350, 1605–1616.
Cai, W.W., Mao, J.H., Chow, C.W., Damani, S., Balmain, A., Bradley, A.,
2002. Genome-wide detection of chromosomal imbalances in tumors
using BAC microarrays. Nat. Biotechnol. 20, 393–396.
Clarke, P.A., George, M.L., Easdale, S., Cunningham, D., Swift, R.I.,
Hill, M.E., Tait, D.M., Workman, P., 2003. Molecular pharmacology of
cancer therapy in human colorectal cancer by gene expression profiling.
Cancer Res. 63, 6855–6863.
Coller, H.A., Grandori, C., Tamayo, P., Colbert, T., Lander, E.S.,
Eisenman, R.N., Golub, T.R., 2000. Expression analysis with
oligonucleotide microarrays reveals that MYC regulates genes involved
in growth, cell cycle, signaling, and adhesion. Proc. Natl. Acad. Sci.
USA 97, 3260–3265.
Compton, C.C., 2002. Surgical pathology of colorectal cancer. Humana
Press, Totowa, New Jersey.
Conti, J.A., Kemeny, N.E., Saltz, L.B., Huang, Y., Tong, W.P., Chou,
T.C., Sun, M., Pulliam, S., Gonzalez, C., 1996. Irinotecan is an active
agent in untreated patients with metastatic colorectal cancer. J. Clin.
Oncol. 14, 709–715.
Coudry, R., Meireles, S., Stoyanova, R., Carpino, A., Cooper, H.A.,
Engstrom, P.F., Clapper, M.L., 2004. Gene expression profiling
of formalin-fixed, paraffin-embedded archival tissue following laser
capture microdissection. Proc. Am. Assoc. Cancer Res. 45, LB-79.
Crnogorac-Jurcevic, T., Efthimiou, E., Nielsen, T., Loader, J., Terris,
B., Stamp, G., et al., 2002. Expression profiling of microdissected
pancreatic adenocarcinomas. Oncogene 21, 4587–4594.
Cvitkovic, E., Bekradda, M., 1999. Oxaliplatin: a new therapeutic option
in colorectal cancer. Semin. Oncol. 26, 647–662.
Dan, S., Tsunoda, T., Kitahara, O., Yanagawa, R., Zembutsu, H.,
Katagiri, T., Yamazaki, K., Nakamura, Y., Yamori, T., 2002. An
integrated database of chemosensitivity to 55 anticancer drugs and
gene expression profiles of 39 human cancer cell lines. Cancer Res.
62, 1139–1147.
Dannenberg, P.V., 2002. Molecular markers of chemotherapy resistance
in colorectal cancer. Humana Press, Totowa, New Jersey.
de Gramont, A., Vignoud, J., Tournigand, C., Louvet, C., Andre, T.,
Varette, C., et al., 1997. Oxaliplatin with high-dose leucovorin and 5fluorouracil 48-h continuous infusion in pretreated metastatic colorectal
cancer. Eur. J. Cancer 33, 214–219.
Dhanasekaran, S.M., Barrette, T.R., Ghosh, D., Shah, R., Varambally,
S., Kurachi, K., Pienta, K.J., Rubin, M.A., Chinnaiyan, A.M., 2001.
Delineation of prognostic biomarkers in prostate cancer. Nature 412,
Ding, L., Chu, S., Kunitake, S., MacNair, J., Cheung, E., Ghosh, M., et
al., 2004. Gene expression profiling of microdissected human breast
cancer cells from formalin-fixed paraffin-embedded sections using
microarrays. Proc. Am. Assoc. Cancer Res. 45, LB-142. Abstract.
Dong, S., Wang, E., Hsie, L., Cao, Y., Chen, X., Gingeras, T.R., 2001.
Flexible use of high-density oligonucleotide arrays for single-nucleotide
polymorphism discovery and validation. Genome Res. 11, 1418–1424.
Elsaleh, H., Iacopetta, B., 2001a. Microsatellite instability is a predictive
marker for survival benefit from adjuvant chemotherapy in a
population-based series of stage III colorectal carcinoma. Clin.
Colorectal Cancer 1, 104–109.
Elsaleh, H., Joseph, D., Grieu, F., Zeps, N., Spry, N., Iacopetta, B., 2000.
Association of tumour site and sex with survival benefit from adjuvant
chemotherapy in colorectal cancer. Lancet 355, 1745–1750.
Elsaleh, H., Powell, B., McCaul, K., Grieu, F., Grant, R., Joseph, D.,
Iacopetta, B., 2001b. P53 alteration and microsatellite instability have
predictive value for survival benefit from chemotherapy in stage III
colorectal carcinoma. Clin. Cancer Res. 7, 1343–1349.
Erlander, M.G., Ellis, M.J., Olson, J., Wittliff, J.L., Smith, B., Younger,
J., Sgroi, D.C., 2003. Global gene expression profiling from formalinfixed paraffin-embedded (FFPE) tissues Implications for retrospective
and prospective clinical studies. Proc. Am. Soc. Clin. Oncol. 22, 125.
Abstract 498.
Femino, A.M., Fay, F.S., Fogarty, K., Singer, R.H., 1998. Visualization
of single RNA transcripts in situ. Science 280, 585–590.
Ferrara, N., Hillan, K.J., Gerber, H.P., Novotny, W., 2004. Case history:
Discovery and development of bevacizumab, an anti-VEGF antibody
for treating cancer. Nat. Rev. Drug Discov. 3, 391–400.
Findlay, M.P., Cunningham, D., Morgan, G., Clinton, S., Hardcastle,
A., Aherne, G.W., 1997. Lack of correlation between thymidylate
synthase levels in primary colorectal tumours and subsequent response
to chemotherapy. Br. J. Cancer 75, 903–909.
Fodde, R., Smits, R., Clevers, H., 2001. APC, signal transduction and
genetic instability in colorectal cancer. Nat. Rev. Cancer 1, 55–67.
Frederiksen, C.M., Knudsen, S., Laurberg, S., Orntoft, T.F., 2003.
Classification of Dukes’ B and C colorectal cancers using expression
arrays. J. Cancer Res. Clin. Oncol. 129, 263–271.
Gillespie, J.W., Ahram, M., Best, C.J., Swalwell, J.I., Krizman, D.B.,
Petricoin, E.F., Liotta, L.A., Emmert-Buck, M.R., 2001. The role of
tissue microdissection in cancer research. Cancer J. 7, 32–39.
Gitan, R.S., Shi, H., Chen, C.M., Yan, P.S., Huang, T.H., 2002.
Methylation-specific oligonucleotide microarray: a new potential for
high-throughput methylation analysis. Genome Res. 12, 158–164.
Goh, H.S., Yao, J., Smith, D.R., 1995. p53 point mutation and survival
in colorectal cancer patients. Cancer Res. 55, 5217–5221.
Golub, T.R., Slonim, D.K., Tamayo, P., Huard, C., Gaasenbeek, M.,
Mesirov, J.P., et al., 1999. Molecular classification of cancer: class
discovery and class prediction by gene expression monitoring. Science
286, 531–537.
Grem, J.L., Danenberg, K.D., Behan, K., Parr, A., Young, L., Danenberg,
P.V., Nguyen, D., Drake, J., Monks, A., Allegra, C.J., 2001. Thymidine
kinase, thymidylate synthase, and dihydropyrimidine dehydrogenase
profiles of cell lines of the National Cancer Institute’s Anticancer Drug
Screen. Clin. Cancer Res. 7, 999–1009.
Hanash, S., 2003. Disease proteomics. Nature 422, 226–232.
Herman, J.G., Merlo, A., Mao, L., Lapidus, R.G., Issa, J.P., Davidson,
N.E., Sidransky, D., Baylin, S.B., 1995. Inactivation of the
CDKN2/p16/MTS1 gene is frequently associated with aberrant DNA
methylation in all common human cancers. Cancer Res. 55, 4525–
Hollstein, M., Sidransky, D., Vogelstein, B., Harris, C.C., 1991. p53
mutations in human cancers. Science 253, 49–53.
Hwang, P.M., Bunz, F., Yu, J., Rago, C., Chan, T.A., Murphy, M.P., et
al., 2001. Ferredoxin reductase affects p53-dependent, 5-fluorouracilinduced apoptosis in colorectal cancer cells. Nat. Med. 7, 1111–1117.
Iacopetta, B., Elsaleh, H., Zeps, N., 2003. Microsatellite instability in
colon cancer. N. Engl. J. Med. 349, 1774–1776, author reply 1774–
Ishkanian, A.S., Malloff, C.A., Watson, S.K., DeLeeuw, R.J., Chi, B., Coe,
B.P., et al., 2004. A tiling resolution DNA microarray with complete
coverage of the human genome. Nat. Genet. 36, 299–303.
J.M. Mariadason et al. / Drug Resistance Updates 7 (2004) 209–218
Johnston, P.G., Benson, 3rd, A.B., Catalano, P., Rao, M.S., O’Dwyer, P.J.,
Allegra, C.J., 2003. Thymidylate synthase protein expression in primary
colorectal cancer: lack of correlation with outcome and response to
fluorouracil in metastatic disease sites. J. Clin. Oncol. 21, 815–819.
Johnston, P.G., Drake, J.C., Trepel, J., Allegra, C.J., 1992. Immunological
quantitation of thymidylate synthase using the monoclonal antibody TS
106 in 5-fluorouracil-sensitive and -resistant human cancer cell lines.
Cancer Res. 52, 4306–4312.
Kihara, C., Tsunoda, T., Tanaka, T., Yamana, H., Furukawa, Y., Ono,
K., et al., 2001. Prediction of sensitivity of esophageal tumors
to adjuvant chemotherapy by cDNA microarray analysis of geneexpression profiles. Cancer Res. 61, 6474–6479.
Kinsella, A.R., Smith, D., Pickard, M., 1997. Resistance to
chemotherapeutic antimetabolites: a function of salvage pathway
involvement and cellular response to DNA damage. Br. J. Cancer 75,
Leppa, S., Sistonen, L., 1997. Heat shock response—pathophysiological
implications. Ann. Med. 29, 73–78.
Levsky, J.M., Shenoy, S.M., Pezo, R.C., Singer, R.H., 2002. Single-cell
gene expression profiling. Science 297, 836–840.
Lindblad-Toh, K., Winchester, E., Daly, M.J., Wang, D.G., Hirschhorn,
J.N., Laviolette, J.P., et al., 2000. Large-scale discovery and genotyping
of single-nucleotide polymorphisms in the mouse. Nat. Genet. 24,
Longley, D.B., Harkin, D.P., Johnston, P.G., 2003. 5-fluorouracil:
mechanisms of action and clinical strategies. Nat. Rev. Cancer 3,
Ma, X.J., Salunga, R., Tuggle, J.T., Gaudet, J., Enright, E., McQuary, P., et
al., 2003. Gene expression profiles of human breast cancer progression.
Proc. Natl. Acad. Sci. USA 100, 5974–5979.
Machover, D., Diaz-Rubio, E., de Gramont, A., Schilf, A., Gastiaburu, J.J.,
Brienza, S., et al., 1996. Two consecutive phase II studies of oxaliplatin
(L-OHP) for treatment of patients with advanced colorectal carcinoma
who were resistant to previous treatment with fluoropyrimidines. Ann.
Oncol. 7, 95–98.
Mariadason, J.M., Arango, D., Shi, Q., Wilson, A.J., Corner, G.A.,
Nicholas, C., et al., 2003. Gene expression profiling-based prediction of
response of colon carcinoma cells to 5-fluorouracil and camptothecin.
Cancer Res. 63, 8791–8812.
Maxwell, P.J., Longley, D.B., Latif, T., Boyer, J., Allen, W., Lynch, M., et
al., 2003. Identification of 5-fluorouracil-inducible target genes using
cDNA microarray profiling. Cancer Res. 63, 4602–4606.
Metzger, R., Danenberg, K., Leichman, C.G., Salonga, D., Schwartz,
E.L., Wadler, S., et al., 1998. High basal level gene expression
of thymidine phosphorylase (platelet-derived endothelial cell growth
factor) in colorectal tumors is associated with nonresponse to 5fluorouracil. Clin. Cancer Res. 4, 2371–2376.
Mirjolet, J.F., Barberi-Heyob, M., Merlin, J.L., Marchal, S., Etienne,
M.C., Milano, G., Bey, P., 1998. Thymidylate synthase expression and
activity: relation to S-phase parameters and 5-fluorouracil sensitivity.
Br. J. Cancer 78, 62–68.
Moertel, C.G., Fleming, T.R., Macdonald, J.S., Haller, D.G., Laurie, J.A.,
Tangen, C.M., et al., 1995a. Fluorouracil plus levamisole as effective
adjuvant therapy after resection of stage III colon carcinoma: a final
report. Ann. Intern. Med. 122, 321–326.
Moertel, C.G., Fleming, T.R., Macdonald, J.S., Haller, D.G., Laurie, J.A.,
Tangen, C.M., et al., 1995b. Intergroup study of fluorouracil plus
levamisole as adjuvant therapy for stage II/Dukes’ B2 colon cancer. J.
Clin. Oncol. 13, 2936–2943.
Mori, Y., Selaru, F.M., Sato, F., Yin, J., Simms, L.A., Xu, Y., et al., 2003.
The impact of microsatellite instability on the molecular phenotype of
colorectal tumors. Cancer Res. 63, 4577–4582.
Mori, Y., Yin, J., Sato, F., Sterian, A., Simms, L.A., Selaru, F.M.,
et al., 2004. Identification of genes uniquely involved in frequent
microsatellite instability colon carcinogenesis by expression profiling
combined with epigenetic scanning. Cancer Res. 64, 2434–2438.
Morin, P.J., 2003. Drug resistance and the microenvironment: nature and
nurture. Drug Resist. Update 6, 169–172.
Nita, M.E., Nagawa, H., Tominaga, O., Tsuno, N., Fujii, S., Sasaki, S.,
Fu, C.G., Takenoue, T., Tsuruo, T., Muto, T., 1998. 5-Fluorouracil
induces apoptosis in human colon cancer cell lines with modulation
of Bcl-2 family proteins. Br. J. Cancer 78, 986–992.
Notterman, D.A., Alon, U., Sierk, A.J., Levine, A.J., 2001. Transcriptional
gene expression profiles of colorectal adenoma, adenocarcinoma, and
normal tissue examined by oligonucleotide arrays. Cancer Res. 61,
Novoradovskaya, N., Whitfield, M.L., Basehore, L.S., Novoradovsky, A.,
Pesich, R., Usary, J., et al., 2004. Universal Reference RNA as a
standard for microarray experiments. BMC Genomics 5, 20.
Perou, C.M., Jeffrey, S.S., van de Rijn, M., Rees, C.A., Eisen, M.B.,
Ross, D.T., et al., 1999. Distinctive gene expression patterns in human
mammary epithelial cells and breast cancers. Proc. Natl. Acad. Sci.
USA 96, 9212–9217.
Pollack, J.R., Sorlie, T., Perou, C.M., Rees, C.A., Jeffrey, S.S., Lonning,
P.E., Tibshirani, R., Botstein, D., Borresen-Dale, A.L., Brown, P.O.,
2002. Microarray analysis reveals a major direct role of DNA copy
number alteration in the transcriptional program of human breast
tumors. Proc. Natl. Acad. Sci. USA 99, 12963–12968.
Pomeroy, S.L., Tamayo, P., Gaasenbeek, M., Sturla, L.M., Angelo, M.,
McLaughlin, M.E., et al., 2002. Prediction of central nervous system
embryonal tumour outcome based on gene expression. Nature 415,
Rampino, N., Yamamoto, H., Ionov, Y., Li, Y., Sawai, H., Reed, J.C.,
Perucho, M., 1997. Somatic frameshift mutations in the BAX gene in
colon cancers of the microsatellite mutator phenotype. Science 275,
Ribic, C.M., Sargent, D.J., Moore, M.J., Thibodeau, S.N., French, A.J.,
Goldberg, R.M., et al., 2003. Tumor microsatellite-instability status as
a predictor of benefit from fluorouracil-based adjuvant chemotherapy
for colon cancer. N. Engl. J. Med. 349, 247–257.
Saha, S., Bardelli, A., Buckhaults, P., Velculescu, V.E., Rago, C., St Croix,
B., et al., 2001. A phosphatase associated with metastasis of colorectal
cancer. Science 294, 1343–1346.
Saito, H., Tsujitani, S., Oka, S., Kondo, A., Ikeguchi, M., Maeta,
M., Kaibara, N., 1999. The expression of thymidine phosphorylase
correlates with angiogenesis and the efficacy of chemotherapy using
fluorouracil derivatives in advanced gastric carcinoma. Br. J. Cancer
81, 484–489.
Salonga, D., Danenberg, K.D., Johnson, M., Metzger, R., Groshen, S.,
Tsao-Wei, D.D., et al., 2000. Colorectal tumors responding to 5fluorouracil have low gene expression levels of dihydropyrimidine
dehydrogenase, thymidylate synthase, and thymidine phosphorylase.
Clin. Cancer Res. 6, 1322–1327.
Sarris, M., Lee, C.S., 2001. nm23 protein expression in colorectal
carcinoma metastasis in regional lymph nodes and the liver. Eur. J.
Surg. Oncol. 27, 170–174.
Sax, J.K., Fei, P., Murphy, M.E., Bernhard, E., Korsmeyer, S.J., El-Deiry,
W.S., 2002. BID regulation by p53 contributes to chemosensitivity.
Nat. Cell Biol. 4, 842–849.
Schaner, M.E., Ross, D.T., Ciaravino, G., Sorlie, T., Troyanskaya, O.,
Diehn, M., et al., 2003. Gene expression patterns in ovarian carcinomas.
Mol. Biol. Cell. 14, 4376–4386.
Scherf, U., Ross, D.T., Waltham, M., Smith, L.H., Lee, J.K., Tanabe, L., et
al., 2000. A gene expression database for the molecular pharmacology
of cancer. Nat. Genet. 24, 236–244.
Schwartz, E.L., Baptiste, N., Wadler, S., Makower, D., 1995. Thymidine
phosphorylase mediates the sensitivity of human colon carcinoma cells
to 5-fluorouracil. J. Biol. Chem. 270, 19073–19077.
Selaru, F.M., Xu, Y., Yin, J., Zou, T., Liu, T.C., Mori, Y., et al., 2002.
Artificial neural networks distinguish among subtypes of neoplastic
colorectal lesions. Gastroenterology 122, 606–613.
Seoane, J., Le, H.V., Massague, J., 2002. Myc suppression of the p21(Cip1)
Cdk inhibitor influences the outcome of the p53 response to DNA
damage. Nature 419, 729–734.
J.M. Mariadason et al. / Drug Resistance Updates 7 (2004) 209–218
Shannon, B.A., Iacopetta, B.J., 2001. Methylation of the hMLH1,
p16, and MDR1 genes in colorectal carcinoma: associations with
clinicopathological features. Cancer Lett. 167, 91–97.
Sorlie, T., Perou, C.M., Tibshirani, R., Aas, T., Geisler, S., Johnsen, H.,
et al., 2001. Gene expression patterns of breast carcinomas distinguish
tumor subclasses with clinical implications. Proc. Natl. Acad. Sci. USA
98, 10869–10874.
St Croix, B., Rago, C., Velculescu, V., Traverso, G., Romans, K.E.,
Montgomery, E., Lal, A., Riggins, G.J., Lengauer, C., Vogelstein, B.,
Kinzler, K.W., 2000. Genes expressed in human tumor endothelium.
Science 289, 1197–1202.
Stuart, R.O., Wachsman, W., Berry, C.C., Wang-Rodriguez, J., Wasserman,
L., Klacansky, I., et al., 2004. In silico dissection of cell-type-associated
patterns of gene expression in prostate cancer. Proc. Natl. Acad. Sci.
USA 101, 615–620.
Takahashi, Y., Nagata, T., Ishii, Y., Ikarashi, M., Ishikawa, K., Asai, S.,
2002. Up-regulation of vitamin D3 up-regulated protein 1 gene in
response to 5-fluorouracil in colon carcinoma SW620. Oncol. Rep. 9,
Tay, S.T., Leong, S.H., Yu, K., Aggarwal, A., Tan, S.Y., Lee, C.H., et al.,
2003. A combined comparative genomic hybridization and expression
microarray analysis of gastric cancer reveals novel molecular subtypes.
Cancer Res. 63, 3309–3316.
Van de Vijver, M.J., He, Y.D., van’t Veer, L.J., Dai, H., Hart, A.A.,
Voskuil, D.W., et al., 2002. A gene-expression signature as a predictor
of survival in breast cancer. N. Engl. J. Med. 347, 1999–2009.
Van ’t Veer, L.J., Dai, H., van de Vijver, M.J., He, Y.D., Hart, A.A., Mao,
M., et al., 2002. Gene expression profiling predicts clinical outcome
of breast cancer. Nature 415, 530–536.
Vasilevskaya, I., O’Dwyer, P.J., 2003. Role of Jun and Jun kinase in
resistance of cancer cells to therapy, 6: 147–156.
Violette, S., Poulain, L., Dussaulx, E., Pepin, D., Faussat, A.M., Chambaz,
J., et al., 2002. Resistance of colon cancer cells to long-term 5fluorouracil exposure is correlated to the relative level of Bcl-2 and
Bcl-X(L) in addition to Bax and p53 status. Int. J. Cancer 98, 498–
Wang, T.L., Diaz, Jr., L.A., Romans, K., Bardelli, A., Saha, S.,
Galizia, G., et al., 2004. Digital karyotyping identifies thymidylate
synthase amplification as a mechanism of resistance to 5-fluorouracil
in metastatic colorectal cancer patients. Proc. Natl. Acad. Sci. USA
101, 3089–3094.
Wang, W., Marsh, S., Cassidy, J., McLeod, H.L., 2001. Pharmacogenomic
dissection of resistance to thymidylate synthase inhibitors. Cancer Res.
61, 5505–5510.
Wilson, A.J., Velcich, A., Arango, D., Kurland, A.R., Shenoy, S.M.,
Pezo, R.C., Levsky, J.M., Singer, R.H., Augenlicht, L.H., 2002. Novel
detection and differential utilization of a c-myc transcriptional block
in colon cancer chemoprevention. Cancer Res. 62, 6006–6010.
Yan, P.S., Perry, M.R., Laux, D.E., Asare, A.L., Caldwell, C.W., Huang,
T.H., 2000. CpG island arrays: an application toward deciphering
epigenetic signatures of breast cancer. Clin. Cancer Res. 6, 1432–1438.
Zembutsu, H., Ohnishi, Y., Tsunoda, T., Furukawa, Y., Katagiri, T.,
Ueyama, Y., et al., 2002. Genome-wide cDNA microarray screening to
correlate gene expression profiles with sensitivity of 85 human cancer
xenografts to anticancer drugs. Cancer Res. 62, 518–527.
Zhang, L., Yu, J., Park, B.H., Kinzler, K.W., Vogelstein, B., 2000. Role
of BAX in the apoptotic response to anticancer agents. Science 290,
Zou, T.T., Selaru, F.M., Xu, Y., Shustova, V., Yin, J., Mori, Y., et al., 2002.
Application of cDNA microarrays to generate a molecular taxonomy
capable of distinguishing between colon cancer and normal colon.
Oncogene 21, 4855–4862.