| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
1 Department of Pathology and Laboratory Medicine, 2 Curriculum in Genetics and Molecular Biology, 3 Department of Genetics, 4 Lineberger Comprehensive Cancer Center, and 5 Center for Environmental Health and Susceptibility, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina; 6 Departments of Genetics, The Norwegian Radium Hospital, Montebello, Oslo, Norway; 7 Department of Medical and Molecular Genetics, Indiana University School of Medicine, Indianapolis, Indiana; 8 Department of Medicine, Section of Oncology, Haukeland University Hospital, Bergen, Norway; and 9 Department of Cell Biology, University of Texas Southwestern Medical Center, Dallas, Texas
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
amplification or deletion (3)
, but the mechanisms of chemoresistance are still poorly understood. To better understand variations in clinical responses to treatment, recent studies have used gene expression patterns to identify major biological subtypes of breast cancer. These studies identified a previously unrecognized tumor subtype with characteristics of breast basal epithelium (4, 5, 6, 7, 8)
; Basal-like tumors are estrogen receptor (ER)-negative, do not overexpress HER2, and they have a poor prognosis compared with tumors derived from luminal epithelium (5
, 7)
.
Basal and luminal breast tumors are often treated with the same chemotherapeutic agents, but little is known about how each cell type responds to these drugs. To improve our understanding of how basal and luminal epithelium differ in their responses to chemotherapy, we selected two representative cell lines from each of these breast epithelial cell types to study; two human mammary epithelial (HME) cell lines immortalized by the overexpression of the catalytic subunit of telomerase (hTERT) represent the basal subtype, and two breast tumor-derived cell lines (MCF-7 and ZR-75–1) represent the luminal subtype (9)
. All four cell lines express wild-type p53 protein. True to their corresponding tumor subtypes, the HME lines are ER-negative and the luminal cancer cell lines are ER-positive. We treated all four cell lines with DOX and 5FU and performed expression profiling to identify patterns of response.
Transcriptional profiling is a powerful approach for investigating cellular responses to drugs. This approach has led to greater understanding of pathway inhibition and off-target drug effects (10) , the response of yeast to genotoxic agents and environmental stresses (11) , and the effects of different kinds of DNA-damaging agents in human cells (12) . Our analyses showed that the transcriptional responses of the basal and luminal cell lines to chemotherapeutics are quite distinct. We also correlated our in vitro data with in vivo data on breast tumors sampled before and after treatment with DOX or 5FU/mitomycin C (4 , 5 , 7) , and we identified commonalities. Taken together, these in vitro and in vivo data sets illustrate that cell type is an important determinant of response to commonly used chemotherapeutics.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Cytotoxicity Assay.
A mitochondrial dye conversion assay (Cell Titer 96, Promega) was used to quantitate cell line responses to chemotherapeutics. Five thousand cells were seeded per well of a 96-well plate. The cells were allowed to adhere overnight and then the media was replaced with fresh media containing a range of drug doses (DOX, 0–10 µM; 5FU, 0–10 mM). After 36 h of drug treatment, 15 µl of tetrazolium dye solution were added, and culture was incubated at 37°C for 1 h before adding Cell Titer 96 Stop Solution. Dye conversion products were allowed to solubilize in a humidified chamber overnight, and absorbance was measured at 570 nm (minus background absorbance at 650 nm).
Estimating the IC50.
The IC50 for 36 h of treatment for each drug in each cell line was estimated using nonlinear regression (SAS Statistical Software, Cary, NC) and the following relationship:
 |
Collection of mRNA for Microarray Experiments.
Cell lines were grown in 150-mm dishes to 70–80% confluence and then were treated for 3, 12, 24, or 36 h with DOX (doxorubicin hydrochloride) or 5FU (Sigma) at the IC50 concentration. Cells were harvested by scraping and mRNA was isolated using a Micro-FastTrack kit (Invitrogen). To generate feeding control (sham) mRNA samples for each cell line, cells were treated with medium only, in parallel with drug-treated samples. Individual harvests of treated or sham mRNA were not pooled before microarray analysis. However, a reference mRNA sample was generated for each of the four cell lines by harvesting untreated mRNA from each cell line at 80% confluence and then pooling four harvests together (i.e., four MCF-7 harvests were pooled and served as the reference mRNA for all MCF-7 experiments), using each cell line as its own reference controlled for baseline differences between the cell lines.
Microarray Experiments.
Syntheses of labeled cDNA were performed as described previously (4)
, with reference cDNAs labeled with Cy3-dUTP and treated and sham cDNAs labeled with Cy5-dUTP. Each cDNA sample mix was hybridized overnight to an oligonucleotide microarray created in the University of North Carolina at Chapel Hill Genomics Core Facility (http://genomicscore.unc.edu/). These microarrays were created by spotting the Compugen Human oligomers library representing 18,861 human genes (http://www.labonweb.com/chips/libraries.html) onto coated microarray slides (Corning no. 40016). All of the microarray raw data tables are available at the University of North Carolina Microarray Database (https://genome.unc.edu/) at the supporting website for this article,10
and have been deposited in the Gene Expression Omnibus under the accession number of GSE763 (submitter C. Perou). The direction of gene expression change was verified by real-time reverse transcription PCR for a subset of samples using commercially available primers (Applied Biosystems) for p21waf, ferredoxin reductase, prostate differentiation factor, and inhibitor of DNA binding 3. To normalize the target sample variation, we used the average of three control genes: splicing factor 3A subunit 1 (SF3A1), pumilio homolog 1 (PUM1), and ß-actin. SF3A1 and PUM1 were selected as control genes because they had the lowest variation across the tumor data set presented in Perou et al. (4)
. Sham-adjusted real-time-PCR values were regressed on the average of sham-adjusted log2(red/green ratio) array values for each gene; the regression yielded a positive slope of 4.3, Pearson r = 0.75 (data not shown).
SAM Using Cell Line Data.
Genes that were significantly induced or repressed were identified using the significance analysis of microarrays (SAM) package Add-In for Microsoft Excel (15)
. Before conducting SAM, genes were excluded that did not have a mean signal intensity greater than twice the median background value for both the red and green channels in at least 70% of the experiments. For genes that passed these filtering criteria, the log-base-2 of median red intensity over median green intensity was calculated.
The gene expression changes in the 3 h time points were very modest (data not shown); therefore, this time point was excluded from all analyses. To identify genes whose steady-state expression was altered, we combined the 12-, 24-, and 36-h time points for each cell line and treatment group into a single class. This eliminated artifacts caused by random temporal variation in steady-state RNA levels. Two or three replicate arrays were used for each treatment condition for each cell line.
To identify a general stress response pattern, DOX- and 5FU-treated experiments were combined into a single class and compared against sham experiments for each cell line (i.e., MCF-7 DOX- and 5FU-treated versus MCF-7 sham). Missing data were imputed using SAM with 100 permutations and 10 k-nearest neighbors. A two-class unpaired SAM analysis was conducted on the imputed data set. The SAM 
values were adjusted to obtain the largest gene list that gave a false discovery rate of less than 5%.
SAM Using Breast Tumor Data.
All of the tumor data were published previously (4
, 5
, 7)
except for data from five new tumors samples collected after chemotherapy, which are now publicly available at the Stanford Microarray Database (http://genome-www5.stanford.edu/). This breast tumor dataset encompassed two different cohorts of breast cancer patients, one of which received neoadjuvant DOX and a second of which received neoadjuvant 5FU and mitomycin C (16
, 17) ; in both cohorts, we obtained samples of the tumors before therapy and at the time of surgical resection (after therapy sample). All before- and after-samples were labeled with Cy5-dUTP, mixed with Cy3-dUTP-labeled Stanford common reference sample, and hybridized to cDNA microarrays produced at Stanford University (4)
. The gene expression patterns of all of the before-samples were compared with the gene expression patterns of all of the after-samples using a two-class, unpaired SAM analysis. A total of 81 before- and 50 after-samples were assessed, representing all of the tumor subtypes identified in Sørlie et al. (5)
. Consistent with the in vitro data analyses, SAM values were adjusted to obtain the largest gene list that gave a false discovery rate of less than 5%.
To study the genes differentially regulated in basal or luminal tumor subtypes separately, we also classified each tumor into one of two groups using the intrinsic list of Sørlie et al. (7) : one group contained those tumors that represented the luminal epithelium-derived tumors (both Luminal A and B for a total of 51 before- and 30 after-samples) and a second group representing the basal subtype (for a total of 11 before- and 10 after-samples). Each of these two groups was then analyzed using a two-class unpaired SAM analysis; gene expression patterns of before-samples were compared with gene expression patterns of after-samples, and false discovery rates were estimated.
Hierarchical Clustering of Gene Expression Responses.
Average linkage hierarchical cluster analysis using Pearson correlation was conducted using the program Cluster, and the data were visualized in Treeview (18
, 19)
. To visualize the gene expression patterns for the luminal cell lines, the data from the union of the genes identified by SAM for MCF-7 and ZR-75–1 were identified, combined into a nonredundant list, and clustered. These clusters illustrate the fold change relative to control levels for each gene. Following the same procedure, data from the union of the gene sets identified for ME16C and HME-CC were extracted, combined into a nonredundant list, and clustered. Cluster analysis was also performed using the top 100 genes identified by SAM for distinguishing between luminal and basal cell lines responses to DOX-treatment and 5FU-treatment. For all of the clusters, genes were excluded that did not have a mean intensity greater than twice the median background for both the red and green channel in at least 80% of the experiments. For the breast tumor data, similar gene filtering, SAM, and clustering analyses were performed.
Western Blot Analysis.
Cells were treated for 36 h with DOX or 5FU at the 36-h IC50 concentration. Cells were rinsed with PBS and then harvested with M-PER Mammalian Protein Extraction reagent (Pierce) containing Halt Protease Inhibitor and 5 mM EDTA (Pierce). Protein concentrations were determined using Micro BCA Protein Assay Reagent kit (Pierce). Lysates were combined with 2x Laemmli Sample Buffer (Bio-Rad) containing ß-mercaptoethanol and were boiled for 5 min. Forty µg of protein were electrophoresed on a 4–20% Tris-HCl Criterion precast gel (Bio-Rad) and transferred to a Hybond-P membrane (Amersham Biosciences) by electroblotting. The blots were probed with antibodies against p21waf1 (Neomarkers; Ab-11) and ß-actin (Abcam, AC-15). Blots were washed three times with Tris-buffered saline supplemented with 0.1% TWEEN and then were probed with antimouse IgG horseradish peroxidase-linked whole antibody from sheep (Amersham). The blots were rewashed, and detection was by enhanced chemiluminescence (SuperSignal West Pico Chemiluminescent substrate; Pierce).
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
.
|
Chemotherapeutic-Induced Gene Expression Patterns in Luminal Cell Lines.
Differences between basal and luminal cell lines responding to treatment were immediately evident given the absolute number of genes whose expression was altered when treated experiments were compared with sham experiments (Table 2)
. In each luminal cell line, 
10-fold more genes were altered in response to drug. To visualize these expression changes, we combined the SAM-supervised lists for the two luminal cell lines and performed a hierarchical clustering analysis (Fig. 1
and Supplemental Fig. 1 for the complete cluster diagram with all gene names). Each cell line had a unique expression response to chemotherapy that was distinct enough to cause the two treated luminal lines to cluster into different dendrogram branches (Fig. 1B)
. Clusters of genes that distinguish between MCF-7 and ZR-75–1 responses can be found in Supplemental Fig. 1.
|
|
. These genes had slightly increased expression in the sham experiments because of feeding but had greatly diminished expression during drug treatment. This cluster included well-characterized cell cycle regulators (20)
such as cyclin A2, cyclin B1, cell division cycle 2, and many genes involved in specific phases of the cell cycle such as Ki-67, ribonucleotide reductase M2, polo-like kinase, and topoisomerase IIA. This cluster also included pituitary tumor-transforming 1, a gene that is overexpressed in many cancers, is tumorigenic in vivo, and has been shown to bind p53 (22)
. The gene product of serine/threonine kinase 6 (STK6) is also present and has cell-cycle-dependent expression, with maximum expression in G2-M (23)
; in addition, STK6 has been shown to bind chromosome 20 open reading frame 1 (24)
, which is also repressed and in this cluster. Squalene epoxidase was down-regulated in the luminal cell lines and is a gene that was differentially expressed between luminal and basal tumors in vivo (5)
.
A large cluster of genes that include DNA-damage and stress-response genes was up-regulated in response to treatment in the luminal lines (Fig. 1D)
. p21waf1 and the DNA-damage response gene GADD45 were induced strongly in both lines. Also present in this cluster were a number of genes involved in xenobiotic metabolism including carboyxlesterase 2, epoxide hydrolase, and ferredoxin reductase. The latter two of these genes, along with p21waf1 and GADD45, are all known to be p53-regulated (25
, 26)
. Induction of xenobiotic metabolism genes may represent a stereotyped adaptive response of the cell to DNA damage.
Chemotherapeutic-Induced Gene Expression Patterns in Basal Cell Lines.
A much smaller list of genes showed significantly altered expression in the ME16C or HME-CC basal cell lines (Fig. 2
and Supplemental Fig. 2). Using the combined list of genes that were significantly altered in either basal cell line in a hierarchical clustering analysis showed that the basal lines did not cluster as distinctly as the luminal cell lines (Fig. 2B)
. Within the treated branch, some time points for the HME-CC line clustered on separate branches, but the drug-treated ME16C experiments all grouped together. This suggests that the changes induced in basal cells treated with chemotherapeutics were subject to more temporal variation. The changes also appeared more subtle; strong signatures like those observed in the luminal cell lines were not nearly as evident in these basal cell lines. We identified a small cluster of genes that was slightly induced in the sham experiments, but that was down-regulated in the treated experiments (Fig. 2C)
. Many of these genes are involved in cellular differentiation including integrin-ß4, collagen type XII1, COX2, and core promoter element-binding protein. A proliferation signature (similar to Fig. 1C
) was not identified in the treated basal lines. However, a set of genes involved in the DNA damage and/or stress response was identified (Fig. 2D)
and similar to the luminal cell lines (Fig. 1D)
, p21waf1 was induced, although less dramatically. Several xenobiotic metabolism genes were also up-regulated including the p53-regulated genes ferredoxin reductase and quinone oxidoreductase homolog, as well as glutathione-S-transferase 
(GST-). Inhibitor of DNA binding 3, an inhibitor of differentiation (27)
, was also induced in both basal cell lines.
|
and Supplemental Fig. 3). The grouping of the cell lines identified two primary dendrogram branches (Fig. 3B)
, one representing only basal cell lines and one representing predominantly luminal cell lines. A few of the late time points for the basal cell lines fell within the luminal branch but remained distinct on their own secondary branches because of their unique expression profiles. This finding again illustrates the temporal variability across the basal cell lines time points.
|
. This gene set included retinoblastoma 1, ribonucleotide reductase M2, MCM4, chromosome 20 open reading frame 1 and pituitary tumor-transforming 1, all of which regulate cell proliferation or have cell cycle-dependent expression (20)
. A cluster of genes whose expression was induced in luminal lines and repressed in basal lines is shown in Fig. 3E
, whereas the gene set in Fig. 3F
was induced in both cell types, but was more highly induced in luminal cells versus basal cells. Among these genes was X-box binding protein 1 (XBP1), a gene whose expression was previously shown to be highly expressed in luminal tumors in vivo (5)
. XBP1 is a transcription factor involved in mediating the unfolded protein response (28)
, which may represent a stress response that is more prominent in secretory luminal cells. HER2 also appeared to be induced more distinctly in luminal cells treated with chemotherapeutics (Fig. 3E)
. HER2 has been extensively studied in breast cancer, and it has been shown that MCF-7 cells that overexpress HER2 retained their proliferative advantage after DOX treatment in a human breast cancer xenograft model (29)
. Fig. 3F
illustrates that the DNA damage response was much more dramatic in the luminal cell lines, with expression of p21waf1 and prostate differentiation factor being highly up-regulated in luminal cells and less dramatically induced in basal cell lines. We also confirmed the cell-type-specific differences in basal versus luminal induction of p21waf1 on the protein level by Western blot (Fig. 4)
.
|
in which the samples are arranged by tumor subtype as defined in Sørlie et al. (Ref. 5
; not clustered); genes were clustered and all fold-changes are displayed relative to the median gene expression level. A total of 28 cDNA clones representing 23 genes were more highly expressed in the after-samples relative to the before-samples (no genes were significantly lower in the after-samples). These findings agree (at least 13 genes in common) with a similar analysis performed on a subset of these data (30)
. Among these 23 genes were the AP-1 coactivators FOS and JUN, p21waf1, and a number of other genes involved in wound healing including connective tissue growth factor and matrix metalloproteinase 9.
|
, we normalized the expression ratio in each after-sample to its paired before-sample (displayed in black) on a gene-by-gene basis. This allowed visualization of the changes caused by chemotherapy in each patient. These changes are difficult to discern in Fig. 5A
because of the diversity of initial expression values in the before-samples. Relative to the paired before-samples, nearly all of the after-samples in all four of the tumor subtypes showed induced expression of these genes, despite their diverse expression ranges in Fig. 5A
. A few of the tumors had anomalous behavior, underscoring the individuality of tumor responses even within a subtype.
Based on our cell line data, we hypothesized that there might also be tumor subtype-specific responses; therefore, we conducted analyses on the before-samples versus the after-samples for the basal and luminal subtypes separately. Using 81 luminal tumor samples and SAM analysis, we identified 14 genes that were changed in expression after treatment. Using 21 basal tumor samples, we identified nine genes that were induced after treatment (Table 3)
. In this analysis, there was a five-gene overlap between the basal and luminal gene lists. A number of genes that were seen in the combined analysis were significantly altered in only one of the subtypes. For example, p21waf1 was present only on the luminal list and core promoter element-binding protein (COPEB) was present only on the basal list. Next, we compared each cell type’s in vivo list with its corresponding in vitro list and identified four genes that were altered in luminal tumors and luminal cell lines (p21waf1, elongin A, prostate differentiation factor, and thrombospondin 1). COPEB was the only gene that was significantly altered in both the basal tumors and cell lines.
|
and 5
). For example, when the basal cell line list was used to cluster all of the tumor samples, p21waf1 and MDM2 clustered together and showed higher expression in the luminal tumors, whereas COPEB and GST- showed higher expression in the basal tumors (Supplemental Fig. 4). When the luminal cell line list was used to cluster all of the tumor samples, subtype-specific responses were also evident; for example, the basal tumors showed high expression of the proliferation signature both before and after chemotherapy. This is consistent with the in vitro findings because the proliferation signature was unchanged in the basal cell lines after chemotherapy treatment. |
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-positive breast tumors, whereas HME lines (finite life span or immortalized) have expression similarities with basal breast tumors (33)
.
In our model of breast cancer, basal and luminal epithelial cells have unique transcriptional responses to the chemotherapeutics DOX and 5FU. The two luminal cell lines showed similar response patterns to one another including the strong induction of DNA damage/stress response genes, notably p21waf1 (Figs. 1D
and 3F
). The basal cell lines showed a much less dramatic induction of p21waf1 (Fig. 2D
and 3F
). All four of our cell lines are wild type for p53 by sequence analysis and express p53 protein (data not shown); therefore, the differences in p21waf1 expression cannot be attributed to differences in p53 status. p21waf1 is involved in the G1 checkpoint response, and others have reported an impaired G1 checkpoint in HME cell lines (34)
. Consistent with a strong cell cycle checkpoint in the luminal cell lines, MCF-7 and ZR-75–1 cells also repressed a large set of proliferation genes (Fig. 1C)
. This suggests that their G1 checkpoint in response to DNA damage is intact (20)
. The two basal cell lines did not repress the proliferation signature, but they did down-regulate genes involved in differentiation (Fig. 2C)
.
The basal cell lines that we used for this study were hTERT-immortalized HME cells, whereas the luminal cell lines were derived from human tumors. Telomerase expression is a hallmark of breast cancer (35)
, but increased telomerase expression is one of many changes that are observed as cells progress toward a malignant state (36)
. Although breast tumor derived cell lines of luminal origin are widely studied, analogous lines of basal origin have not yet been identified. We acknowledge that comparison of breast cancer lines versus immortalized breast lines represents a starting point for investigations of these cell types. Future comparisons using additional cell lines, and preferably cancer cell lines of basal origin, may yield more data of greater significance. We note that some of the expression differences observed between the basal and luminal cell lines could be due to differences in tumorigenicity. However, we found that our cell lines recapitulated some of the cell type differences seen in vivo in response to these same agents (Table 3
; Supplemental Figs. 4 and 5). The overlap observed between the tumors and cell lines is significant, especially considering three differences between these data sets: (a) the tumor data were acquired using cDNA microarrays and a common reference sample whereas the cell lines were assayed using 60mer oligonucleotide arrays and a cell-line specific reference (untreated pooled reference); (b) the cell lines were all p53 wild type, whereas 40% of the tumors were p53 mutant. Thus, the in vivo analysis is more likely to have excluded some p53-dependent responses to chemotherapy; and (c) the tumors represent a heterogeneous cell population and the cell lines represent only a single cell type.
A strength of tumor profiling studies is that they capture the heterogeneity of tumors in their natural environment. However, this heterogeneity makes it difficult to study the chemotherapy responses of specific cell types. The role of each cell type in a tumor can begin to be dissected using cell-line models, preferably with multiple cell lines representing each cell type. Cell lines are as unique as the tumors from which they were derived, but common response patterns can only become identifiable when looking at multiple cell lines in concert. This was illustrated in a recent study of 60 cell lines and 60,000 compounds (33 , 37) in which relationships between sets of cell lines, sets of genes, and toxicant sensitivity were identified. In the work presented here, we used four cell lines with two cell lines representing each of two tumor subtypes. Characterizing common responses and interindividual variation in these cell lines will help to identify those responses that are stereotypical for each cell type.
Recent studies have demonstrated that DNA-damaging agents induce generic stress responses. In 2000, Gasch et al. (11) showed that yeast displayed a stereotypic pattern of gene expression when exposed to a wide range of stresses including heat shock, growth factor deprivation, and treatment with hydrogen peroxide. These authors termed the stereotypic response the "environmental stress response (ESR)." The environmental stress response included repression of growth-related genes and genes encoding ribosomal proteins and induction of genes involved in DNA damage response and metabolism. These results are in agreement with our finding that a major response to treatment included repression of genes involved in cell growth and induction of DNA damage response genes. Our work with breast cell lines corroborates other recent human cell line studies that have demonstrated common stress responses after DNA-damaging treatments (12 , 26 , 38, 39, 40) . In this article, we have demonstrated that some of the changes seen in vitro were also observed in vivo.
Finally, we note that DOX and 5FU have distinct mechanisms of action (41
, 42)
. For example, DOX is thought to target topoisomerase IIA blocking the G2-M transition and 5FU targets thymidylate synthase blocking S-phase progress. In our experiments with luminal cell lines, both drugs affected gene expression in all phases of the cell cycle (Fig. 1C)
. These cell cycle genes serve as proliferation markers and are not specific to a single mode of action. The specific mechanisms of action of DOX and 5FU may be evident in a subset of genes expressed in our experiments and subsequent analyses will attempt to identify this gene set. However, to fully validate toxicant-specific gene sets, it must also be demonstrated that the gene set predicts mode of action for independent data sets on mechanistically similar drugs. Our primary objective for this work was to understand how cell types differed in their stress response patterns, which are the dominant gene expression responses to DNA damage. The identification of cell-type specific stress responses in vitro and in vivo has implications for understanding the biological response to therapy.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.orig); M. Troester and K. Hoadley contributed equally to the work.
Requests for reprints: Charles Perou, Lineberger Comprehensive Cancer Center, University of North Carolina at Chapel Hill, Campus Box 7295, Chapel Hill, NC 27599-7295. Phone: (919) 843-5740; Fax: (919) 843-5718; E-mail: cperou{at}med.unc.edu
10 All figures and text can be obtained at the supporting website for this article, https://genome.unc.edu/pubsup/TOX/. 
Received 1/13/04. Revised 3/10/04. Accepted 4/ 8/04.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  H. Q. Ta, K. S. Thomas, R. S. Schrecengost, and A. H. Bouton A Novel Association between p130Cas and Resistance to the Chemotherapeutic Drug Adriamycin in Human Breast Cancer Cells Cancer Res., November 1, 2008; 68(21): 8796 - 8804. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Lu, K. A. Becker, M. J. Hagen, H. Yan, A. L. Roberts, L. A. Mathews, S. S. Schneider, H. T. Siegelmann, K. J. MacBeth, S. M. Tirrell, et al. Transcriptional Responses to Estrogen and Progesterone in Mammary Gland Identify Networks Regulating p53 Activity Endocrinology, October 1, 2008; 149(10): 4809 - 4820. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Goncalves, E. Charafe-Jauffret, F. Bertucci, S. Audebert, Y. Toiron, B. Esterni, F. Monville, C. Tarpin, J. Jacquemier, G. Houvenaeghel, et al. Protein Profiling of Human Breast Tumor Cells Identifies Novel Biomarkers Associated with Molecular Subtypes Mol. Cell. Proteomics, August 1, 2008; 7(8): 1420 - 1433. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Blancafort, M. P. Tschan, S. Bergquist, D. Guthy, A. Brachat, D. A. Sheeter, B. E. Torbett, D. Erdmann, and C. F. Barbas III Modulation of drug resistance by artificial transcription factors Mol. Cancer Ther., March 1, 2008; 7(3): 688 - 697. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. J.H. Arlander, B. T. Greene, C. L. Innes, and R. S. Paules DNA Protein Kinase Dependent G2 Checkpoint Revealed following Knockdown of Ataxia-Telangiectasia Mutated in Human Mammary Epithelial Cells Cancer Res., January 1, 2008; 68(1): 89 - 97. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Meslin, A. Hamai, P. Gao, A. Jalil, N. Cahuzac, S. Chouaib, and M. Mehrpour Silencing of Prion Protein Sensitizes Breast Adriamycin-Resistant Carcinoma Cells to TRAIL-Mediated Cell Death Cancer Res., November 15, 2007; 67(22): 10910 - 10919. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Meslin, R. Conforti, C. Mazouni, N. Morel, G. Tomasic, F. Drusch, M. Yacoub, J. C. Sabourin, J. Grassi, S. Delaloge, et al. Efficacy of adjuvant chemotherapy according to Prion protein expression in patients with estrogen receptor-negative breast cancer Ann. Onc., November 1, 2007; 18(11): 1793 - 1798. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Diallo-Danebrock, E. Ting, O. Gluz, A. Herr, S. Mohrmann, H. Geddert, A. Rody, K.-L. Schaefer, S. E. Baldus, A. Hartmann, et al. Protein Expression Profiling in High-Risk Breast Cancer Patients Treated with High-Dose or Conventional Dose-Dense Chemotherapy Clin. Cancer Res., January 15, 2007; 13(2): 488 - 497. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Lacroix, R.-A. Toillon, and G. Leclercq p53 and breast cancer, an update. Endocr. Relat. Cancer, June 1, 2006; 13(2): 293 - 325. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. E. Hochreiter, H. Xiao, E. M. Goldblatt, S. M. Gryaznov, K. D. Miller, S. Badve, G. W. Sledge, and B.-S. Herbert Telomerase Template Antagonist GRN163L Disrupts Telomere Maintenance, Tumor Growth, and Metastasis of Breast Cancer. Clin. Cancer Res., May 15, 2006; 12(10): 3184 - 3192. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. S. Oh, M. A. Troester, J. Usary, Z. Hu, X. He, C. Fan, J. Wu, L. A. Carey, and C. M. Perou Estrogen-Regulated Genes Predict Survival in Hormone Receptor-Positive Breast Cancers J. Clin. Oncol., April 10, 2006; 24(11): 1656 - 1664. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. Chen, J. Houldsworth, A. B. Olshen, G. Nanjangud, S. Chaganti, E. S. Venkatraman, J. Halaas, J. Teruya-Feldstein, A. D. Zelenetz, and R. S. K. Chaganti Array comparative genomic hybridization reveals genomic copy number changes associated with outcome in diffuse large B-cell lymphomas Blood, March 15, 2006; 107(6): 2477 - 2485. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Yuan, Y. Xu, J.-H. Woo, Y. Wang, Y. K. Bae, D.-S. Yoon, R. P. Wersto, E. Tully, K. Wilsbach, and E. Gabrielson Increased Expression of Mitotic Checkpoint Genes in Breast Cancer Cells with Chromosomal Instability Clin. Cancer Res., January 15, 2006; 12(2): 405 - 410. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. C. Mallory, G. Crudden, A. Oliva, C. Saunders, A. Stromberg, and R. J. Craven A Novel Group of Genes Regulates Susceptibility to Antineoplastic Drugs in Highly Tumorigenic Breast Cancer Cells Mol. Pharmacol., December 1, 2005; 68(6): 1747 - 1756. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Camphausen, B. Purow, M. Sproull, T. Scott, T. Ozawa, D. F. Deen, and P. J. Tofilon From The Cover: Influence of in vivo growth on human glioma cell line gene expression: Convergent profiles under orthotopic conditions PNAS, June 7, 2005; 102(23): 8287 - 8292. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Bertucci, P. Finetti, J. Rougemont, E. Charafe-Jauffret, N. Cervera, C. Tarpin, C. Nguyen, L. Xerri, R. Houlgatte, J. Jacquemier, et al. Gene Expression Profiling Identifies Molecular Subtypes of Inflammatory Breast Cancer Cancer Res., March 15, 2005; 65(6): 2170 - 2178. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Wesierska-Gadek, M. Gueorguieva, and M. Horky Roscovitine-induced up-regulation of p53AIP1 protein precedes the onset of apoptosis in human MCF-7 breast cancer cells Mol. Cancer Ther., January 1, 2005; 4(1): 113 - 124. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. D. Ramirez, S. Sheridan, L. Girard, M. Sato, Y. Kim, J. Pollack, M. Peyton, Y. Zou, J. M. Kurie, J. M. DiMaio, et al. Immortalization of Human Bronchial Epithelial Cells in the Absence of Viral Oncoproteins Cancer Res., December 15, 2004; 64(24): 9027 - 9034. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. O. Nielsen, F. D. Hsu, K. Jensen, M. Cheang, G. Karaca, Z. Hu, T. Hernandez-Boussard, C. Livasy, D. Cowan, L. Dressler, et al. Immunohistochemical and Clinical Characterization of the Basal-Like Subtype of Invasive Breast Carcinoma Clin. Cancer Res., August 15, 2004; 10(16): 5367 - 5374. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
