Correspondence

A Gene Signature in Breast Cancer

N Engl J Med 2007; 356:1887-1888May 3, 2007

Article

To the Editor:

Liu et al. (Jan. 18 issue)1 report on a 186-gene “invasiveness” gene signature (IGS) that discriminates between normal breast epithelium and tumorigenic breast-cancer cells that are characterized by CD44 expression and low or undetectable levels of CD24. This signature is associated with survival among patients with breast cancer. We profiled 200 breast tumors with the use of DNA microarrays and determined their molecular subtype. We used the 186-gene IGS to classify the basal and the luminal samples. The two subtypes were well separated. We found that 89% of genes that were overexpressed in cancer stem cells were coordinately overexpressed in basal samples. This association provides a link between an intrinsic feature of breast cancers (basal subtype) and cancer stem cells, reinforcing the relevance of the 186-gene IGS. The result suggests that basal-cell breast cancers are enriched in tumorigenic breast-cancer stem cells or maintain a similar transcriptional profile owing to a block of differentiation, which partially explains the poor prognosis for patients with such tumors.

Francois Bertucci, M.D., Ph.D.
Nathalie Cervera
Daniel Birnbaum, M.D., Ph.D.
Institut Paoli-Calmettes, 13009 Marseille, France
bertuccif@marseille.fnclcc.fr

1 References

1. 1

   Liu R, Wang X, Chen GY, et al. The prognostic role of a gene signature from tumorigenic breast-cancer cells. N Engl J Med 2007;356:217-226
   Full Text | Web of Science | Medline

To the Editor:

Liu et al. conclude that the 186-gene IGS they developed is new, since it shows little overlap with other breast-cancer signatures. However, the IGS they discuss is similar to a “genetic grade” signature defined and characterized by Ivshina et al. in a data set from more than 400 patients.1 Both signatures seem to be prognostic for tumors of intermediate grade only. To address this issue, we tested the independence of the IGS and the 264 genes identified by Ivshina et al., further motivated by a recent article suggesting dependency among several breast-cancer signatures in spite of little overlap.2 On the basis of our data, the IGS and 264 genetic-grade classifiers had a correlation of 0.81 (P<0.001) and were in agreement on the classification of 85% of tumors (Figure 1Figure 1Correlation between the Invasiveness Gene Signature (IGS) and the Genetic-Grade Signature.). Since the 264-gene signature is largely a proliferation marker, ruling out dependency on proliferation now seems to be important in demonstrating the novelty of the IGS.

Kristian Wennmalm, M.D.
Karolinska Institutet, S-171 76 Stockholm, Sweden
kriwen@ki.se

Lance D. Miller, Ph.D.
Genome Institute of Singapore, Singapore 13872, Singapore

Jonas Bergh, M.D., Ph.D.
Karolinska Institutet, S-171 76 Stockholm, Sweden

2 References

1. 1

   Ivshina AV, George J, Senko O, et al. Genetic reclassification of histologic grade delineates new clinical subtypes of breast cancer. Cancer Res 2006;66:10292-10301
   CrossRef | Web of Science | Medline

2. 2

   Fan C, Oh DS, Wessels L, et al. Concordance among gene-expression-based predictors for breast cancer. N Engl J Med 2006;355:560-569
   Full Text | Web of Science | Medline

To the Editor:

Liu et al. report that the status of the estrogen receptor and differentiation in tumors outperformed the prognostic value of the IGS among patients in their study. We suggest that their data may be limited by both the paucity of samples from which this signature was derived and the heterogeneous histologic features, tumor origin, and treatment. The current findings should be validated before incorporating this signature into clinical trials.

Wendy A. Woodward, M.D., Ph.D.
Anthony Lucci, M.D.
Massimo Cristofanilli, M.D.
University of Texas M.D. Anderson Cancer Center, Houston, TX 77030
wwoodward@mdanderson.org

for the Advanced Research Center for Micrometastatic Disease

Author/Editor Response

We agree with Bertucci and colleagues that basal-type carcinomas probably represent a subgroup of breast cancers in which the cancer cells are blocked at an early stage of differentiation.

We agree with Woodward and colleagues that the estrogen receptor is an important prognostic factor. Nonetheless, as shown in Table 3 of our article, the IGS is associated with the prognosis in estrogen-receptor–positive patients. Basal-cell breast cancers are generally estrogen-receptor–negative, so such tumors were removed from our analysis. Therefore, the prognostic power of the IGS is not simply based on the identification of estrogen-receptor–negative or basal-cell breast cancers. We also agree that the IGS was based on the isolation of breast-cancer stem cells from only a few samples and that analysis of a much larger group of patients could result in an improved signature. Nonetheless, the fact that a signature derived from just a few breast-cancer stem cells has prognostic power speaks to the value of investigating this population of cancer cells.

Dr. Wennmalm and colleagues ask whether the IGS has prognostic power if proliferation genes are removed. To test how important the proliferation-related genes are to the gene signature, we removed all 15 genes involved in cell proliferation or the cell cycle from the IGS and used the remaining genes as a signature for predicting the outcome of patients in the Netherlands Cancer Institute database. The remaining gene signature was still significantly associated with death (P<0.001; hazard ratio, 1.34) and metastasis (P<0.001; hazard ratio, 1.29).

Michael F. Clarke, M.D.
Stanford University, Stanford, CA 94305

Rui Liu, Ph.D.
University of Michigan, Ann Arbor, MI 48109

Xinhao Wang, Ph.D.
Oncomed Pharmaceuticals, Mountain View, CA 94043

Citing Articles (7)

Citing Articles

1. 1

   Felipe de Sousa E Melo, Selcuk Colak, Joyce Buikhuisen, Jan Koster, Kate Cameron, Joan H. de Jong, Jurriaan B. Tuynman, Pramudita R. Prasetyanti, Evelyn Fessler, Saskia P. van den Bergh, Hans Rodermond, Evelien Dekker, Chris M. van der Loos, Steven T. Pals, Marc J. van de Vijver, Rogier Versteeg, Dick J. Richel, Louis Vermeulen, Jan Paul Medema. (2011) Methylation of Cancer-Stem-Cell-Associated Wnt Target Genes Predicts Poor Prognosis in Colorectal Cancer Patients. Cell Stem Cell 9:5, 476-485
   CrossRef

2. 2

   Thomas Stricker, Daniel V.T. Catenacci, Tanguy Y. Seiwert. (2011) Molecular Profiling of Cancer—The Future of Personalized Cancer Medicine: A Primer on Cancer Biology and the Tools Necessary to Bring Molecular Testing to the Clinic. Seminars in Oncology 38:2, 173-185
   CrossRef

3. 3

   Nam P. Nguyen, Fabio S. Almeida, Alex Chi, Ly M. Nguyen, Deirdre Cohen, Ulf Karlsson, Vincent Vinh-Hung. (2010) Molecular biology of breast cancer stem cells: Potential clinical applications. Cancer Treatment Reviews 36:6, 485-491
   CrossRef

4. 4

   Chiang, Anne C., Massagué, Joan, . (2008) Molecular Basis of Metastasis. New England Journal of Medicine 359:26, 2814-2823
   Full Text

5. 5

   Wendy A. Woodward, Erik P. Sulman. (2008) Cancer stem cells: markers or biomarkers?. Cancer and Metastasis Reviews 27:3, 459-470
   CrossRef

6. 6

   Danila Coradini, Manuela Gariboldi, Maya Fedeli, Marco A. Pierotti, Maria Grazia Daidone. (2008) Invasiveness gene signature predicts a favorable outcome also in estrogen receptor-positive primary breast cancers treated with adjuvant tamoxifen. Breast Cancer Research and Treatment 111:2, 389-390
   CrossRef

7. 7

   G Storci, P Sansone, D Trere, S Tavolari, M Taffurelli, C Ceccarelli, T Guarnieri, P Paterini, M Pariali, L Montanaro, D Santini, P Chieco, M Bonafé. (2008) The basal-like breast carcinoma phenotype is regulated by SLUG gene expression. The Journal of Pathology 214:1, 25-37
   CrossRef