Purpose : The aim of this study was to classify breast carcinomas based on variations in gene expression pattern derived from cDNA microarrays and to correlate tumor characteristics to clinical outcome.
Methods : A total of 7 pairs of breast tumors and control tissues were taken at the time of primary surgery for array analysis. Then, performed microarray experiments in breast cancer tissues with the cDNA microarray spotted 3,063 clones of genes, were analyzed by hierachical clustering.
Results : Thirteen genes were over expressed in tumor samples regardless of their histopathological features and ER status, those were including, vitamin A responsive gene, proliferating cell nuclear antigen (PCNA), and signal transducer and activator of transcription 1 (STAT1). Twenty-four genes were down-regulated in tumor sites, those were including, discoidin domain receptor family mamber 2, crystallin alpha B, and myosin light polypeptide kinase. We also identified the differentially expressed genes between ER positive and negative tumors. PCNA, FLJ20500, STAT1, signal recognition particle 9 kD, and proteasome activator subunit 2 were more predominantly expressed in ER negatives. Serine protease 23, vitamin responsive gene, fibronectin 1, and SERPINA1 genes were more highly expressed in ER positive tumors. We further classified the patients according to their gene expression patterns with Cluster program. Clustering results divided patients into two distinct groups, the first group consists of only estrogen receptor (ER) negative tumors and they showed more higher gene expression levels of cell replication and cycle, invasion and metastasis, those considered poor prognosis signature. The other group mostly consists of ER positive tumors.
Conclusion : These results support the feasibility and usefulness of this systematic approach to studying variation in gene expression patterns in human cancers as a means to dissect and classify solid tumors. We believe, this gene expression profile will outperform all currently available clinical parameters in predicting disease outcome.
|