The Breast Cancer Epithelial-Stromal Interactome
Eun-Yeong Oh, Nicholas M Knobluch, Andrew H Beck. Beth Israel Deaconess Medical Center & Harvard Medical School, Boston, MA
Background: The significance of tumor microenvironment in breast cancer tumorigenesis and invasion is increasingly recognized. It is, however, largely unknown how tumor epithelial cells interact with the stromal cells to make their microenvironment more advantageous for tumor growth and invasion. Our aim is to discover epithelial-stromal cross-talk partners in breast cancer.
Design: Using the Gene Expression Omnibus, we identified 2 LCM expression profiling datasets. Study 1 consisted of 5 samples captured from normal epithelium (NE1) and stroma (NS1), and 28 samples captured from breast cancer epithelium (BrE1) and stroma (BrS1). Study 2 consisted of 14 samples from each of: normal epithelium (NE2), normal stroma (NS2), breast cancer epithelium (BrE2), and breast cancer stroma (BrS2). We performed a genome-wide epithelial-stromal expression association study to identify pairs of genes showing highly correlated expression (either positive or negative) between epithelial and stromal compartments. This analysis was implemented using the MatrixEQTL package in R. We performed 4 sets of analyses (NE1vs NS1, NE2 vs NS2, BrE1 vs BrS1, and BrE2 vs BrE2) and identified epithelial-stromal cross-talk pairs consistently identified as showing highly correlated expression in breast cancer epithelial and stromal compartments but not in the normal breast epithelial and stromal compartments.
Results: We identified 10 breast cancer epithelial-stromal cross-talk pairs at an FDR of < 0.02%. 9 of the interactions indicated a positive association between epithelial and stromal expression levels, and 1 of the pairs indicated an inverse association. For genes in these 10 pairs, we assessed their association with survival using a publicly available breast cancer expression profiling meta-dataset of 2898 breast cancer patients with clinical follow-up. This analysis shows that for the pairs NPM1(epi)-FOSL2(str), S100A8 (epi)-S100A8(str), and CYP4F8(epi)-PAX1(str), both members of each of the gene pairs are significantly associated with survival (all P's < 0.02), further suggesting that these epithelial-stromal cross-talk partners may play important roles in breast cancer invasion and progression.
Conclusions: This sysetmatic analysis of the breast cancer interactome identified 10 novel candidate epithelial-stromal cross-talk partners, including 3 pairs of markers associated with breast cancer survival. These findings provide new insights into the biology of the breast cancer microenvironment and represent novel candidate therapeutic targets for manipulating epithelilal-stromal interactions in breast cancer treatment.
Monday, March 4, 2013 9:30 AM
Poster Session I Stowell-Orbison/Surgical Pathology/Autopsy Awards Poster Session # 33, Monday Morning