Global Gene Expression Analyses of Three BCC Subsets, Based on the Relative Level of Oct4A

Despite education and aggressive treatment, breast cancer (BC) remains a clinical problem. BC cells (BCCs) can migrate early to metastatic sites where they may exist in cellular dormancy for decades. Presently, there are no consensus markers for cancer stem cells (CSCs) that are involved in tumor initiation and progression, and drug resistance. The current designation of CSCs might comprise similar tumor initiating cells, but at different developmental phase. In order to understand these differences, we developed a working hierarchy of BCCs. We initiated the studies in which three BCC subsets were selected based on the relative expressions of the stem cell-linked genes, Octamer4A (Oct4A). The sorted BCCs were subjected to array analyses using Affymetrix gene chip. Hierarchical clustering indicated distinct gene expression among the three subsets. Differential gene expressions of membrane proteins validated three novel genes, TMEM-98, GPR64 and FAT4. These three genes, in combination of known markers for CSCs, CD44, CD24, aldehyde dehydrogenase 1 (ALDH1) and Oct4A, were used to stratify BCCs led to a working hierarchy of BCCs. The validity of the hierarchical BCCs was applied to blood samples from patients, during relapse, and before and after treatment. These studies resulted in the patients grouped with distinct BCCs in the circulation. The relevance of the latter findings are discussed with regards to prediction of treatment response and time of BC relapse. The findings require a larger cohort of patients in a prospective multi-center study. The stratification could be important to understand treatment response, strategies for alternative approaches, and an understanding of the interaction between particular BCC subsets and the tissue microenvironment. MDA-MB-231 cells were transfected with pEGFP1-Oct3/4 and then selected with neomycin. Three groups of cells were isolated with the FACSDiva cell sorter. The top and lowest 5% of cells according to GFP expression were designated as Oct4hi and Oct4low while the Oct4med cells were in the middle of the two extreme subsets. Cells were counted and processed in quadruplicate in a random order. Due to limited reagents only duplicates of the Oct4med cells were processed.