Association of gene copy number alterations in formalin-fixed, paraffin-embedded tumors with prognostic molecular and pathologic tumor subtypes in early-stage breast cancer.

Abstract

572 Background: Human breast cancer is being redefined as a set of discrete molecular phenotypes based on gene expression profiling. The contribution of copy number alterations (CNAs) to these subtypes as the underlying genetic determinants of distinct pathways has been suggested, but less demonstrated. Methods: To address the relationship between CNAs and tumor subtype in early breast cancers, we investigated the use of a molecular inversion probe (MIP) based high density single nucleotide polymorphism (SNP) array from Affymetrix to identify genome-wide DNA CNA's in 971 formalin-fixed, paraffin-embedded stage I/II breast tumors. Results: Heirarchical clustering of CNAs yielded 3 MIP clusters (A, B, and C) that were significantly correlated with luminal A, B and triple negative (TN) molecular subtypes approximated by surrogate tumor markers. The luminal A subtype had the least genetic instability, whereas TN subtype demonstrated significant CNAs across the entire genome. Significant CN gains in chromosomes 1q, 8q, and 16p with losses in chromosomes 8p and 8q were seen irrespective of subtype. Characterization of chromosomal segments showing CN gains/losses in luminal A tumors supported aberrations in genes involved in estrogen receptor expression and function, whereas CNAs in TN breast cancers were primarily associated with genes involved in inflammation, angiogenesis, and cell adhesion. CNAs were significantly greater for larger tumors and high grade tumors whereas there were no significant differences in CN with respect to stage and lymph node status (positive or negative). Conclusions: Using novel MIP technology, we show that CNAs, as patterns, largely align with the luminal and 'basal-like' (TN) tumor subtypes. Thus, our data provide additional evidence that acquired somatic events may underlie, at least in part, the differences in gene expression patterns observed between molecular phenotypes. These studies support the role of genomic instability as a strong determinant of pathologic characteristics associated with metastatic potential, even among early stage breast cancers independent of tumor subtype. No significant financial relationships to disclose.