An integrated analysis of three distinct IBC/non-IBC affymetrix gene expression data sets to study the transcriptional heterogeneity both between IBC and non-IBC and within IBC.

Abstract

10571 Background: Studies have applied expression profiling to inflammatory breast cancer (IBC). Most of these studies were underpowered, mainly because IBC sample sizes were small. Here, we present an integrated analysis of 3 distinct gene expression data sets of IBC and nIBC samples – thus with enhanced power - to further uncover the molecular biology of IBC. Methods: 3 Affymetrix expression data sets of 137 IBC and 252 nIBC samples were integrated, normalized, summarized and quality-controlled. Samples were classified according to the molecular subtypes. IBC-specific heterogeneity was investigated using clustering, coupled with silhouette analysis. Supervised analyses were performed in non-stage-, stage-, and molecular subtype-matched approaches. IBC-specific activated pathways, miRNA-families, and transcription factors were identified with a target gene analysis approach. Results: 4 Robust IBC sample clusters were identified, associated with the molecular subtypes, predominated by the combined Basal-like, ErbB2+, and Luminal B subtypes (~70% vs. ~40% in non-IBC; P<0.0001). When comparing IBC to nIBC, stage-matched and non-stage-matched differences were identified (global test, P<0.0001). When comparing IBC with nIBC samples within each of the molecular subtypes, differences persisted only within the luminal A and normal-like subtypes (global test, P<0.0001 and P=0.046, resp.). Target gene analysis identified two molecular pathways (TGFb and INFa), 16 transcription factors (e.g. NKX2.2, FOXM1), and 10 miRNA families that were differentially activated in IBC and nIBC (FDR<0.01) but not between the 4 IBC sample clusters. Conclusions: We show that IBC is indeed transcriptionally heterogeneous and that differences between IBC and nIBC are predominated by the differences in the distribution of the molecular subtypes. Taking this into account, we were able to more accurately spot IBC-specific changes in pathway, miRNA, and transcription factor activation.