Abstract S2-5: Molecular Profiling of Aromatase Inhibitor-Treated Post-Menopausal Breast Tumours Identifies Determinants of Response

• Published in: Cancer research (Chicago, Ill.) Vol. 70; no. 24_Supplement; pp. S2 - S2-5
• Main Authors: Dunbier, AK, Ghazoui, Z, Anderson, H, Smith, IE, Dowsett, M.
• Format: Journal Article
• Language: English
• Published: 15.12.2010
• DOI: 10.1158/0008-5472.SABCS10-S2-5
• ISSN: 0008-5472

Abstract Aim: To identify molecular mechanisms and predictors of response to aromatase inhibitors (AIs) in primary ER+ breast cancer. Background: Although AIs are highly effective suppressants of estrogen synthesis in postmenopausal women, up to 50% of patients derive little or no clinical benefit from AI treatment. The proliferation marker Ki67 provides a valuable pharmacodynamic marker for response in the presurgical setting and 2 week measurements of this marker have been shown to predict recurrence-free survival1. However, understanding of the precise molecular effects of AIs and causes of resistance is limited. We present, to our knowledge, the largest study of the transcriptional effects of AI treatment in the neoadjuvant setting. Methods: Baseline and 2wk post-treatment core-cut tumor biopsies were obtained from 112 postmenopausal women with stage I to IIIB ER+ early breast cancer who received single agent neoadjuvant anastrozole2. RNA extracted from biopsies was analysed on Illumina 48K microarrays. Genes which changed expression upon anastrozole treatment were identified using paired class comparison of the 81 matched baseline and 2wk pairs from which gene expression data was available. Correlates of response were determined using multiple testing corrected Spearman correlation analysis. Pathway analysis was performed on the ranked list of correlated genes. IHC for Ki67, ER and PgR was performed centrally on FFPE sections from duplicate core biopsies. Results: At a false discovery rate (FDR) of 0.01, 1154 genes were differentially expressed (792 down-regulated and 362 up-regulated). Proliferation-associated genes and classical estrogen-dependent genes such as TOP2A, CCNB2, TFF1, and PDZK1 were strongly downregulated (FDR<1x10-7). Upregulated genes included collagens and chemokines such as COL16A1 and CCL14. Pathway analysis revealed cell cycle, estrogen receptor and apoptosis-related pathways to be amongst the most significantly altered (P<0.005). However, the response of all genes upon treatment varied considerably between patients. Investigation of genes expressed in pre-treatment biopsies which correlated with response to AI as measured by the decrease in Ki67 over two weeks revealed a strong inflammatory signature. Higher expression of pro-inflammatory genes such as TNF, CXCR3 and IL2RA was associated with poorer response, while expression of GATA3 was associated with more favourable response (P<0.001). Multivariable analysis revealed that the relationship of TNF with response was independent of estrogen receptor expression. Conclusions: Consistent with the variable clinical benefit from AI treatment, the molecular response to AI treatment shows substantial hetereogeneity. Poor Ki67 response to anastrozole treatment is associated with higher baseline expression of an inflammatory signature. Further evaluation of this signature has the potential to identify patients who would benefit from additional therapies. 1. Dowsett et al., JNCI, 2007, 99, 167-70. 2. Smith et al., JCO, 2007, 3816-22. Supported by the Mary-Jean Mitchell Green Foundation and Breakthrough Breast Cancer. Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr S2-5.