Abstract P6-04-09: Lack of response to aromatase inhibitors involves distinct mechanisms

Abstract

Abstract Background: Estrogen is a key hormone involved in cancer cell development. It forms a complex with estrogen receptor alpha (ER) which binds to estrogen response elements (EREs) in the promoter regions of genes under its transcriptional control. Letrozole inhibits local production of estrogen. Despite the prevalent use of such drugs little is known about their effect at molecular and transcriptomic level. A better understanding of the molecular mechanisms underlying why approximately 75% of ER rich cancers respond might help to elucidate the mechanisms characterising a lack of response in some patients. Methods: A new gene expression dataset was generated from sequential core biopsy/theatre samples (before treatment, 10–14 days on treatment and at surgery) taken from ER+ patients undergoing neo-adjuvant letrozole treatment. Clinical response to treatment was assessed by changes in tumour volume based on 3D ultrasound (performed by a single operator). Results: On letrozole the majority of ER+ tumours respond with dramatic early transcriptomic changes. A significant reduction in proliferation is seen through down-regulation of key cell cycle control genes such as CyclinD1, CyclinA, CyclinB1, CyclinB2 and CDK2, and genes involved with the initiation phase of DNA replication such as the MCM complex. Several of these (including CyclinD1, CyclinB2 and MCM2) have been shown to contain candidate ERE sequences in their promoter region and therefore down-regulation of these would be concordant with deprivation of estrogen. There is also a significant up-regulation of genes including collagens, lamanins, integrins and others involved with intra-cellular adhesion and extra-cellular matrix (ECM) remodelling. Preliminary analysis suggests the involvement of a local immune response as a potential mechanism for tumour cell death, as genes involved with antigen presentation, leukocyte trans-endothelial mediation, natural killer cell mediated cytotoxicity and T-cell mediated cell death are significantly up-regulated within the clinically responding group. Two distinct subsets were seen in non-responders. The first subgroup shows a ‘classical non-response profile’ with very little change in expression of genes involved with cell cycle, ECM remodelling, cellular adhesion and antigen presentation between baseline and 14-days of treatment. The second subgroup have a gene expression profile similar to the responding group with a down-regulation of cell cycle and up-regulation of genes involved with ECM remodelling, cellular adhesion and antigen presentation. Whether these are true non-responders is not clear. Conclusion: Approximately half of clinically non-responding patients have molecular profiles similar to responding patients whilst the other half have a distinct pattern of expression with significantly less change in genes associated with cell cycle control, ECM remodelling, intra-cellular adhesion and antigen presentation. On-going work will elucidate long term outcomes in these two groups. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P6-04-09.