Abstract 786: In-vivo and in-vitro breast cancer profiling to delineate adaptive endocrine resistance: A pilot study

A substantial fraction of estrogen receptor (ER) positive breast cancers fail to respond to endocrine treatment, i.e. tamoxifen and aromatase inhibitors. While tamoxifen failure has been partially attributed to primary resistance, for example CYP2D6 deficiency (Schroth et al JAMA 2009), adaptive resistance during ER blockade/estrogen deprivation plays a prominent role. Little is known about shared molecular patterns driving both resistance types. Long-term estrogen deprived (LTED) breast cancer can be induced to undergo apoptosis by pulse estrogen (Song et al JNCI 2001; Jordan JCO 2008). We hypothesize that it should be possible to identify predictive markers for adaptive endocrine resistance and sensitivity to apoptosis to recognize endocrine refractory tumors. To study a putative interface of primary and acquired resistance we performed gene expression profiling with low-density arrays (Fluidigm BioMark system) in tumors of 12 breast cancer patients treated with tamoxifen. They were selected to model CYP2D6 independent primary resistance: half of the patients had functional CYP2D6 but were tamoxifen resistant due to relapse (median time-to-event 3.8 years). The other half of patients were CYP2D6 deficient but tamoxifen sensitive due to the absence of relapse (median follow-up 13.5 years). Expression profiles have been compared to LTED MCF-7 cells (3 and 6 months estrogen deprived) and MCF-7 cells grown in the presence of estrogen thereby modeling adaptive resistance. Gene targets were selected based on literature i) involved in endocrine resistance (N=62), ii) playing a role in transport and estrogen supply (N=30), and iii) being p53 targets for apoptosis and cell-cycle check points (N=46). We observed both shared and specific expression phenotypes for adaptive and primary resistance. The proliferation genes Cyclin B1 and STK15 were downregulated in LTED MCF-7 cells in line with lowered proliferation rates and, unexpectedly, Cyclin B1 expression was also decreased in tamoxifen insensitive tumors suggesting an involvement of G2/M cell cycle transition. The chemokine receptor CXCR4 was upregulated in both LTED cells and resistant tumors, confirming its prominent role in breast cancer progression. The 25-hydroxyvitamin D3 hydroxylase CYP27B1 was upregulated in LTED, but not in resistant tumors suggesting a different role between the two models. Other potential markers involve differentiation, inflammatory response, cell cycle/apoptosis, and estrogen disposi