Investigating tamoxifen resistance in
the luminal B estrogen receptor positive breast cancer subtype:
Tailoring treatment in hormone responsive breast cancer.
Loi
S, Haibe-Kains B, Phillips WA, Lallemand
F, Campbell IG, Piccart MJ, Sotiriou C and McArthur GA
Background: We recently
reported that the level of expression of proliferation genes in
estrogen receptor-positive (ER+) breast cancer (BC) is a powerful
predictor of prognosis, and high levels of proliferation are associated
with tamoxifen resistance (Loi et al, JCO 2007). This study aimed to
explain the biological basis for these observations using global gene
expression profiling.
Methods: 246 ER+ BC samples
from women treated with adjuvant tamoxifen only were analyzed with gene
expression arrays and evaluated using gene set enrichment analysis
(GSEA). 173 samples were sequenced for PIK3CA mutations. Ingenuity
Pathway Analysis was used to generate interaction networks. MCF-7 BC
cells treated heregulin (HRG) was used as an in vitro model of ERBB2
pathway activation.
Results: We found that gene
sets suggesting ERBB2 and PI3K pathway activation were significantly
enriched in the tamoxifen resistant subtype. HRG treated MCF7 cells
displayed phosphorylation of HER2/3 without HER2 overexpression.
Treatment with HRG overcame tamoxifen induced cell cycle arrest
(p<0.01) and increased anchorage-independent colony formation
(p<0.01), validating the hypothesis that ERBB2 signaling without
overexpression contributes to tamoxifen resistance. Samples were
sequenced for PIK3CA mutations to investigate PI3K pathway activation.
Mutations were found in 45 samples (26%). There was no significant
association with mutation and lymph node status, tumor grade or age.
Tumor size had a borderline association (p=0.056). Contrasting gene
expression profiles of mutation positive samples and wild type revealed
81 significantly differentially expressed genes (p=0.03). A
gene-expression predictor consisting of 35 genes based on gene
interactions was then constructed. The PIK3CA associated gene set could
predict clinical outcome in 1,228 independent ER+ BC samples HR: 1.39
(95%CI 1.23-1.58) p<0.001.
Conclusions: Using gene
expression profiling, we have identified signaling pathways in ER+ BC
that could be targeted in the clinical setting. Furthermore, we propose
gene signatures of activated ERBB2 and PI3K pathways that may help
stratify patients for therapy selection.