Favorable outcome associated with an IGF-1 ligand signature in breast cancer.

The insulin-like growth factor (IGF) axis is fundamentally important in cell growth, development and cancer. We used genomic technologies to better characterize the activity of the IGF axis in human breast cancer and to identify predictors of response to IGF targeted therapies. Analysis of the gene expression patterns and pathway analysis in 204 clinically annotated primary breast cancers were performed and compared to levels of mRNA for IGF ligands and receptors. Pathway activation scores were calculated by Pearson correlation (+1, -1). Network analysis was performed using Ingenuity software. IGF-1 ligand levels were strongly negatively correlated (P < 10(-6)) with a published IGF-IR activation signature. A signature of high IGF-1 ligand was associated with better prognosis (P = 0.025-1.5 × 10(-8)) in several public datasets. Pathway analysis revealed upregulation of pathways associated with breast differentiation (adipocyte growth factors, PPAR-gamma) and down-regulation of proliferation pathways (AKT/MAPK) in the IGF-1 ligand high group. Of note, the IGF-1 ligand signature was anti-correlated with IGFIR receptor levels (P = 0.07). In conclusion, a breast tumor-derived signature of high IGF-1 ligand is associated with favorable outcome, in contrast to a previously reported IGF-IR activation signature. The prognostic value of the IGF-I ligand signature is validated in three independent datasets. These signatures should be applied in study of IGF1-R targeted therapy.