Abstract
194 Background: The urokinase plasminogen-activator (uPA) and its inhibitor (PAI-1) predict the benefit of node-negative breast cancer patients from chemotherapy. The determination of uPA and PAI-1 need(s) large amounts of tumor tissue. In contrast, microarray-based gene expression profiling (MGEP) requires only small quantities of FF tissue and is applicable to FFPE archival tissue. Identification of MPEG-derived surrogate markers for ELISA-based uPA/PAI-1 may facilitate their analysis together with MPEG-based predictive genes and metagenes. Methods: Three groups of breast cancer biopsies were analysed. Group A: 136 FF tissues from 2008-2009, group B: 85 FF tissues from 2010, group C: 20 independent FFPE tissues. Gene expression was assessed by Agilent 4x44K microarrays. Results: For group A a significant correlation between protein expression of uPA and uPA gene (PLAU) expression (r = 0.627 p < 0.001) and PAI-1 and PAI-1 genes (SERPINE1)(r = 0.281 p = 0.001) was found. Top sets of genes correlated with uPA/PAI-1 protein expression were extracted. For group B additional sets of top correlated genes were identified. The correlation of individual genes with uPA/PAI-1 protein expression was highest for their genes themselves (PLAU: r = 0.694 p < 0.001; SERPINE1 (r = 0.469 p < 0.001). Subsequently consensus gene sets were identified with a good performance for both biopsy sets. Only genes which were significantly correlated with their counterparts in group C were selected, resulting in a final 11-gene model (uPA metagene) and a 6-gene model (PAI-1 metagene). They significantly correlated with uPA/PAI-1 protein expression in all groups of biopsies (uPA: group A: r = 0.621, p < 0.001; group B: r = 0.771, p < 0.001; group C: r = 0.665, p = 0.005; for PAI-1: group A: r = 0.569, p < 0.001; group B: r = 0.687, p < 0.001; group C: r = 0.683, p = 0.008). Conclusions: Reliable MPEG-derived metagenes as surrogate markers for ELISA-based uPA/PAI-1 expression were identified.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have