Quantitative nuclease protection assay (qNPA) for gene expression analysis on breast cancer core biopsies.

Abstract

49 Background: qNPA employs in situ hybridization of detection probes to cross linked mRNA, making it ideal for formalin fixed paraffin embedded (FFPE) tissue. It has been shown to measure gene expression in archived lymphoid and lung cancer tissue. We assessed the feasibility of qNPA to measure differentially expressed genes in pretreatment FFPE core breast biopsies among pathologic responders (pR) and nonresponders (pNR) to preoperative chemotherapy. Methods: We included preoperative breast cancer patients treated at our institution from 2003-09 with FFPE core biopsies. mRNA expression of 170 genes, representing oncogenic pathways or associated with anthracycline and taxane response was measured by qNPA (HTG, Tucson, AZ). Data was normalized to 3 housekeeper genes and average of 3 biologic replicates reported. Seven genes below detection in > 50% samples were excluded. Expression values of 163 unique genes were analyzed for pR vs pNR with dChip software. Empirical FDR was estimated using 1,000 permutations of sample labels. Results: Treatment and response: Sample failure: 6/57 (10%). pR vs pNR did not separate on hierarchical clustering. FLJ12650 and IGFBP2 showed lower expression in pR vs pNR with fold changes of 4.09 and 2.40, respectively (p < 0.01; median FDR: 1/163). FLJ12650 was significant (p < 0.01, median FDR: 0/163) when patients receiving anthracycline ± taxane were analyzed (groups 1-3) and showed a trend (p < 0.05) in group 1 alone. Conclusions: qNPA for limited available FFPE tissue from core biopsies is feasible with acceptable sample failure rates. Small sample size and number of analyzed genes limited definitive conclusion about informative genes in our study. Our FLJ12650 results, a gene coding for membrane Na+/K+ ATPase interacting protein, are consistent with previous findings of overexpression in pNR to anthracyclines + taxanes (Hess et al, JCO 2006, Vol 24; 4236) using fresh tissue. Future qNPA validation of predictive markers, identified by whole transcriptome analysis in a homogenous cohort may provide more definitive results. [Table: see text]