Development and Analytical Validation of a Quantitative Tissue-Based Assay for Phospho-Smad2 (Ser465/467) in Glioblastoma Multiforme: A Potential Biomarker of TGF-β Pathway Activation.
Timothy R Holzer, Mark Gustavson, Robert Pinard, Jason Christiansen, Lonnie Graves, Scott P Myrand, Sunil K Kadam, Michael MF Lahn, John T Brandt, Jonathan M Yingling, Aejaz Nasir. Eli Lilly and Co., Indianapolis, IN; HistoRx, Branford, CT
Background: Using immunohistochemical assays, we and others have shown that high p-Smad2 expression is associated with shorter survival in patients with glioblastoma multiforme (GBM). This study focuses on: 1) Development and optimization of a robust assay for quantification of p-Smad2 in archival GBM tissues from one institution and 2) Cross-platform validation of adverse prognostic significance of high nuclear p-Smad2 expression on an independent GBM cohort.
Design: The study group included 108 adult patients (Male:Female/68:34; Mean age: 53.7 years) with primary GBM treated with standard of care therapy at a single academic medical center. A multi-core GBM tissue-microarray (TMA) was stained for p-Smad2, using fluorescence immunohistochemistry (AQUA® technology). For each tumor, nuclear p-Smad2 (Ser465/467) expression was quantified as an AQUA score. Scores were compared among redundant cores from each case, using Pearson's correlation coefficient and GLM-ANOVA. An optimal AQUA score cut-point to differentiate cases with low and high p-Smad2 expression was determined with respect to 2-year overall disease-specific survival, using X-Tile™ graphical cut-point analysis program. Kaplan-Meier survival analysis and Cox Proportional Hazards modeling were performed using the low and high p-Smad2 expression groups of patients.
Results: Of the 108 GBM cases, 89 were evaluable by AQUA analysis. AQUA scores for p-Smad2 expression in redundant cores showed high similarity (Pearson's R = 0.97; GLM ANOVA P = 0.54), suggesting good reproducibility. Using the optimal cut-point above, 62 (70%) GBMs were categorized as low and 27 (30%) GBMs as high p-Smad2 expressers. The high p-Smad2 patient subset showed a significant (p=0.013) decrease in 2-year disease-specific survival. The median survival of this group was 10 months vs. 13 months in patients with low p-Smad2 expression. Despite adjusting for age, a critical determinant for survival in GBM, high p-Smad2 expression remained significantly associated with decreased survival [HR = 1.92 (95%CI: 1.09 – 3.40); p=0.025].
Conclusions: Using a quantitative AQUA-based assay, high p-Smad2 expression was further substantiated as an adverse prognostic factor in primary GBM patients. Since p-Smad2 is a member of the TGF-β pathway, a quantitative assay may be clinically relevant for TGF-β beta inhibitor clinical trials.
Category: Special Category - Pan-genomic/Pan-proteomic approaches to Cancer
Monday, February 28, 2011 1:00 PM
Poster Session II # 235, Monday Afternoon