|
Back to 2011 Program
Molecular Profiling of Erlotinib Resistance in an In-Vitro Bladder Cancer Model
William C Faust, Marc Manganiello, Justin Zbrzezny, Christina Deliyiannis, Jason Gee, John Libertino, Antonia Holway, Kimberly R Christ
Lahey Clinic, Cambridge, MA
Introduction
We previously reported differential sensitivity of 17 urothelial carcinoma of the bladder (UCB) cell lines to the EGFR inhibitor erlotinib where lines displaying EMT characteristics showed greater resistance. In this study we evaluated the correlation between microRNA (miRNA) expression levels and erlotinib resistance in an in-vitro model of UCB.
Methods
Erlotinib sensitivity was determined by clonogenic assay in 46 UCB cell lines randomly divided into training and test sets. MiRNA expression levels were determined by microarray analysis and confirmed by qRT-PCR. Multilogistic regression analysis and the Random-Forest Algorithm were used to identify microRNAs predictive of sensitivity.
Results
In the training group, 62 miRNAs were significantly different between the 16 sensitive and 14 resistant cell lines. In the resistant group, 38 miRNAs were up-regulated and 24 miRNAs were down-regulated. A predictive model using two miRNAs, resulted in the misclassification of 1 resistant and 2 sensitive lines. Sensitivity and specificity was 93% and 87.5%, respectively, for the detection of resistance while the area under the receiver operator characteristic curve was 0.9554. In the test set of cell lines, the classifier had a PPV of 50% and a NPV of 100%.
DISCUSSION
MiRNAs are a powerful new tool in the molecular diagnosis and treatment of UCB. We have found a group of previously uncharacterized miRNAs that accurately predicts the response of UCB cell lines to erlotinib treatment. Next steps involve bringing this molecular information to the clinic, and using molecular profiles to guide chemotherapeutic treatment decisions in patients with UCB.
Back to 2011 Program
|
