Virginia R. Litle1, Andrew Feber1, Liqiang Xi1, James D. Luketich2, Arjun Pennathur2, Rodney J. Landreneau2, Maoxin Wu1, Scott J. Swanson1, Tony E. Godfrey1; 1Cardiothoracic Surgery, Mount Sinai Medical Center, New York, NY; 2University of Pittsburgh, Pittsburgh, PA
Objective: Micro RNA (miRNA) expression in tumors has been associated with patient outcome, cell differentiation, proliferation and apoptosis. The purpose of our study was to examine miRNA expression in esophageal cancer and premalignant lesions to gain a better understanding of esophageal tumor biology and identify potential markers for diagnosis and progression.Methods: miRNA was isolated from 30 tissue specimens comprising of 10 esophageal adenocarcinomas (AC), 10 squamous cell carcinomas (SCC), 6 normal squamous epithelium (NSE), 3 Barrett's esophagus (BE) and 1 high grade dysplasia (HGD). miRNA expression was analyzed using Ambion Bioarrays containing probes for 328 human miRNA's. Data were analyzed using NCI, BRB-array tools V3.5.
Results: Unsupervised hierarchical clustering resulted in three major branches on which, samples are predominantly clustered based on histology. One branch consists of nine SCC samples and only one SCC is not included. A second branch consists of all ten AC specimens as well as one BE and the lone HGD sample. The remaining branch consists of six NSE samples, two BE and one SCC. Thus, this arm represents "normal" tissue with the exception of the one misclassified SCC. This representation shows that the SCC tumors are more similar to NSE than the AC samples which are themselves more similar to the BE. Pairwise comparisons between sample types revealed several miRNA's which may be markers of tumor progression. Mir_203 and mir_205 are expressed 3-10 fold lower in SCC and AC compared with NSE while mir_21 expression is 2-4 fold higher in both tumors versus NSE. Other miRNA's, such as mir_192 appear to be more specific to AC. Interestingly both mir203 and mir_192 have been shown to influence proliferation in lung cancer cells and one predicted target of mir_203 is a gene that has already been implicated in esophageal cancer (ECRG4).
Conclusion: miRNA expression profiles clearly distinguish esophageal tumor histologies and discriminate normal tissue from tumor. Our data also provide molecular evidence for the origin of AC from BE and show that HGD is molecularly more similar to AC than BE. Furthermore, one BE sample clustered with the AC samples and this may indicate a premalignant molecular phenotype not seen at the morphologic level. Overall, this study shows that miRNA expression may provide insight into, and markers for, disease diagnosis and progression in esophageal malignancies.
Back to Annual Meeting
Back to Program Outline
