Identification of a microRNA Signature Useful in the Differential Diagnosis between Burkitt Lymphoma and DLBCL
S Montes-Moreno, L di Lisio, N Martinez, G Gomez, H Rodriguez, SM Rodrigez-Pinilla, P Mckelvie, JL Rios, J Menarguez, JC Cigudosa, MA Piris. CNIO, Madrid, Spain; HGUGM, Madrid, Spain
Background: MicroRNAs (miRNAs) have been associated with B cell development, precise differentiation stages and lymphomagenesis, and proposed to constitute lymphoma-type specific markers. We have tested whether microRNA profiling can be used for the differential diagnosis between Burkitt Lymphoma (BL) and Diffuse Large B-cell Lymphoma (DLBCL), analyzing also a set of B cell lymphoma cases with features intermediate between BL&DLBCL.
Design: A screening series of 42 cases (30 DLBCL and 12 BL) was studied for miRNA expression using a one colour microarray containing probes for 470 human miRNAs. SAM analysis was performed to identify a first set of 43 differentially expressed candidate microRNAs (FDR<0.05 and fold change>1.5 log2). RT-PCR using total RNA extracted from FFPE tissue was then performed to refine and validate this result in a larger cohort of 85 (40 BL & 45 DLBCL) new cases. Class prediction analysis by different statistical algorithms (Random Forests, KNN) was performed and a predictor signature was derived. The predictor was applied to a series of BCL intermediate between BL&DLBCL (12 cases).Target prediction was performed using both Pictar and Targetscan databases.
Results: A signature of 23 miRNAs was found to be differentially expressed between BL and DLBCL (FDR < 0.05). Of note MYC regulated microRNAs (miR92, miR26b, miR29b, let7f) were present in the signature. A restricted signature of 20 miRNAs was derived from class prediction analysis and performed well (error rates ranging 0.12-0.16) in the distinction between BL and DLBCL. Furthermore, when Intermediate BCL cases where included in this analysis, clustered together BL and the higher accuracy was obtained when this cases were considered as BL (all but one were properly classified). Target prediction using current databases indicate that specific genes commonly related with BL or DLBCL gene expression signatures such as CD10, MYBL1, BCL2, PIM1 and FOXP1 have been found to be targets of this microRNAs.
Conclusions: A microRNA signature has been found to be differentially expressed between BL and DLBCL. The presence of MYC regulated microRNAs in this signature demonstrates its value as markers of MYC overexpression. This signature can be considered an alternative tool for the classification of these B cell lymphoma types including cases of B cell lymphoma with features intermediate between BL&DLBCL.
Monday, March 22, 2010 8:15 AM
Platform Session: Section B, Monday Morning