Acute lymphoblastic leukemia (ALL) is an heterogeneous disease comprising several subentities that differ for both immunophenotypic and molecular characteristics. Over the years, the biological understanding of this neoplasm has largely increased. Gene expression profiling has allowed to identify specific signatures for the different ALL subsets and permitted the identification of pathways deregulated by a given lesion. MicroRNAs (miRNAs) are small noncoding RNAs, which play a pivotal role in several cellular functions. In this study, we investigated miRNAs expression profiles in a series of adult ALL cases by microarray analysis. Unsupervised hierarchical clustering largely recapitulated ALL subgroups. Furthermore, we identified miR-148, miR-151, and miR-424 as discriminative of T-lineage versus B-lineage ALL; ANOVA highlighted a set of six miRNAs-namely miR-425-5p, miR-191, miR-146b, miR-128, miR-629, and miR-126-that can discriminate B-lineage ALL subgroups harboring specific molecular lesions. These results were confirmed and extended by quantitative-PCR on a further cohort of cases. Finally, we used Pearson correlation analysis to combine miRNA and gene expression profiles. The distribution of correlation coefficients generated by comparing the expression of every miRNA/gene pair in our data set shows enrichment of both positively and negatively correlated pairs over background distributions obtained using randomized data. Moreover, a clear enrichment for predicted miRNA:target pairs is observed at negative correlation coefficient intervals. Signal-to-noise ratio highlighted several miRNA/gene pairs with a possible role in the disease. In fact, gene set enrichment analysis of genes composing the selected miRNA/gene pairs displays over-representation of functional categories related to cancer and cell-cycle regulation.
(c) 2009 Wiley-Liss, Inc.