Characterization of alternative mRNA 3'-processing in mouse progenitor B-cell lymphoma

The molecular bases of cancer development are highly complex, and defy easy characterization. Studies of gene deregulation in cancer have typically focused on transcriptional control, but an emerging picture shows that post-transcriptional processing and regulation can also cause critical changes in gene expression. Using a mouse model of pre-B cell acute lymphoblastic leukemia (preB-ALL), the major cancer of early childhood, I have investigated connections between mRNA processing and lymphoid neoplasia.;Mice doubly deficient in p53 and critical DNA repair factors develop lymphomas that fall into three subsets with distinct survival curves but similar histology and overall gene expression profiles. A probe-level analysis of the microarray data reveals characteristic differences in mRNA processing amongst the three tumor subtypes. I observed significant changes of mRNA processing in ∼400 genes, many of which are implicated in cancer pathogenesis. My approach specifically seeks to identify 3'-untranslated region (3'-UTR), which aberrantly exclude or include regulatory elements from the mRNA (e.g., miRNA or RNA-binding protein target sequences), creating the potential to corrupt post-transcriptional regulation of these genes. Preliminary sequence analysis of alternatively processed genes shows putative evidence of alterations in miRNA activity. Additionally, increased protein expression of a critical polyadenylation trans-factor in the tumors suggests that the polyadenylation machinery also plays an active role in the alternative processing.;Importantly, genes that show exclusive differential processing in each tumor subtype define signatures that can be used for tumor classification. An internal cross-validation analysis demonstrated that these signatures correctly classified subtypes with 55-80% accuracy. This suggests that alternative mRNA 3'-end processing could provide powerful new diagnostic or prognostic markers. Moreover, my data leads to novel insights into the molecular mechanisms of tumorigenesis. Finally, my findings emphasize the critical importance of measuring changes in transcript processing, as well as overall levels, to understand altered expression during tumorigenesis.