Identification of Alternative Splicing Events Regulated by the BCR-ABL Oncoprotein

Post-transcriptional modifications, including alternative splicing, are critical in determining the protein inventory of normal human cells. A number of malignancy-related and apoptosis-related genes have de-regulated alternative splicing due to mutation or deletion of genomic splice sites or through other, less well understood mechanisms. Chronic myelogenous leukemia (CML) is characterized by a translocation that produces the dominant acting BCR-ABL oncoprotein, which is responsible for genomic instability, resistance to apoptosis, cellular proliferation, differentiation abnormalities, and many other acquired cellular changes. It is likely that additional undiscovered properties of the BCR-ABL oncogene contribute to CML pathogenesis. Imatinib, an inhibitor of the BCR-ABL tyrosine kinase, induces a high rate of complete cytogenetic remission in CML. However, the ability of imatinib to eradicate the leukemia stem cell is unknown. Moreover, additional treatment targets will need to be identified to optimize therapies for imatinib-refractory patients. We used gene array analysis to investigate whether alternative splicing may be important in the biology and pathogenesis of CML. Two analyses were conducted, examining the splicing patterns of a total of 587 genes. 570 of the genes have been implicated in apoptosis, and 17 of the genes play a role in hematologic malignancies. In the first experiment, splicing profiles of all 587 genes (5832 independent splice events) were evaluated using a customized ExonHit array. Splicing profiles of three CML cell lines were compared to a well characterized primary thymus control tissue. Dye swaps confirmed experimental consistency. Bioconductor software was used to select events with greater than two-fold differential expression. We have identified 124 candidate splice forms in 91 genes that are differentially regulated in at least two of the three CML cell lines compared with control hematopoietic tissue. 88 of the 91 are implicated in apoptosis, while the remaining three are among the genes implicated in hematologic malignancies. The candidate genes identified by these means include WT1, RUNX1, PFDN5, and BRCA1. In a second experiment, the same 587 genes were examined in two CML cell lines treated with imatinib and compared to untreated controls. Prior pilot experiments identified a dose of imatinib that inhibited BCR-ABL but did not cause apoptosis during the course of the experiment. Using the same experimental procedures and analysis, we identified 41 candidate splice forms in 41 genes that were differentially expressed in imatinib-treated CML cell lines. Among these candidate genes, 40 are from the group of apoptosis-related genes, and one is from the group of genes related to hematologic malignancies. Candidate genes include MAPK3, FES, DCC, and TGFB1. Combining the two experiments, a group of 122 genes and 163 splice events were identified as potential BCR-ABL directed, imatinib-sensitive candidates. We performed validation experiments for one of these, WT1. Imatinib treatment consistently altered the ratios of specific splice forms present in CML cells. These findings are consistent with BCR-ABL dependent splicing of WT1 in CML. Because WT1 isoforms have differential effects on the survival, proliferation, and differentiation of leukemia cells, regulation of WT1 splicing by BCRABL may play an important role in the pathogenesis of CML. Overall, our results suggest that the BCR-ABL oncoprotein may cause significant dysregulation of mRNA splicing as part of the process of neoplastic transformation. We are currently validating selected genes among the 122 candidates in primary CML samples, performing bioinformatics analysis of the specific tumor-specific splice site anatomy to gain insights into potential upstream regulators, and investigating apoptosis regulation by selected candidate splice regulated targets.