Genome-Wide Aberrant Splicing in Patients with Acute Myelold Leukemia (AML) Is Associated with Altered Expression of Splicing Factors

Abstract 652

Long-term survival of patients with acute myeloid leukemia (AML) is poor, and new forms of therapy are needed. Many genetic lesions have been identified and studied, and most patients have chromosome translocations or other mutations that promote self-renewal of leukemic stem cells, block differentiation, enhance growth, and block apoptosis. Only a few of these mutations result in druggable targets (e.g., PML-RARa, Kit, PDGFR, FLT3 for instance). In addition to genetic lesions, epigenetic abnormalities have been shown to be very common in AML, and provide opportunities for novel treatments. Using genome-wide approaches to identify alternative splicing, we have recently shown that AML cells have a high level of aberrantly regulated genome-wide alternative splicing (AS) as a frequent epigenetic event. By comparing samples from 62 AML patients with 10 normal donors (NDs) we identified 428 genes differentially spliced in AML. A list of differentially spliced genes includes 50 oncogenes and 52 tumor suppressor genes, as well as genes encoding proteins involved in cell proliferation and differentiation, and apoptosis. We evaluated splicing event frequency in AML compared to NDs and we observed that on average 527 (range 137–1657) genes were identified as differentially spliced in any given patient, out of 62 analyzed. Also, we found that any given differentially spliced gene, of the 3,108 detected, were spliced on average in 26 (range 1–54) AML patients. Thus, splicing aberrations are highly recurrent in AML patients.

To identify the causes of aberrant splicing in AML, we evaluated transcript levels of the 24 major splicing factors (SFs) that are involved in the first and second splicing transesterification reactions. These splicing factors are important proteins involved in spliceosomalassembly. Expression levels of these SFs were evaluated in 20 AML patients exhibiting high levels of AS. Quantitative RT-PCR analysis showed significant (up to 30 fold) upregulation of U2AF2 (P<2.00E-07), PTBP (P=3.00E-04) and SFRS12 (P=0.002) SF transcript levels in AML patient samples compared to CD34+ cells from NDs. In preliminary studies, we also detected elevated expression of U2AF2 and PTBP proteins in several patient samples. These results suggest the intriguing possibility that aberrant splicing in AML may be the result of alterations of these SFs. To test this hypothesis we generated stably transfected HEK293 cell lines overexpressing U2AF2 or PTBP. We have developed a synthetic semi-quantitative splicing assay to evaluate the effects of overexpression of these SFs.

We have obtained a minigene cassette of the p53 inducible PIG3 gene based on previous splicing studies. The minigene cassette was cloned between RFP (red fluorescent prtoein) and GFP (green fluorescent prtoein) in such a way that translation of the normally spliced transcript results in expression of RFP and GFP, while aberrant splicing results in the expression of RFP only. Production of a similar minigene cassette that includes exons/introns of a gene that is subjected to aberrant splicing in AML (NOTCH2, FLT3 and CD13) is in progress. In studies, completed so far, with the PIG3 minigene cassette construct transiently transfected into the HEK293 cells lines, overexpression of PTBP increased aberrant splicing of the PIG3 minigene. Similar studies testing the effects of elevated levels of U2AF2 and PTBP on NOTCH2 and other genes misspliced in AML (such as FLT3 and CD13) will be presented. Our results indicate that aberrant splicing could be an important event in AML, and development of an in vitro, synthetic splicing assay will enable us to better understand the underlying causes of this process in AML.