DIPG-31. MOLECULAR MECHANISMS AND FUNCTIONAL IMPACT OF ABERRANT SPLICING IN DIFFUSE MIDLINE GLIOMAS

Fewer than 1% of children diagnosed with diffuse-midline glioma (DMG) survive for more than 5 years, because no effective therapies exist for these patients. Here, we sought to identify and characterize mechanisms of aberrant splicing (AS) in primary DMG tumors. We observed transcriptome-wide AS (9,805 differential splicing variations in 4,734 genes), and identified a DMG-specific splicing signature, that included known cancer genes. We hypothesize that AS of cancer genes play a role in DMG tumor formation. Assessing whether splicing factor dysregulation impacted known cancer transcripts, we discovered several splicing factors, including SRRM4, SRRM3 and RBFOX3 to be down-regulated in DMG. Additionally, we found an enrichment of binding motifs for these proteins within flanking regions of these mis-spliced exons. We also observed recurrent significant exon inclusion in tumor suppressor SMARCA4, an integral member of the SWI/SNF family of proteins involved in chromatin remodeling. Further, we identified AS in 16 of the 27 members of the SWI/SNF complex, including increased skipping of exon 7 in DPF2, representing a complete mRNA transcript switch in DMG. Since SRRM4, SRRM3 and RBFOX3 are known regulators for neural-specific microexons, we focused on microexon splicing changes, hypothesizing that these regulators may be driving microexon mis-splicing in these tumors. We identified 245 known microexons lost or gained in DMG. Moreover, a quarter of which were observed in known cancer genes, with the most frequent splice event causing gain of a clathrin-binding site in the tumor suppressor BIN1 with a concurrent loss of an out-of-frame microexon in the oncogene BAK1, presumably activating it. Altogether, our results suggest that aberrant splicing may be an alternative mechanism driving DMG tumorigenesis and we are currently molecularly validating a subset of these events with the overall goal of identifying novel therapeutic targets for DMG tumors.