FET fusion oncoproteins hijack the LASR complex to rewire alternative splicing in sarcomas

The FET (FUS, EWSR1, TAF15) genes are commonly involved in chromosomal translocations resulting in their fusion with various transcription factors (TF) genes. These genomic abnormalities are hallmarks of several sarcomas and leukemias, and are found along few other alterations in these neoplasms. The chimeric proteins encoded by these fusion genes share a similar architecture, with a strong aminoterminal transactivation domain derived from FET proteins, and a carboxyterminal DNA-binding domain derived from the TF partner. As this structure is reminiscent of that of a TF, the oncogenic potential of FET fusion proteins has been first attributed to their ability to reprogram transcription. However, this transcriptional role is not sufficient to fully explain how these oncoproteins single-handedly drive various cancers. Indeed, growing evidence points towards novel post-transcriptional roles for FET fusions, notably in the alternative splicing of pre-mRNA. Such a function has previously been studied for the prototypical FET fusion EWSR1::FLI1, the main driver of Ewing sarcoma. Interestingly, RBFOX2, an RNA-binding protein (RBP) governing alternative splicing as part of a large assembly of splicing regulators (LASR) complex, has been identified as a key functional partner of EWSR1::FLI1.
In this project, we aim to determine whether the pre-mRNA alternative splicing function observed for EWSR1::FLI1 could be a shared mechanism promoting FET fusion-driven oncogenesis. RNA-sequencing of various FET-translocated sarcoma cell lines showed that thousands of alternative splicing events are induced when the expression of the corresponding fusion is prevented. This suggests that FET fusions inforce a specific splicing landscape in their corresponding sarcoma. In addition, we showed that a representative panel of FET fusions promoted exon inclusion of a reporter minigene, but only when directly tethered onto its pre-mRNA, suggesting that the control of splicing by FET fusions might be direct and might rely on its recruitment onto pre-mRNA. As FET fusions lack canonical RNA-binding domains, we hypothesized that endogenous recruitment could be mediated indirectly via an RBP. By performing luciferase-based protein complementation assays, we demonstrated that almost all FET fusions could interact with RBFOX2. Surprisingly, RBFOX2 appeared to preferentially interact with the C-terminal domains of the fusions, which are derived from a wide range of unrelated TFs. RBFOX2 is known to function as part of a hetero-multimeric splicing complex called LASR. Based on our genome-wide splicing analysis analysis, we found that RNA-binding motifs of RBFOX2 and other members of the LASR complex were found enriched in the proximity of differentially regulated cassette exons. We further validated that members of LASR co-immunoprecipitated with FET fusions. Finally, we confirmed the importance of RBFOX2 and other LASR components in the modulation of splicing of endogenous and cancer-related gene targets of FET fusions.
Altogether, our work provided evidence supporting a direct role in splicing for FET fusions by interacting with the LASR complex. Although the mechanisms by which the oncoproteins and LASR collaborate to promote oncogenesis remain unclear, we have established a moonlighting function for these driver fusion proteins that could be crucial in our understanding of the tumorigenesis of multiple neoplasms.