(P 305) A NEW MOUSE MODEL FOR TESTING SPLICE-SWITCHING THERAPIES IN RHABDOMYOSARCOMA

Objective: Recent advances in high throughput sequencing technologies have uncovered a surprising number of alternatively spliced variants associated with tumorigenesis, implicating regulated splicing in the tumor phenotype. We have shown that the alternatively spliced shorter A variant of the insulin receptor gene (INSR) is highly expressed in pediatric rhabdomyosarcoma (RMS) and can circumvent blockade of the IGF-1R receptor to allow persistent IGF-2 signaling making these cancers resistant to treatment. We have designed antisense oligonucleotides with splice-switching capabilities (splice-switching oligonucleotides, SSOs) to target the INSR pre-mRNA. We hypothesize that these SSOs will restore the normal INSR-B isoform, decrease IGF-2 signaling effects that are facilitated by the shortened INSR-A isoform, and prove to be a promising therapeutic opportunity in mouse models, in the short term, and in human clinical trials and beyond, in the long-term. However, it is necessary to develop animal models that will avail SSO testing in relevant pre-clinical models. Hence, we have turned to the mouse as a powerful tool to generate such a model.

Methods: To generate a new mouse model for RMS that will avail testing of splice-altering therapies in tumors in their native environment (as opposed to cell lines and xenografts), we performed genome editing technologies to induce subtle mutations in intronic splice elements to induce expression of alternatively spliced Insr A in the context of its full-length Insr B counterpart. To test the ability of the mutations to affect the splicing of exon 11, we generated a mouse splice-reporter minigene that includes exon 10, 11 and 12 with intervening introns. We then used site-directed mutagenesis to induce specific mutations that were predicted to disrupt the constitutive splicing pattern of Insr. We generated our mouse mouse mutant line harboring the MBNL G113A/G116A mutation along with a mutation that destroys the protospacer adjacent motif (PAM) to prevent Cas9 cleavage of the template and increase targeting efficiency. Because the G116A and PAM G95A mutants each introduce a unique restriction site, were able screen the resultant mice with PCR followed by restriction digestion. For confirmation, we also sequenced PCR amplicons that span the target mutation site.

Results: We have demonstrated that mouse embryos predominantly express the Insr A isoform, consistent with the splicing INSR ratio seen in human embryonic tissues. Furthermore, we reveal that the muscle tissue expresses the full-length Insr B isoform as early as post-natal day 1, which persists into adulthood. The mouse Insr pre-mRNA splicing is likewise responsive to hypoxia as hypoxic stress induces decreased levels of the mouse Insr B isoform in multiple mouse cell lines. Furthermore, we have established that the Insr splice regulatory elements are conserved in mice and that SSO targeting the CUGBP negative splice regulatory element can restore full-length Insr splicing to yield increased levels of the Insr B isoform. We illustrate that the mutations in the INSR minigenes induced increased levels of the transcript skipping exon 11, with the MBNL G113A/G116A mutant expressing 43% of the smaller transcript, compared to just 4% in the wild-type minigene. Furthermore, Introducing MBNL binding site mutations into the Insr genomic locus results in increased skipping of INSR exon and leads to the expression of stress markers in the muscle.

Conclusion: Our INSR splice-sensitized mouse model can be crossed to existing RMS models to sensitize mice to INSR splicing and RMS. By obtaining an Insr splice-sensitive mouse model we will be able to obtain pre-clinical response data and defining mechanisms of action and biodistribution and targeting of SSOs, we will be well-suited to devise novel, clinically translatable strategies for enhancing cancer therapies.