Abstract A006: Transcriptomic landscape of fusion-driven CNS tumors for the design and validation of junction-specific antisense oligonucleotide-mediated targeting of fusion oncogenes

Abstract Introduction: Aside from standard surgical resection followed by chemo/radiation therapy, there is a dearth of effective therapies for tumors of the central nervous system (CNS) despite increasing understanding of the molecular mechanisms driving oncogenesis and progression. Antisense oligonucleotides (ASOs) have emerged as a powerful method to target disease-causing transcripts through sequence-specific translational blocking or degradation. However, while it is known that oncogenic fusions drive some CNS tumors, few studies have investigated and characterized the landscape of oncogenic fusions and variation in breakpoints, pertinent for the translation of ASOs into a viable therapy. Methods: We performed retrospective analysis of CNS tumors from 1690 patients profiled at the Children’s Hospital of Philadelphia as part of routine clinical molecular workup. Additionally, we have developed an ASO-based therapeutic platform to design ASOs which target the fusion junctions and validate their efficacy in eliminating the targeted fusion transcripts. Results: Our cohort spanned both pediatric (980/1690) and adult (710/1690) ages (range = 0 - 91) and histologic subtypes. In total, 21% of CNS tumor patients (354/1690) harbored an oncogenic fusion; Grade 1 = 38% (218/567), Grade 2 = 6% (12/211), Grade 3 = 12% (22/176), and Grade 4 = 13% (75/570). Fusions were most common in glial neoplasms (27%, 328/1213) followed by ependymal tumors driven by ZFTA rearranged supratentorial ependymoma (13%, 20/150), and mesenchymal tumors (28%, 5/18). The most common fusions included BRAF in Grade 1 tumors and MET and EGFR in Grade 4 tumors. NTRK1/2/3 fusions, which have promiscuous partners spanned all grades, and FGFR1/2/3 fusions were only found in Grade 1 or 4 tumors, most often fused to TACC1 or TACC3. BRAF, NTRK2, and MET fusions had the highest number of unique partners. Investigating unique transcripts, nearly all fusions had conserved breakpoints except for KIAA1549::BRAF fusions, which had 15 unique transcripts. Interestingly, while ZFTA::RELA fusions had relatively conserved breakpoints, all tumors expressed at least 2 unique transcripts (1 major), suggesting co-existence of different chimeric isoforms. MYB::QKI fusions, identified in low-grade gliomas, was selected as a pilot target to evaluate the ASO-based therapeutic platform. A list of five ASOs were tested in mammalian cells expressing MYB::QKI fusion reporters. The most effective ASO was found to reduce MYB-QKI reporter fluorescent signal by ∼83% and fusion transcript level by ∼63% compared to the negative ASO, as measured by cellular imaging and RT-qPCR, respectively. Conclusion: In the current landscape, where therapeutic options for fusion-driven tumors in the CNS are limited, this ASO-based therapeutic platform presents a new avenue for developing personalized nucleic acid-based therapies aimed at improving clinical outcomes for individuals with CNS tumors. Citation Format: Hyojeong Hwang, Jinhua Wu, Jeffery Schubert, Meha Patel, Natali Naveh, Netta Golenberg, Minjie Luo, Yiming Zhong, Lea F Surrey, Marilyn M Li, Monica Pomaville, Kathy Fange Liu, Derek Wong. Transcriptomic landscape of fusion-driven CNS tumors for the design and validation of junction-specific antisense oligonucleotide-mediated targeting of fusion oncogenes [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Fusion-Positive Cancer: From Discovery to Therapy; 2026 Jan 13-15; Philadelphia PA. Philadelphia (PA): AACR; Cancer Res 2026;86(1_Suppl):Abstract nr A006.