(P 15) EZHIP: A NOVEL EPIGENETIC DRIVER OF OSTEOSARCOMA PATHOGENESIS

Objective: Osteosarcoma (OS) is the most common malignant bone tumor. OS incidence rates peak during the pubertal growth spurt as mesenchymal progenitors expand and differentiate to form new bone, suggesting that OS is rooted in altered development. Comprehensive genomic studies did not improve clinical care in the past 30 years. Conversely, the OS epigenome remains largely uncharted. Given the pivotal role of epigenetics in orchestrating development, we sought to investigate epigenomic perturbations in Osteosarcomagenesis.

Methods: We evaluated the expression of cancer-driving histone mutations “oncohistones”, the oncohistone-mimic EZHIP and post-translational modifications in 2 independent cohorts (>75 patients) by immunohistochemistry. Next, we interrogated the functional consequences of EZHIP expression in patient-derived cells, mouse orthotopic models and human mesenchymal cells (hMSC). CUT&RUN and RNAseq were used to study the epigenome and transcriptome and results were validated via functional assays including live-cell imaging and in-vitro differentiation. Finally, in-vitro and in-vivo assessment of FDA-approved drugs was performed.

Results: EZHIP is ectopically expressed in ~22% of OS resulting in the loss of H3K27me3. Importantly, acute H3K27me3 loss significantly correlated with poor response to neoadjuvant chemotherapy. At a cohort-level, relapsed tumors exhibited significantly lower H3K27me3 levels compared to primaries, highlighting its important role in OS progression. CRISPR/Cas9 knockout and lentiviral overexpression showed that EZHIP is a prominent oncogene required for in-vivo tumorigenesis in NSG mice. Transcriptomic and functional investigation revealed that EZHIP activates key oncogenic pathways in hMSC and OS cells. Moreover, EZHIP promotes the expansion of pre-osteoblastic osteo-CAR programs that lock hMSC in a stem-like pre-malignant state at the expense of normal differentiation which becomes blocked. Lastly, EZHIP expression imparts sensitivity to EZH2 inhibitors, including Tazemetostat, in orthotopic models.

Conclusion: EZHIP and H3K27me3 loss are novel mechanisms in osteosarcoma (OS) pathogenesis. While H3K27me3 loss is a poor prognostic marker, it also presents a therapeutic opportunity especially in chemoresistant OS, through the use of Tazemetostat and other EZH2 inhibitors. Further investigation combining Tazemetostat with current standards of care for patients with refractory tumors is warranted.