(P 84) CSF1R OVEREXPRESSION ACCELLERATES OSTEOSARCOMA TUMOR FORMATION AND GENERATES TUMORS WITH DECREASED IMMUNE CELL INFILTRATION

Objective: Understanding the molecular events responsible for transcriptional variation (heterogeneity) within osteosarcoma tumors is challenging, especially considering the extreme inter-tumor genomic and transcriptional heterogeneity in this rare and deadly cancer. The objective of this study was to investigate the underlying mechanisms of transcriptional variation in osteosarcoma using genetically engineered mouse models.

Methods: To address this objective in an unbiased manner, we employed a genetically engineered mouse model where an oncogenic transposon was mobilized using the Osx1-Cre transgene. This model allowed for the accelerated formation of osteosarcoma tumors in both wild-type and Trp53R270H background mice, regardless of their sex. RNA-Seq data was generated from the resulting tumors, including those formed without transposon mobilization in Trp53R270H mutant mice. By analyzing the RNA-Seq data, we identified transposon-gene fusions that acted as drivers of tumor formation. Systematic analyses of transcript levels using GCESS dimensional reduction revealed conserved transcriptional variation in the genetically engineered mouse tumors, as well as in endogenously derived human and canine osteosarcoma tumors. By comparing the observed transcriptional patterns with the recurrent fusions, we generated hypotheses regarding the role of the driver events in tumor formation, which were further validated via generation of additional genetically engineered mice.

Results: Recurrent fusions were observed in the transposon-accelerated mouse tumors, and these fusions varied based on genetic background, sex, and original chromosome location of the transposon. However, no strong differences were observed in the tumor transcriptional patterns based on the Sleeping Beauty (SB) acceleration, genetic background, sex, or original chromosome location of the transposon. Comparison of the fusions with the observed transcriptional patterns suggested that overexpression of Csf1r was oncogenic and immunosuppressive, due to a decrease in the presence of immune-related transcripts. Csf1r is primarily expressed by immune cells, as indicated by single-cell RNA-Seq datasets, and it is known to regulate hematopoietic proliferation and differentiation. We hypothesized that Csf1r overexpression in tumors might interfere with Csf1-mediated signals, leading to the reduced presence of immune cells observed in the tumors. To test this hypothesis, we generated Csf1r overexpressing mice. These showed accelerated tumor formation on a Trp53R270H mutant background and the tumors exhibited decreased levels of immune cells. Additionally, we found a strong correlation between CSF1 ligand, CSF1R, and overall immune transcript levels in human, canine, and murine osteosarcoma, except for the murine tumors that overexpress Csf1r.

Conclusion: In conclusion, our findings indicate that overexpression of Csf1r leads to the generation of osteosarcoma tumors with reduced immune presence. This is significant because previous studies have shown that reduced immune presence in osteosarcoma and other sarcoma types is associated with poor patient outcomes. By therapeutically increasing CSF1 levels within tumors, we may promote the proliferation and differentiation of immune cells, potentially improving patient outcomes. This approach offers advantages over personalized therapies, as it may benefit all patients rather than just those with specific targets