Abstract A35: From map to model: Genetic evolution of diffuse intrinsic pontine glioma provides the framework for accurate tumor modeling

• Published in: Cancer research (Chicago, Ill.) Ill.), 2020-06, Vol.80 (11_Supplement), p.A35-A35
• Main Authors: Lubanszky, Evan, Ryall, Scott, Siddaway, Robert, Kim, Byunjin, Bridge, Taylor, Pajovic, Sanja, Hawkins, Cynthia
• Format: Article
• Language: English
• ISSN: 0008-5472; 1538-7445
• DOI: 10.1158/1538-7445.CAMODELS2020-A35

Diffuse intrinsic pontine glioma (DIPG) is a devastating pediatric high-grade glioma (HGG) that diffusely penetrates the brainstem and is the leading cause of pediatric brain tumor death. Around 80% of DIPGs harbor lysine 27-methionine (K27M) substitutions in histone H3. H3K27M inhibits the activity of Polycomb repressive complex 2 (PRC2) that deposits H3K27me3, leading to widespread H3K27me3 loss, but how this leads to cancer is unclear. Further, despite clonal analysis indicating that H3K27M is the tumor-initiating mutation in DIPG, in vivo models of H3K27M-driven cancer have failed to generate tumors without additional oncogenic drivers. Thus, the early events driven by H3K27M alone remain obscure as does the most relevant cell of origin. To address the role of H3K27M in tumor initiation we generated a transgenic murine model with a tetracycline-repressible, Cre-inducible H3F3A-K27M transgene within the ROSA26 locus. This allows for both spatial and temporal control of K27M expression, via the cre-inducible LSL cassette and tetracycline treatment, respectively. These mice were crossed with the brain specific -cre drivers Nestin-, Olig2-, or GFAP-, targeting neural stem cells, oligodendrocyte progenitor cells, and glial progenitor cells, respectively. H3F3A K27M/-cre double-positive mice were born significantly below mendelian ratios, the most severe being those crossed with Olig2-cre (Nestin- 18/100; 0.18, GFAP- 15/114; 0.132, Olig2- 9/106; 0.08). All mice displayed a degree of ataxia and a smaller overall stature compared to littermate controls, with the most severe phenotype being particularly evident in Olig2-cre mice. Additionally, both Nestin- and GFAP-cre crosses experienced decreased survival post-weaning age (p21) (Nestin- 13/18 alive, GFAP- 10/15 alive); however, this decrease was greatly exaggerated in Olig2-cre (1/9 alive). Together, these data suggest that the displayed phenotypes are dependent upon the cell population in which H3F3A K27M is expressed. We also found that in addition to spatial control, temporal control of H3F3A K27M expression is critical to survival. Doxycycline (dox) was administered to breeding pairs for each H3F3A K27M/-cre line to repress transgene expression. Dox-H3F3A K27M/-cre double-positive mice were born at higher mendelian ratios than their no-dox counterparts (Nestin- 26/79; 0.33, GFAP- 14/50; 0.28, Olig2- 17/52; 0.327). Litters remained on dox treatment until either p0 or p21. Delayed H3F3A K27M expression led