ID #489 Coordination of Three Mechanisms Defines Driver Mutations in Replication Repair Deficient Pediatric High-Grade Gliomas
Yuan Chang, Nicholas Fernandez, Jose Dimayacyac, Lucie Stengs, Melissa Edwards, Logine Negm, Vanessa Bianchi, Ayse Bahar Ercan, Sheradan Doherty, Cynthia Hawkins, Anirban Das, Nuno Miguel Nunes, Uri TaboriAbstract
Background
Replication repair deficiency (RRD) results from defects in the mismatch repair system and/or DNA polymerase ε/δ proofreading. Germline RRD predisposes to development of pediatric high-grade gliomas (HGG) characterized by high mutation burden (10-1000 mutations/Mb). The genes and mutations driving cancer and oncogenic profiles in RRD-HGG remain unknown.
Methods
RRD-HGGs from the International Replication Repair Deficiency Consortium were analyzed using clinical data, whole exome, RNA sequencing, DNA methylation profiling, and ddPCR, with paired primary and recurrent tumors used to assess clonal evolution.
Results
Across 225 primary and recurrent RRD-HGG, we identified 79 commonly mutated genes exhibiting distinctive characteristics. RRD-HGG harbored canonical adult glioma drivers such as IDH1 and PTEN, while lacking common pediatric drivers. In parallel, chromatin remodelling and DNA epigenetic modification pathways were overrepresented with genes such as DNMT3A, SETD2 and ARID1A/B, converging on stemness-associated gene expression. RAS/MAPK and AKT/MTOR pathway mutations were also enriched, enabling exploration of therapeutic vulnerabilities. Interestingly, mutational spectra of individual genes were not random, but determined by RRD-related mutational signatures and, thereby, constrained to specific recurrent amino acid changes which can be used for neoantigen-targeted vaccine-based therapy. Furthermore, these mutations recurrently emerge throughout tumor evolution, as observed by specific hotspot mutations in TP53 and IDH1 detected at low allele frequencies in 51.2% and 40.6% of both primary and relapse tumors, respectively, resulting from ongoing RRD signature-driven mutagenesis. Importantly, a third mechanism also shapes the mutational spectrum of RRD_HGG, as recurrent mutations were enriched for non-immunogenic neoantigens, highlighting the role of immune editing as a driver of tumor evolution in hypermutant cancers. Finally, for specific mutations, we estimated the contribution of biological driver strength, RRD mutational signature and immune evasion in determining their enrichment in gliomagenesis.
Conclusion
Taken together, our findings provide evidence for a coordinated tripartite mechanism where RRD-signature driven mutational processes provide the fundamental pool of available mutations upon which functional and immune selection jointly act to ultimately define the biologically unique oncogenic molecular driver landscape of pediatric RRD HGGs.