Immunogenomic profiling reveals targets for gene therapy in pediatric brain tumors
Abstract Background Pediatric brain tumors, particularly pontine diffuse midline glioma (pDMG), remains lethal with limited therapeutic options. Improved stereotactic biopsy techniques and advances in bioinformatics are progressively enabling deeper exploration of immunological vulnerabilities empowering novel strategies, including adoptive cell and gene therapies (ACGTs). We aimed to integrate genomic, transcriptomic, and immunological analyses to identify actionable pathways, surface antigens, and neoantigens that could inform next-generation immunotherapies, including Chimetic Antigen Receptor (CAR)-T cells and T cell Receptor (TCR)-T cell strategies. Methods Primary pDMG samples (n = 6) underwent whole-genome and RNA sequencing. Transcriptional drug response profiling was used to define targetable transcriptional dependencies. Surface antigen expression and immune cell composition were assessed to evaluate suitability for ACGTs. Neoantigen prediction employed PIOR, integrating somatic variant calling, Human Laukocyte Antigen (HLA) binding, and expression data. Prioritized neoantigens were synthesized and used to stimulate healthy donor T cells. Activated CD137+ T cells were sorted for bulk TCR sequencing to identify clonally expanded TCRs. Results Transcriptional drug response profiling revealed heterogeneous but actionable pathway dependencies. B4GALNT1 expression varied across tumors, identifying a subset with GD2 levels compatible with CAR-T targeting. Tumor microenvironment profiling showed enrichment of dendritic cells and M2 macrophages, with scarce CD8+ T cells and NK cells. Across samples, 31 somatic variants were identified including alterations in ACVR1, H3K27M, and TP53. Several predicted neoantigens induced T cell activation and clonal expansion. Conclusion This integrated profiling approach identifies targetable pathways, surface antigens, and neoantigens in pDMG, supporting the development of CAR-T and TCR-T therapies. These insights also suggest potential applicability to other cancers harboring shared mutations.