IMMU-17. Comprehensive immunological gene expression profiling of pediatric brain tumors

Abstract

Immunotherapy, predominantly through immune checkpoint inhibition (ICI), has had incredible success in treating some metastatic cancers, however, outside of rare cases of mismatch repair deficient (MMRD) gliomas, brain tumors have not had consistent responses to ICI. This can be attributed to a variety of factors including a low tumor mutation burden, lack of T cell infiltrates, and the CNS immune privilege. There are numerous strategies to target the tumor immune microenvironment (TIME) beyond ICI, include CAR-T cells, tumor vaccines, and myeloid cell modulation. The investigation of these depends critically on detailed characterization of the cell populations and interactions in the CNS TIME. We developed a 103 gene NanoString immune-oncology gene expression panel that includes markers reflecting selected cell types, therapeutic targets, and cellular pathways, as well as the 18-gene Tumor Inflammation Signature, a well validate biomarker for ICI response. We have used this to characterize over 500 brain tumors, including a diverse set of 227 pediatric low-grade gliomas (LGG), 86 MMRD gliomas, 47 diffuse intrinsic pontine gliomas (DIPG), 26 ependymomas, 36 medulloblastomas, 70 adult gliomas, and 35 non-tumor brain samples. Our results demonstrate a broad range of immunologic states, including within groups of tumors with the same genetic driver alteration. In pediatric LGG with BRAF V600E, there was clear histologic correlation with immune status, as glioneuronal tumors had substantial upregulation of T cell markers and regulatory genes, while diffuse astrocytomas had a near normal immune profile. In DIPG there was strong upregulation of macrophage markers, contradicting prior reports that have characterized these tumors as immunologically neutral. In a set of MMRD gliomas treated with ICI we identified several differentially expressed genes correlating with therapeutic response, including CCL4, CXCL9, and HGPD. In sum, this provides a characterization of diverse immune activation states across pediatric gliomas and other brain tumors.