Molecular cartography of initiating and infiltrating glioma cells using multisector sampling and spatial transcriptomics

The National Institute of Neurological Disorders and Stroke awarded Translational Genomics Research Institute $305,960 on April 15, 2026, under the Extramural Research Programs in the Neurosciences and Neurological Disorders program (CFDA 93.853) to perform molecular cartography of glioma cell initiation and infiltration patterns using spatial transcriptomics.

The recipient will conduct spatial transcriptomics on 254 image-mapped samples from 22 distinct tumors to characterize diffusion patterns of glioma cells infiltrating brain tissue, testing the hypothesis that different tumor cell types (NPC-like, OPC-like, and APC-like cells) exhibit distinct dissemination patterns and microenvironment interactions within cortical and subcortical regions. The work will include tumor-microenvironment interaction analysis stratified by individual patient and across anatomical locations. The recipient will develop an agent-based model using 179 samples from 20 individuals with integrated imaging, DNA methylation, and DNA sequencing data to reconstruct historical tumor development and forecast future progression in space and time. The resulting four-dimensional model of tumor cell dissemination will support individualized local treatment planning and prognostication.

Performance occurs in Phoenix, Arizona, with a period of performance from April 15, 2026, through March 31, 2028. This is a Project Grant, an assistance type.

Abstract

Diffuse dissemination of tumor cells into the brain is one of the key problems impacting optimal treatment of gliomas. We hypothesize that distinct types of tumor cells harbor characteristic diffusion patterns. Specifically, we suspect that NPC-like cells predominate in the cortical layers and interact with the laminar extension of neuronal cells in the cortex, while OPC-like cells dominate subcortically and associate with axons and APC-like cells associate with vascular and immune cells. In order to provide four-dimensional reconstructions of tumor development and test our hypotheses, we will perform spatial transcriptomics on n = 254 image-mapped samples from n = 22 distinct tumors (aim 1). Tumor-microenvironment interactions will be evaluated separately for each patient and by comparing distinct locations across patients. Finally, we will develop an agent-based model using n = 179 samples from n = 20 individual with integrated imaging, DNA methylation and DNA sequencing information to reconstruct historical and forecast future tumor development in space and time (aim 2). The findings of these studies will allow for a high-resolution dissection of the identity and location of ancestral stem- like tumor cell populations, as well as the identification of sites of diffuse infiltration. A four-dimensional model of tumor cell dissemination will enable individualized local treatment planning and prognostication.