FY25 Spatial Transcriptomic Characterization of Pediatric Brain Tumors in Kentucky

Summary: Kentucky children suffer from a high cancer burden relative to other children in the United States. For the ten-year period ending in 2018, Kentucky ranks 7th highest in cancer incidence, with the 3rd highest rate of invasive brain and central nervous system tumors. Within the state, our published research has demonstrated that children living in the Appalachian and north central regions of the state experience between a 74% and 104% increased risk of developing a brain tumor. Pediatric brain tumors are characterized by high intratumoral and intertumoral heterogeneity. This heterogeneity is a leading cause of therapeutic resistance and treatment failure. In this project, we propose to comprehensively characterize intratumoral and intertumoral heterogeneity in a morphological context for pediatric brain tumors. We will perform spatial transcriptomic (ST) analysis on Formalin-Fixed Paraffin-Embedded (FFPE) tissue samples from 100 patients, where a true population-based sample will be obtained from treating facilities in and outside of Kentucky using the resources of the Kentucky Cancer Registry’s Virtual Tissue Repository. We will also develop new bioinformatics methods to enhance the study of tumor-associated cell type interactions. Within a tumor, our analysis will characterize the intratumoral heterogeneity at pathological, cellular, and molecular levels, and elucidate the interactions between tumor cells and their microenvironment. Across patients, our analysis will identify molecular subtypes of tumors based on ST profiles, and delineate unique ST patterns in the high-risk regions in Kentucky. We hypothesize that the large-scale spatial transcriptomic analysis coupled with advanced bioinformatics methods will offer a new dimension of insight on the mechanisms of pediatric brain tumor development, contribute to deciphering the molecular basis of high incidence rate in the Appalachian and north central regions, and provide critical information for developing new treatments for this devastating cancer. Significance: Transcriptomic and genomic studies have revolutionized the diagnosis and treatment of pediatric brain tumors. However, typical transcriptomic analyses, i.e. bulk and single-cell RNAseq, ignore the morphological information of tumor tissue, and thus are insufficient to fully characterize tumor heterogeneity. In recent years, a new technology, named spatial transcriptomic (ST) analysis, has been developed that allows transcriptomic profiling at close to single cell resolution with morphological context in individual tissue sections. As a representative of this technology, the Visium by 10x Genomics is able to simultaneously assess 5,000 spots within a user-specified 6.5mmx6.5mm capture area of a tissue section to obtain the gene expression profile of each of the spots. Visium has been used to probe heterogeneity for other solid tumor types. In this project, we will leverage the Visium technology and advanced bioinformatics methods to comprehensively study intratumoral and intertumoral heterogeneity of pediatric brain tumors within a morphological context. To our knowledge, our project will be the first population-based ST study for pediatric brain tumors in Kentucky and in the United States. It will substantially advance our understanding of the interaction between tumor cells and their microenvironment, identify unique spatial gene expression patterns in the high-risk regions in Kentucky, and provide invaluable information for the development of novel treatments. Specific Aims: In Aim 1, we will perform the Visium spatial gene expression profiling experiment on FFPE tissue specimens from 100 pediatric brain tumor patients in Kentucky. The specimens will be identified by the Kentucky Cancer Registry’s Virtual Tissue Repository. In consultation with neuro-pathologist Dr. Janna Neltner, tissue imaging, processing, and transcriptome library preparation will be performed by Dr. Douglas Harrison using the Visium FFPE v2 spatial assay. Transcriptomic profiling will be performed by Dr. Shulin Zhang using a NovaSeq 6000 sequencer. In Aim 2, we will investigate intratumoral and intertumoral heterogeneity based on the ST data in Aim 1 and using advanced bioinformatics methods. Within each tumor, we will estimate the cell type composition in each spot and cluster spots based on their expression profiles. We will then examine the consistency between the clusters and pathologist-annotated histological regions. We will also identify highly expressed genes and enriched biological pathways for each cluster/pathologist-annotated region such as tumor core or infiltrative tumor margin. In addition, we will study the co-localization and interaction of cell types. Across patients, we will identify molecular subtypes of pediatric brain tumors based on spatial gene expression profiles. We will also delineate unique spatial gene expression patterns between high and low risk regions of the state. In addition, we will integrate the ST data with bulk RNAseq and whole exome sequencing data we already obtained from a previous PCRTF-funded project. Results will be interpreted by physician scientist Dr. John Villano. In Aim 3, we will develop novel bioinformatics methods to enhance the analysis of ST data. Because the ST technology is new, many data analysis methods are still borrowed from those developed for bulk or single cell RNAseq, which do not adequately utilize the spatial information. We will use Bayesian statistical techniques to develop new bioinformatics methods to more accurately and efficiently perform spot clustering and cell type co-localization and interaction analyses. Those methods will not only enhance the analysis of this project but also are broadly applicable to many other ST studies. Study results and sequencing data will be shared with the ACCELERATE Consortium for broader research use. Budget: The budget will be no more than the recommended $250,000 per year for two years.