Genetic Characterization of Canine Brain Tumors

National Center For Advancing Translational Sciences

Investigators

Linked publications & trials

Abstract

Previously we ran bioinformatic analysis on RNA-SEQ data from canine tumor samples associated with glioma, choroid plexus, brain histiocytic sarcoma, non-brain histiocytic sarcoma, and meningioma. Analysis identified 12 gene co-expression modules, each differentially activated for specific cancer types. Biological pathways associated with the genes of each module were determined. The highest expressing genes from each module were chosen as biomarker candidates and therapeutic targets for further study. This year we generated RNA-SEQ data on an additional set of 116 meningioma samples spanning 10 histological types: Grade 1 (Meningothelial, transitional, fibrous, angiomatous, psamommatous, syncytial, microcystic), Grade 2 (Chordoid, atypical), and Grade 3 (anaplastic). Using bioinformatic analysis, gene expression values were determined for each meningioma sample. The samples were then clustered according to similarity of gene expression. The clustering by gene expression was, for the most part, consistent with clustering according to histological subtype. However, there were some exceptions. For example, one sample cluster based on gene expression, grouped the following histological subtypes together: Meningioma Psammomatous, Meningioma Synsytial, Meningioma Transitional, and Meningioma Meningothelial. Upon further investigation of these samples based on clustering by genes, it was found that they exhibited one common theme, that is the presence of psammoma bodies. As such, the histology assignment for these samples was changed to psam, indicating the presence of psammoma bodies. This and other inconsistencies has prompted the team to review the histology for all samples, in the context of the clusters based on gene expression. This review of the histological assignments is important since the goal is to find targets and therapies relevant to the various sub-types of meningioma and, as such, gene targets should be appropriately assigned to subtype. This work is currently ongoing. Similar data on matched human cancers will also be analyzed, identifying common genes between canine and human samples with respect to the various meningioma subtypes. Future directions include tissue and blood proteomics studies of these common genes. Tissue-expressed proteins may provide actionable targets to drive drug discovery campaigns, while proteins in the blood may represent potential non-invasive diagnostic biomarkers for cancer detection. The eventual goal is to test drug candidates in canines and if efficacious, to test the drugs in humans. Since up-front work is done to ensure targets are common to canines and humans with respect to disease, drug studies in humans are more likely to be successful.