Mosaic Analysis with Double Markers Reveals Tumor Cell of Origin in Glioma

Cancer results from molecular mutations occurring in specific cell types, thus determining the cell-of-origin is critical for effective cancer treatment. Inferring such information from terminal tumors can be misleading because malignant tumor cells tend to acquire aberrant properties. Animal models are widely used because one can initiate mutations in specific cell types. However, cell-of-mutation that harbors initial molecular changes may not directly transform, rather merely passes along mutations to its progeny cell lineage, which then serves as cell-of-origin and finally transforms into malignancy. Such a problem is exemplified in the glioma field: glioma can be induced either by mutating tumor suppressor genes in neural stem cells (NSCs) or by over-expressing oncogenes in restricted progenitor cells such as oligodendrocyte precursor cells (OPCs). However, for the NSC glioma model, it remains unknown whether NSCs directly transform or simply pass along mutations to OPC lineage for malignancy. Here we use a genetic system termed Mosaic Analysis with Double Markers (MADM) to study the earliest stage of gliomagenesis from mutated NSCs. At an in vivo single-cell resolution, we unbiasedly analyzed tumorigenic potential of NSCs and all cell lineages derived from them. At pathologically undetectable pre-transforming phases, we found no significant aberrant growth of mutant NSCs and their derivatives except for the OPC lineage. We then tracked the tumor progression from earliest stage and confirmed the expansion of OPCs lead to gliomagenesis. Consistently, transcriptome profiling reveals that terminal-stage tumor cells display salient OPC features and resemble human proneural subtype of glioblastoma multiform (GBM). Most importantly, introducing the same mutations directly into OPCs was sufficient for malignant transformation and these tumor cells are indistinguishable in their gene expression profiles from those in which the first mutations were introduced into NSCs. Our findings strongly implicate OPCs as the transforming cell type for glioma even when initial mutations could occur in NSCs, and highlight the importance of analyzing early phases of tumorigenesis to distinguish cell-of-origin from cell-of-mutation.