Modeling Patient-Derived Glioblastoma with Cerebral Organoids.

The prognosis of patients with glioblastoma (GBM) remains dismal, with a median survival of approximately 15 months. Current preclinical GBM models are limited by the lack of a "normal" human microenvironment and the inability of many tumor cell lines to accurately reproduce GBM biology. To address these limitations, we have established a model system whereby we can retro-engineer patient-specific GBMs using patient-derived glioma stem cells (GSCs) and human embryonic stem cell (hESC)-derived cerebral organoids. Our cerebral organoid glioma (GLICO) model shows that GSCs home toward the human cerebral organoid and deeply invade and proliferate within the host tissue, forming tumors that closely phenocopy patient GBMs. Furthermore, cerebral organoid tumors form rapidly and are supported by an interconnected network of tumor microtubes that aids in the invasion of normal host tissue. Our GLICO model provides a system for modeling primary human GBM ex vivo and for high-throughput drug screening.

Detailed longitudinal sampling of glioma stem cells in situ reveals Chr7 gain and Chr10 loss as repeated events in primary tumor formation and recurrence.

Intratumoral heterogeneity at the genetic, epigenetic, transcriptomic, and morphologic levels is a commonly observed phenomenon in many aggressive cancer types. Clonal evolution during tumor formation and in response to therapeutic intervention can be predicted utilizing reverse engineering approaches on detailed genomic snapshots of heterogeneous patient tumor samples. In this study, we developed an extensive dataset for a GBM case via the generation of polyclonal and monoclonal glioma stem cell lines from initial diagnosis, and from multiple sections of distant tumor locations of the deceased patient's brain following tumor recurrence. Our analyses revealed the tissue-wide expansion of a new clone in the recurrent tumor and chromosome 7 gain and chromosome 10 loss as repeated genomic events in primary and recurrent disease. Moreover, chromosome 7 gain and chromosome 10 loss produced similar alterations in mRNA expression profiles in primary and recurrent tumors despite possessing other highly heterogeneous and divergent genomic alterations between the tumors. We identified ETV1 and CDK6 as putative candidate genes, and NFKB (complex), IL1B, IL6, Akt and VEGF as potential signaling regulators, as potentially central downstream effectors of chr7 gain and chr10 loss. Finally, the differences caused by the transcriptomic shift following gain of chromosome 7 and loss of chromosome 10 were consistent with those generally seen in GBM samples compared to normal brain in large-scale patient-tumor data sets.