Pineoblastoma is an aggressive paediatric pineal parenchymal tumour (PPT) comprising different DNA methylation subgroups (miRNA1, miRNA2, RB1 and MYC/FOXR2), characterized by distinct molecular landscapes, genetic drivers, demographics and clinical outcome. In the mature pineal gland, a small endocrine gland in the centre of the brain, pinealocytes synthesize melatonin in a circadian manner. Pineoblastoma is thought to arise in the developing pineal gland; however, the molecular mechanisms and cell types that regulate pineal development remain poorly understood. This has limited the generation of adequate pineoblastoma models to study and improve treatment of this aggressive brain cancer.
Consistent with this, in genetically engineered mouse models (GEMMs) that allow conditional lineage-specific perturbations of pineoblastoma subgroup-specific drivers, pineal-region tumours formed only when drivers were induced during the peak developmental window for pinealocyte progenitors (E15–P10). Using this conditional lineage-specific perturbation system, the authors generated multiple pineoblastoma GEMMs that faithfully modelled human pineoblastoma subgroups (with the exception of the MYC model, which was more ambiguous). To understand pineoblastoma pathogenesis, they inferred gene regulatory networks from the Rb1 GEMM and identified a photoreceptor-like transcriptional network that drives tumour initiation. Interestingly, the same network was also present in retinoblastoma (RETB) and group 3 medulloblastoma (G3-MB).
