While genetically engineered mouse models (GEMMs) represent valuable tools to study diseases, human pluripotent stem cell (hPSC)-based models offer several advantages over GEMMs including:
(1) Species identity with human patient tumors.
(2) Robust application to more quickly evaluate candidate genetic drivers.
(3) Synteny of chromosome segments enables modeling of chromosome copy number abnormalities.
Here, we describe hPSC-based models of neuroblastoma (NB). Normal hPSCs were differentiated toward trunk neural crest cells (NCC)1. We next introduced established/candidate drivers of NB, and implanted the resulting trunk NCCs orthotopically into renal capsules of mice. As proof-of-principle, we transduced trunk NCC with MYCN and/or mutant anaplastic lymphoma kinase (ALK F1174L), NB drivers described in human patients and demonstrated in GEMMs and zebrafish 2-11. Within 3 months, 60% of mice developed tumors with ALK F1174L/MYCN, 10% with MYCN, and none with either empty vector or ALK F1174L alone. Tumors were transplantable and demonstrated small round blue cell histology characteristic of poorly differentiated tumors12. Further analysis revealed expression of markers typically found in NB, while lacking markers of the small round blue tumors rhabdomyosarcoma, lymphoma and Ewing sarcoma. We are currently evaluating candidate predisposition genetic events that may cooperate with MYCN, including loss of lncRNAs CASC14, CASC15, and deletion of chromosome 1p 13-16. Our preliminary results suggest loss of the lncRNA or chromosome 1p combined with misexpression of MYCN promotes growth in vitro and/or in vivo. Thus, we present a hPSC-based model of NB that is useful to validate novel candidate genetic drivers and to evaluate personalized therapies.