Neuroblastoma is a commonly lethal solid tumor of childhood and intensive chemoradiotherapy treatment cures ~50% of children with high-risk disease. The addition of immunotherapy using dinutuximab, a monoclonal antibody directed against the GD2 disialoganglioside expressed on neuroblasts, improves survival when incorporated into front-line therapy and shows robust activity in regressing relapsed disease when combined with chemotherapy. Still, many children succumb to neuroblastoma progression despite receiving dinutuximab-based immunotherapy, and efforts to counteract the immune suppressive signals responsible are warranted. Animal models of human cancers provide useful platforms to study immunotherapies. transgenic mice are immunocompetent and develop neuroblastomas at autochthonous sites due to enforced expression in developing neural crest tissues. However, GD2-directed immunotherapy in this model has been underutilized due to the prevailing notion that neuroblasts express insufficient GD2 to be targeted. We demonstrate that neuroblasts in -driven tumors express GD2 at levels comparable to human neuroblastomas but rapidly lose GD2 expression when explanted ex vivo to establish tumor cell lines. This occurs in association with a transition from an adrenergic to mesenchymal differentiation state. Importantly, not only is GD2 expression retained on tumors in situ, treatment with a murine anti-GD2 antibody, 14G2a, markedly extends survival in such mice, including durable complete responses. Tumors in 14G2a-treated mice have fewer macrophage and myeloid-derived suppressor cells in their tumor microenvironment. Our findings support the utility of this model to inform immunotherapy approaches for neuroblastoma and potential opportunities to investigate drivers of adrenergic to mesenchymal fate decisions.