Abstract 3901: Modeling and deciphering COJEC resistance in high risk neuroblastoma

Abstract Neuroblastoma (NB) is a pediatric cancer of the sympathetic nervous system that accounts for 15% of childhood cancer deaths. High-risk (HR) NB tumors are treated with an aggressive multimodal therapy, including the COJEC induction protocol which is a combination of five different chemotherapeutic agents. However, around 15% of primary tumors are refractory from the start and over 60% present treatment resistant relapses. In this project we aim to decipher the mechanisms involved in treatment resistance in HR-NB. Using various NB patient-derived xenograft (PDX) models we designed and tested a clinically-relevant protocol that mimics COJEC treatment and surgical resection. RNA-sequencing, SNP analysis, and histology, were used to explore the molecular mechanisms involved in treatment response. With our COJEC-like protocol we were able to replicate the original patient response in our PDX models. NB PDX1 presented upfront progression and a lack of morphological differentiation. At a molecular level it was characterized by genome duplication and high number of chromosomal aberrations, as well as a high expression of early development/mesenchymal and cell cycle genes. In contrast, NB PDX3 presented an overall good response, leading to a 22% relapse rate after surgery. Relapsed tumors displayed high expression of early development/mesenchymal genes, similar to those seen in refractory PDX1. Meanwhile, the PDX3 tumors collected during surgery that did not relapse (cured) showed low MYCN expression, downregulation of cell cycle and DNA repair genes, along with a strong upregulation of nervous system/nor-adrenergic genes. These genetic signatures derived from the cured PDX3 samples also correlated with good prognosis in large cohorts of HR-NB patients. Despite observing a convergent/parallel evolution of specific genetic aberrations across our PDXs, their presence or absence does not predict prognosis. Instead, the high RNA expression levels of those specific genes correlates with treatment response both in our PDX models and in patient cohorts. Finally, we established 3D tumor organoids derived from treated PDX tumors and confirmed they maintain the molecular characteristics of the parental tumor as well as the acquired COJEC resistance and the tumorigenic capacity, making them an excellent tool to test novel therapeutics against HR-NB. In conclusion, we designed a clinically-relevant COJEC-like protocol that can mimic patient response in PDX models. Early developmental gene signatures appear to correlate with treatment resistance in HR-NB, while the expression of nor-adrenergic signatures after chemotherapy are associated with good prognosis, independently of the accumulation of copy number aberrations. Based on our results, we propose that a combination of DNA and RNA analysis at the surgical resection time point after COJEC could be of clinical interest to evaluate COJEC treatment response and predict future relapses. Citation Format: Adriana Mañas, Kristina Aaltonen, Natalie Andersson, Karin Hansson, Alexandra Seger, Hiroaki Yasui, Katarzyna Radke, Javanshir Esfandyari, David Lindgren, David Gisselsson, Daniel Bexell. Modeling and deciphering COJEC resistance in high risk neuroblastoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 3901.