Effects of COJEC induction on neuroblastoma patient-derived xenografts (PDX).

Abstract

e21503 Background: Neuroblastoma (NB) is a solid pediatric tumor that causes 15% of childhood cancer deaths. High-risk patients have 50 to 70% mortality after 5 years. Currently, the standard way to treat high risk patients is COJEC induction followed by surgery, chemotherapy with stem cell transplants, irradiation and immunotherapy. This intense treatment has significantly increased overall survival, but mortality is still high, survivors are prone to relapse and treatment resistance is a major clinical problem. COJEC induction consists in the cycled administration of the chemotherapeutics Cisplatin, Vincristine, Etoposide, Cyclophosphamide and Carboplatin. These drugs have been long used in the clinic, but a comprehensive preclinical study combining all of them had not been done. The aim of this project was to establish a PDX model for chemotherapy induction in high-risk NB and investigate tumor characteristics and response upon treatment. Methods: NB cells from previously developed high-risk PDX models were injected subcutaneously in nude mice. When tumors reached an appropriate size, they were treated with either a COJEC-like protocol or single agent induction with high dose cisplatin. After treatment, tumors were analyzed by immunohistochemistry, chromosomal copy number analysis and RNA sequencing. Results: Most mice responded well to COJEC induction and 8/10 tumors were reduced. Within the reduced tumors, 4 were subjected to total resection and relapsed after several weeks, mimicking the clinical situation. Cisplatin treatment alone led to no significant tumor reduction and 6/10 tumors were highly resistant. Thus, as for humans, multidrug induction is more successful in reducing tumor size but presents an uneven response with high relapse rate. The responsive tumors show a higher degree of cellular differentiation when compared to resistant tumors. Cisplatin treated tumors have accumulation of chromosomal aberrations when compared to controls, and subclonal dynamics are enhanced in resistant tumors. Conclusions: We established a PDX model for chemotherapy induction in high-risk NB. The results mimic clinical findings seen in high-risk NB patients. The model is useful to understand the mechanisms of treatment resistance and for the testing of novel therapies against high-risk NB.