Liquid biopsy approaches to inform neuroblastoma prognosis and disease monitoring

Abstract Fewer than half of all children with high-risk neuroblastoma become long-term survivors. Currently, it is not possible to predict if a child will be cured with standard therapy or is destined to relapse. Furthermore, standard clinical evaluations lack sensitivity to detect minimal residual disease (MRD) that ultimately leads to recurrence. Thus, there is a critical challenge and an unmet need to develop new precision biomarkers to identify patients who will ultimately have a poor response to the high-intensity therapy and may benefit from alternate approaches. We will develop new biomarkers to guide treatment decisions using cell-free DNA (cfDNA) and a novel, epigenetic-based methodology that will identify underlying biology driving aggressive neuroblastoma. In many cancer types, analysis of cfDNA isolated from peripheral blood has shown promise, revealing biomarkers for diagnosis, prognostication, and tumor surveillance. Cytosines in DNA can either be unmodified, methylated (5- methylcytosine, 5mC), or contain an oxidized form of 5mC, 5-hydroxymethylcytosine (5hmC). Unlike 5mC, elevated 5hmC deposition across a gene body marks active transcription. In this proposal, we will use nano- hmC-seal, a whole-genome methodology for analyzing 5hmC modifications in cfDNA. Recently, we evaluated 5hmC in cfDNA collected serially from children with neuroblastoma and demonstrated that 5hmC profiles correlated with disease burden and patient outcome. Importantly, we also found a cfDNA 5hmC derived biomarker can distinguish patients with superior response to treatment from those at high risk for relapse. 5hmC profiles from cfDNA compared to diagnostic high-risk primary tumors demonstrated cfDNA is derived from clinically aggressive, malignant cells with activation of networks common in relapsed tumors. To prospectively determine the prognostic strength of 5hmC-based cfDNA biomarkers, we will use nano-hmC-seal to generate 5hmC profiles from clinically annotated serial blood samples (liquid biopsies) collected from 400 patients enrolled on the ongoing Children’s Oncology Group High-Risk Neuroblastoma Phase III study (ANBL1531, NCT03126916). We hypothesize that cfDNA 5hmC profiles from children with neuroblastoma will serve as superior biomarkers for response and survival compared to current clinical methods and will reveal transcriptional networks driving relapse. The specific aims are: 1) Evolve and validate biomarkers of poor response at diagnosis; 2) Prospectively identify minimal residual disease (MRD) and predict relapse from serial cfDNA 5hmC profiles; 3) Experimentally confirm candidate networks enriched in cfDNA at relapse. The success of this proposal will lead to: 1) unprecedented diagnostic biomarkers to improve therapeutic decisions; 2) early detection and interventions for patients with relapse causing MRD; 3) identification of epigenetic mechanisms which drive relapse. This work will have a transformative impact by to identifying patients who benefit from early introduction of alternate therapy, improving outcomes for those with aggressive disease.