Abstract
Despite advances in the molecular exploration of paediatric cancers, approximately 50% of children with high-risk neuroblastoma lack effective treatment. To identify therapeutic options for this group of high-risk patients, we combine predictive data mining with experimental evaluation in patient-derived xenograft cells. Our proposed algorithm, TargetTranslator, integrates data from tumour biobanks, pharmacological databases, and cellular networks to predict how targeted interventions affect mRNA signatures associated with high patient risk or disease processes. We find more than 80 targets to be associated with neuroblastoma risk and differentiation signatures. Selected targets are evaluated in cell lines derived from high-risk patients to demonstrate reversal of risk signatures and malignant phenotypes. Using neuroblastoma xenograft models, we establish CNR2 and MAPK8 as promising candidates for the treatment of high-risk neuroblastoma. We expect that our method, available as a public tool (targettranslator.org), will enhance and expedite the discovery of risk-associated targets for paediatric and adult cancers.
Introduction
Neuroblastoma is a pediatric cancer of the sympathetic nervous system. While some cases regress spontaneously, "high-risk" cases are notoriously difficult to treat, with survival rates often falling below 50%. The primary challenge is that standard DNA sequencing often fails to reveal "druggable" mutations in these patients. To bridge this gap, researchers have turned toward mRNA signatures. Instead of looking for a single broken gene, they focus on the overall "transcriptional state" of high-risk tumors, seeking drugs that can flip the switch from a malignant state back to a lower-risk or more differentiated state.
Methodology: The TargetTranslator Pipeline
The study introduced TargetTranslator, a multi-step integrative framework designed to bridge the gap between patient omics and therapeutic discovery.
The pipeline operates through four main stages:
- Signature Estimation: mRNA signatures associated with 16 clinical and genetic risk factors (e.g., MYCN amplification) are estimated using linear least squares.
- Drug Matching: The LINCS/L1000 database is queried to identify compounds that induce gene expression changes opposite to high-risk signatures.
- Target Deconvolution: By cross-referencing these drugs with the STITCH5 protein-interaction network, specific protein targets enriched among top-performing drugs are identified.
- Experimental Validation: Predicted drugs are verified in patient-derived xenograft (PDX) cells using SMART-Seq2 RNA sequencing and tested in vivo using zebrafish and mouse models.
Key Results and Performance
TargetTranslator identified 88 statistically significant protein targets associated with neuroblastoma risk. A formal ROC analysis demonstrated the algorithm's high predictive accuracy, achieving an AUC of 0.916 in predicting drug-induced signature modulation.
Two specific targets emerged as particularly promising candidates:
- CNR2 (Cannabinoid Receptor 2): Activating CNR2 was found to downregulate N-Myc protein levels, a major driver of neuroblastoma aggression. In mouse models, the CNR2 agonist GW405833 significantly impaired tumor growth.
- MAPK8 (JNK1): Inhibiting MAPK8 suppressed the malignant phenotype in high-risk cells. In zebrafish models, the MAPK8 inhibitor AS601245 reduced tumor size to 21% of controls.
Discussion and Future Directions
The study demonstrates that clinical risk can be treated as a targetable molecular state. By focusing on reversing malignancy rather than simple cytotoxicity, the approach identifies treatments that force cells to lose their high-risk identity.
However, limitations remain. Signature consistency varies across different risk factors; for instance, MYCN signatures are highly stable, while 11q deletion signatures are less so. Furthermore, the exact molecular mechanism by which CNR2 activation leads to N-Myc downregulation requires further dissection. Future work will involve expanding the analysis beyond the L1000 library to full-transcriptome drug response data.
The TargetTranslator method is now available as a public tool at targettranslator.org.
Summary updated