Hepatoblastoma is a rare but aggressive liver cancer that predominantly affects children under five years of age. While it is the most common liver cancer in this age group, its global incidence remains low - about 1-2 cases per million children. Despite recent advances in treatment, outcomes remain poor for the 20% of patients with metastatic or unresectable tumors. Thus understanding “why?” is crucial to saving these young lives. 

‍

The SignalWise project is a public-private partnership that unites DeepLife’s cutting-edge AI and the scientific excellence of the MIRCADE (Methods and Innovations for Research in Pediatric Cancers) team at Inserm (the French national institute of public health and research) to accelerate translational research for diseases with high unmet medical needs, such as hepatoblastoma.  

‍

At the project outset, Inserm Transfert highlighted the importance of this collaboration, emphasizing its focus on innovative ways to find novel drug combinations for patients - especially those facing relapse. One year on, together we harness generative AI applied to multi-omics single cell data to perform an unprecedented exhaustive molecular analysis of hepatoblastoma cells. This analysis revealed hidden cellular mechanisms and uncovered causal links between genes, pathways and phenotypes driving tumor behaviour.

‍

By combining DeepLife’s OmicStore and Cell BluePrint technologies with MIRCADE’s in vitro and in vivo models, we are creating digital twins of hepatic tumor cells that will enable scientists to pinpoint actionable biology. Our goal is to identify drug-pathway interactions and develop rational drug combinations that minimize toxicity and overcome resistance. 

‍

‍

Our approach starts with high-resolution digital reconstructions of hepatic tumor cells, built from integrated omics data generated through MIRCADE’s in vitro and in vivo experiments. Using our digital twin model, we have profiled tumor cells across a spectrum of mild to aggressive states and different treatment conditions to reveal mechanisms underlying drug resistance. By analyzing signaling pathway activation across these diverse cellular states, we have identified promising drug-pathway interactions through our cell-specific interactome-based analyses.

‍

‍

Next, we constructed a drug–pathway matrix to systematically identify mechanisms of action by linking compounds to their targeted signaling pathways based on overlap and significance. These insights are then mapped onto our cell-specific interactomes, enabling the generation of testable biological hypotheses. This strategy has already led to the identification of candidate inhibitors, including those that block resistance pathways, providing a foundation for exploring rational drug combinations to tackle resistant tumor populations.

‍

Christophe Grosset, Research Director at the Bordeaux Institute of Oncology and Team Lead for MIRCADE shared his perspective: “The MIRCADE team has a 13-year experience in the field of hepatoblastoma and our overall work led to major biological findings and preclinical advances. These include the stratification of tumors in three distinct subgroups identifiable by a concise four-gene signature (Hooks et al, Hepatology 2018), the demonstration that LHX2 protein plays a tumor suppressive role in adult and pediatric liver cancers (Mosca et al, Liver Cancer 2022), the design of a new neonatal model of hepatoblastoma in mice (Klein, Guillorit et al, Oncology 2025), and the therapeutic benefit of combining EZH2 inhibitor and statin to treat patients with high risk tumor (Khoubai et al, Molecular Cancer 2025, in press).”

‍

‍

“Therefore, our partnership with DeepLife company through the SignalWise project will shed more light on the molecular mechanisms causing hepatoblastoma and bring new therapeutic options to the patients, especially those presenting with an aggressive tumor or lung metastases”, Christophe said. “We are currently pursuing our research efforts in hepatoblastoma through innovative projects granted by the French National Institute of Cancer, with the aim at triggering immune response and at developing more personalized medicine using nanosystems loaded with therapeutic nucleic acids.”

‍

The initial computational insights from this research collaboration are now entering preclinical validation in zebrafish and juvenile mouse models to assess efficacy, target engagement, and histological response. Our roadmap spans from data-driven discovery (2024) to the construction of a full preclinical dossier by the end of 2026 and we are fully committed to transforming these breakthroughs into impactful therapies.

‍

‍

“Collaborations like SignalWise demonstrate the power of combining cutting-edge technology with scientific excellence. At DeepLife, we believe that partnering with leading academic teams like the MIRCADE team, led by Christophe Grosset and supported by Inserm Transfert, is essential to accelerate the development of impactful therapies for children with hepatoblastoma,” said Kevin Carvalho, Head of Partnerships, DeepLife.

‍

SignalWise is a pediatric oncology effort with high translational potential. We are proud to be a part of this collaborative effort combining AI and biology to improve outcomes for the thousands of young patients facing this disease each year in the world. The project team welcomes new partners to join us on this important journey to advance novel hepatoblastoma therapies, extend indications, and create new assets based on our discoveries. Together, let’s build the future of precision pediatric oncology.

‍