Using machine learning and existing medicines, scientists at Lund University have found a drug combination that slowed tumor growth and boosted survival in models of high-risk neuroblastoma, one of the deadliest childhood cancers. The work could pave the way for faster, more affordable treatments.
Using artificial intelligence and a library of existing medicines, researchers in Sweden have identified a drug combination that shows strong promise against one of the deadliest childhood cancers.
The team at Lund University reports that pairing a common cholesterol-lowering statin with a phenothiazine, a drug used for conditions such as migraine and nausea, slowed tumor growth and improved survival in experimental models of high-risk neuroblastoma. The study is published in the journal EMBO Molecular Medicine.
Neuroblastoma is a cancer of immature nerve cells that affects about 5,500 children worldwide each year, most of them under age 5. While some forms are relatively mild, the aggressive, high-risk type has the lowest survival rate of all childhood cancers and is notoriously difficult to treat.
This cancer often responds to chemotherapy at first, only to return in a more dangerous form.
“This disease has a high relapse risk. Chemotherapy drugs often work well initially, but when the cancer comes back, it’s resistant,” Daniel Bexell, head of the Molecular Pediatric Oncology research team at Lund University, said in a news release.
To tackle that problem, Bexell’s group turned to drug repurposing, an approach that looks for new uses for medicines that are already approved for other conditions. Repurposing can dramatically speed up the path to patients because the safety of the drugs is already well understood.
In the new study, the Lund team joined forces with the British AI and biotech company Healx, researchers at Karolinska Institutet, and two childhood cancer charities: the aPODD Foundation in the United Kingdom and ENEA (European Neuroblastoma Association) in Italy.
Using machine learning, the collaborators analyzed large datasets describing how thousands of drugs work in the body, alongside detailed information about genes that are especially important in neuroblastoma. The goal was to predict which existing drugs might disrupt the cancer’s biology.
From these computer-generated shortlists, the Lund researchers then moved into the lab. They tested promising candidates on aggressive neuroblastoma tumors derived from patients, first as single drugs and then in combinations.
“We quickly noted that two types of drugs could have an effect individually. But together the effect was very powerful – a strong synergy,” added first author Katarzyna Radke, who was a doctoral student in the research team.
The standout pair was a statin and a phenothiazine. Statins are widely prescribed to lower blood cholesterol and reduce the risk of heart disease. Phenothiazines are an older class of drugs used in various conditions, including to treat nausea and migraine.
It was already known that statins block the body’s ability to make new cholesterol. The Lund team discovered that the phenothiazine also reduced cholesterol in tumor cells, but through a different, complementary mechanism. When combined, the two drugs drove cholesterol levels in cancer cells down to a critical threshold.
“Together, this led to low cholesterol levels in the tumour cell. Many tumour cells died, and those that survived were sensitive to chemotherapy drugs. We knew that cholesterol was important for the tumour, but were surprised that it had such strong effects,” Bexell added.
Cholesterol is essential for building cell membranes and supporting rapid cell growth, both of which cancer cells rely on. By cutting off this supply from two directions at once, the drug duo appeared to weaken the tumor cells and restore their vulnerability to chemotherapy.
In laboratory experiments and animal studies, the combination treatment slowed tumor growth and improved survival in mice with aggressive, chemoresistant neuroblastoma. The findings suggest that targeting cholesterol metabolism could be a powerful way to overcome resistance in this cancer.
The work is still at an early, preclinical stage. The drugs used in the study were not given to children, and more research is needed before any clinical trials can begin. The team now plans to refine the chemical properties of the two medicines to optimize how they work together against neuroblastoma while remaining safe.
However, the results point to a hopeful path forward for families facing a diagnosis that currently carries a grim prognosis. Because both drugs are already on the market for other uses, successful repurposing could shorten the time and reduce the cost needed to bring a new treatment option to patients.
The study also highlights how machine learning and big data can accelerate discoveries in pediatric oncology, a field where traditional drug development has often lagged behind adult cancers.
By combining AI-driven predictions with careful laboratory testing on patient-derived tumors, the Lund team and its partners have created a model for finding new therapies in existing medicine cabinets.
If future work confirms the safety and effectiveness of this statin–phenothiazine strategy in children, it could add a much-needed weapon against high-risk neuroblastoma and offer renewed hope to young patients whose cancers no longer respond to standard chemotherapy.
Source: Lund University