ML study identifies new glioblastoma treatment strategy
A new precision medicine approach that combines genetic and functional profiling could predict how glioblastoma will respond to therapy.
A new study from scientists at the UCLA Health Jonsson Comprehensive Cancer Center helps explain why glioblastoma becomes resistant to treatment and introduces a new approach that paves the way for more personalised treatment strategies.
The approach, described in Nature Communications, combines genetic profiling with functional profiling to help predict how well glioblastoma will respond to treatments.
“Many cancer treatments are based on the genetic profile of a patient’s tumour. However, genomic features alone can’t always predict how a tumour will respond to a therapy,” said Dr David Nathanson, senior author of the study and Professor of Molecular and Medical Pharmacology at the David Geffen School of Medicine at UCLA. “This study explores a new approach that looks beyond the tumour’s genetic blueprint, in which we combine genetic data with functional tests to show how live cancer cells respond to treatments. This gives us a much clearer picture of what treatments will work and why.”
Glioblastoma is notoriously difficult to treat due to its ability to resist apoptosis and its rapid adaptation to therapies.
While traditional genomic precision medicine uses DNA sequencing to identify genetic mutations in tumours, it doesn’t account for the genetic complexity of brain tumours or the dynamic behaviour of cancer cells.
Machine learning tool
To overcome to this challenge, Nathanson and his team used BH3 profiling to measure the cancer cells’ responses to treatments that aim to trigger cell death in real time.
Through this analysis, the team found that standard therapies like radiation or chemotherapy can change how the tumour’s self-destruct machinery works, but this effect depends on specific genetic features, such as a functional p53 gene.
The researchers then created a machine learning (ML) tool called GAVA, which combines genetic and functional data to help predict how a glioblastoma tumour might respond to a specific treatment combination.
In preclinical models, GAVA helped predict which tumours would respond best to a combination of standard cancer treatments and drugs that block certain proteins. They found that targeting the protein BCL-XL, which plays a major role in helping cancer cells avoid death, could enhance the effectiveness of treatments in certain cases.
The researchers then tested a new experimental drug called ABBV-155, an antibody-drug conjugate that is designed to target BCL-XL in tumours while sparing healthy cells.
“We found that combining standard therapies with ABBV-155 successfully induced tumour shrinkage, which is something we rarely observe in clinically relevant glioblastoma models,” said Nathanson. “The results are incredibly exciting and we are hopeful that this approach will pave the way for a new therapy for patients with this devastating disease.”
The researchers are now working to test the combination in a clinical trial to evaluate its impact on patients.
Diana Spencer, Senior Digital Content Editor, DDW
