In an innovative new study on glioblastoma, scientists used artificial intelligence (AI) to reprogram cancer cells, converting them into dendritic cells (DCs), capable of identifying cancer cells and commanding other immune cells to kill them.
Glioblastoma is the most common brain cancer in adults and also the deadliest, with less than 10% of patients surviving five years after their diagnosis. Although new approaches such as immunotherapy have revolutionized the treatment of other cancers, they have done little for glioblastoma patients. This is partly because these hard-to-reach brain tumors hide behind the blood-brain barrier, where immune cells have difficulty reaching and eliminating them.
But new research, supported in part by the National Institutes of Health and led by the Keck School of Medicine of USC, has harnessed AI to explore which genes control a cell’s fate, whether it develops into a cell heart, lung cells or cancer. cell, for example. Researchers have identified genes that can reprogram glioblastoma cells, converting them into immune cells within the tumor itself so that they effectively target related cancer cells for destruction.
In mouse models of glioblastoma, this approach increased the chance of survival by up to 75%. The results have just been published in Cancer immunology researcha journal of the American Association for Cancer Research.
“This groundbreaking study harnesses the power of AI to transform glioblastoma cells into immune-activating cells, marking a significant advance in cancer immunotherapy,” said the study’s lead author. David Tran, MD, PhDassociate professor of neurological surgery and neurology and head of division of neuro-oncology at Keck School of Medicine. “By turning the cancer’s own cells against it, we are paving the way for more effective treatments and offering new hope to patients fighting this and many other aggressive cancers. »
In addition to their work in animal models, the researchers used AI to identify a set of genes capable of converting human glioblastoma cells into immune cells. In the future, scientists could deliver this genetic material to glioblastoma patients by integrating it into a harmless virus, a tool known as a viral vector.
“Forcing a cell with 20,000 genes to become something else is incredibly complex. Using traditional molecular approaches, this would be almost impossible to achieve,” said Tran, who also co-directs the USC Brain Tumor Center has USC Norris Comprehensive Cancer Center. “AI helps us answer some crucial questions and gives us a powerful way to learn how to manipulate the fate of a cell.”
Controlling the fate of a cell
DCs play a central role in activating the immune response: they take up antigens (for example from a cancer cell) and present them to other immune cells, including armies of T cells, thereby triggering a large-scale attack.
Evidence suggests that DCs can fight glioblastoma, but scientists have yet to find a reliable way to get them past the blood-brain barrier and into the hostile environment of a tumor. By reprogramming the cancer cells already located within the tumor, Tran and his team circumvented this major challenge.
When manipulating the fate of a cell, an important consideration is specificity. Converting healthy brain cells into DCs, for example, could shrink a brain tumor but cause health problems.
“We don’t want to give a patient something that can convert all kinds of cells into CDs,” Tran said.
The machine learning system he and his team developed was able to interrogate tens of thousands of genes and millions of gene-to-gene connections, identify those that can specifically target glioblastoma cells, and reprogram them to look like CDs. This AI-based method differs from previous research, which used what is known as an empirical approach to manually identify genes that control a cell’s fate.
“The high computing power of AI really helps us speed up the discovery process,” Tran said.
Used alongside other immunotherapies, reprogramming glioblastoma cells has significantly improved immune response and survival rates in mouse models of glioblastoma. Combined with immune checkpoint therapy, the new approach improved the chances of survival by 75%. Combined with a conventional DC vaccine, the new approach doubled the chances of survival. (Used alone, neither immune checkpoint therapy nor a DC vaccine increases the chances of survival of patients with glioblastoma.)
Towards clinical trials
In addition to the proof-of-concept study in mice, the researchers used their AI system to identify a set of human genes capable of converting human glioblastoma cells into DC-like cells. Next, they plan to refine this list, package the genetic material into a viral vector, and launch an initial round of safety and effectiveness testing in animal models.
“We want to expand the research, using AI to help us find the best possible combinations as we move toward testing in patients,” Tran said.
If the approach is deemed safe and effective, meaning it improves outcomes in glioblastoma models and does not cause unexpected side effects, Tran and his team will seek approval to begin clinical trials on patients in several years. Eventually, they also hope to use their AI model to find genes that can reprogram other types of cancer cells to behave like DCs.
About this research
Besides Tran, other authors of the study are Tianyi Liu, Son B. Le and Dongjiang Chen of the Division of Neuro-Oncology, Departments of Neurological Surgery and Neurology and the USC Brain Tumor Center, Keck School of Medicine of the USC, University of Southern California; and Dan Jin, Mathew Sebastian, Alberto Riva, and Ruixuan Liu of the University of Florida College of Medicine, Gainesville, Florida.
This work was supported by the National Cancer Institute of the National Institutes of Health (R42CA228875, R01CA238387, P30CA014089, and F30CA232641) and the Bankhead Coley Research Program of the Florida Department of Health (6BC04).
Disclosure: Tran and Le are the inventors of two patent applications (62/952,725 and 62/586,655) based on data in this manuscript.