Machine-Learning Model Can Help Identify Ovarian Cancer Treatment Targets

A study published last month in Nature Metabolism shows that a machine learning (ML)-based computational platform can identify specific metabolic targets in ovarian cancer, which could be used in personalized treatment therapies.

According to the Centers for Disease Control and Prevention (CDC), ovarian cancer is the second most common gynecologic cancer in the US and causes more deaths than any other cancer of the female reproductive system.

Because of this, many researchers investigating treatment targets and personalized therapies for ovarian cancer are turning to cutting-edge approaches like precision medicine and genomics. In this study, researchers used an interdisciplinary approach to develop a machine-learning model to identify metabolic vulnerabilities within certain genes that can impact cancer growth.

According to the press release, these vulnerabilities can occur as a result of genetic alterations that happen in tandem with cancer growth. Cancer mutations are frequent in ovarian cancer cases. These mutations contribute to the aggressiveness of the cancer by giving mutated cells a growth advantage. However, certain genes may get deleted alongside these mutations, which can result in cells that are more vulnerable to treatment.

“When a gene is deleted, metabolic genes, which allow the cancer cells to grow, are also deleted,” explained Deepak Nagrath, PhD, associate professor of biomedical engineering at the University of Michigan and lead author of the study, in the press release. “The theory is that vulnerabilities emerge in the metabolism of cancer cells due to specific genetic alterations.”

Using this theory, the researchers sought to investigate the relationship between metabolism and these genes. The press release states that when genes in charge of regulating metabolic function are deleted, cancer cells alter their own metabolism to compensate for this loss. Because of this, tumor formations and cancer growth can continue at aggressive rates despite genetic changes.

To evaluate this relationship in ovarian cancer, the team used a combination of complex metabolic modeling, machine learning, and optimization theory, which is concerned with minimizing the number of cancer cells at the end of treatment, in cell-line and mouse models.

Through this approach, the researchers gained insight into the ovarian cancer enzyme MTHFD2. In ovarian cancer cells with an impairment in the mitochondria due to the deletion of the UQCR11 gene, there is a critical imbalance of a metabolite known as nicotinamide adenine dinucleotide (NAD+). Using their machine-learning model, the researchers found that MTHFD2 would alter its own function in response to this imbalance, pivoting to create NAD+ in the cells. This results in a vulnerability that could be targeted to selectively kill off cancer cells while leaving healthy cells minimally affected, according to the press release.

“Personalized therapies like this are becoming an increasing possibility for improving efficacy of first-line cancer treatments,” said Abhinav Achreja, PhD, a research fellow and first author of the study, in the press release. “There are several approaches to discovering personalized targets for cancer, and several platforms predict targets based on big data analyses. Our platform makes predictions by considering the metabolic functionality and mechanism, increasing the chances of success when translating to the clinic.”