New research helps model how the immune system shapes cancer development

Understanding the interaction between immune cells and cancer cells has important implications for cancer immunotherapies, including checkpoint inhibitor drugs and cell-based therapies, as well as newer treatments like cancer vaccines.

Over the last decade, computational oncologist Benjamin Greenbaum, PhD, has tried to shed light on how the immune system senses cancer. An essential part of Dr. Greenbaum’s research has focused on genetic patterns in cancer cells that activate a person’s own immune system as cancer evolves. This arm of the immune system recognizes molecular patterns that signal a threat.

Although these patterns, called “pathogen-associated molecular patterns” (PAMPs), arise in the patient’s cells, they can resemble a foreign invader and initiate an immune response.

Repetitive DNA sequences, where certain segments occur again and again, can give rise to PAMPs. These repeated sequences are normally not active, even though they represent about half of the human genome. But when they do get activated, they can produce RNA molecules that resemble patterns found in the genetic code of viruses — a phenomenon called “viral mimicry.”

The Greenbaum lab has been taking a deep dive into modeling viral mimicry, trying to understand what causes it, how the immune system responds to it, and how this could shape cancer cells’ development process.

A Tool To Quantify Viral Mimicry

Dr. Greenbaum’s lab — including members Alexander Soloyov, PhD, Siyu Sun, PhD, and Stephen Martis, PhD — in collaboration with an international team of researchers, has developed a mathematical model to quantify viral mimicry based on methods from statistical physics, machine learning, and the dynamics of evolution. This approach enabled them to understand the signals that cause certain viral mimics to be preserved in the genome, while other PAMPs are lost.

They published their results September 24, 2025 in Cell Genomics.

“Being able to quantify mimicry, to have good methods for assessing what turns it on or off, is going to help us understand how the innate immune system interacts with cells and impacts their evolution, including during cancer evolution,” Dr. Greenbaum says.

They found that specific classes of repetitive DNA elements are particularly good at mimicking viruses. This suggests that the mimicry might have a purpose, such as serving as a general defense against viral infections. These repetitive elements may also help protect developing cells against viral threats.

It’s also possible that mimicry can signal cellular dysfunction — sending up a flare to the immune system because there’s something wrong with a cell that could threaten an organism’s well-being.

“By being quantitative about this topic, we can provide a methodology for answering a lot of questions about how cancer and other cells get the attention of the innate immune system despite these sick cells essentially still being ‘self,’” Dr. Greenbaum says. “Where do these mimics come from, and why do cancer cells still have some of them if they result in increased attention from the immune system?”

For example, in 2024 his lab reported in Immunity that their research tool enabled them to gain insight into how pancreatic cancer cells avoid drawing attacks from immune cells by accommodating DNA repeats called retrotransposons.

The mathematical approach developed by Dr. Greenbaum and colleagues should lend itself to many discoveries about how innate immunity affects cancer development and progression.

“A better understanding of what activates the innate immune system can help us figure out how to improve immunotherapies,” Dr. Greenbaum says. “With cancer vaccines, it could help us learn how to make the vaccines more or less visible to the immune system so that we can better tune immune engagement.”

Moving forward, Dr. Greenbaum says, the precision that the research model provides will help with the understanding of this important aspect of human biology and with the development of new practical applications.