When T cells activate, they become “T helper,” or Th cells, Faeder explains. In this role they encourage other white blood cells to attack the tissue, killing its cells and any invading microbes they may contain. Alternatively, T cells can become “regulatory T cells,” or Treg cells, sending out signals that make other cells ignore the tissue. Th cells help destroy invading pathogens, or cancer cells. But they also can mistakenly lead an attack on healthy tissues, causing autoimmune diseases like type I diabetes or lupus. Their attacks can also cause a transplant recipient to reject a donor organ. Treg cells help the body avoid those problems by recognizing healthy tissues as “self,” and so, good to be left alone.

Scientists had previously studied in some detail how the molecules that carry messages within T cells pass along the friend-or-foe message. In particular, they performed laboratory experiments as well as computer modeling that partly explained how these messengers’ levels ebbed and flowed to create the cell’s decision.

Initially, they found, the same messenger molecules either increase or decrease in much the same way in either case. But in the presence of a strong indicator of infection—a high level of “antigens” that signal a microbe or cancer may be present—these changes are larger, and persist. The T cell becomes a Th cell. With a weaker antigen signal, the changes in the messengers are smaller and temporary, and the cell becomes a Treg cell.

Adding Details, Time

The problem with the model was that it lacked detail—and it didn’t allow investigation of how the timing and strength of the signal would affect how it’s received by the T cell.

As Faeder explains, “We lacked the ability to study how quantitative differences between signals and cells might translate into different outcomes.” Enter BioNetGen, which allowed the MMBioS team to simulate the biochemical signals inside the T cell in much more detail, and with a better attention to these differences.

“When you’re dealing with these molecular interactions, the conventional ways of writing down the model gain complexity pretty quickly, with the number of molecular species growing in a difficult-to-manage way,” Faeder says. “The models turn into ‘hairballs’ in which everything is interacting with everything else.” They become impossible for a human to understand, and too complicated even for powerful computers to calculate.

Faeder and colleagues Michael Blinov and William Hlavacek developed BioNetGen at Los Alamos National Laboratory in the early 2000s to tackle these kinds of complex signaling pathways. The computer-modeling tool works by turning the molecules and their chemical interactions into a set of rules that drive the simulation. Instead of worrying about the details of how the molecules in the system can combine, the rules allow complex situations to be handled in a simple way.

“Rule-based modeling allows the simulation of models with huge numbers of possible states using only a handful of biochemical rules, thus ‘taming’ the hairball,” Faeder says.

A Window to New Details—And a New Actor

A BioNetGen analysis of the actors in the Th/Treg decision point generally confirmed more crude, earlier computer modeling results. Importantly, the virtual T cells in the compuer model acted the same way as real T cells did in the laboratory.

The simulation also opened a window to details that scientists hadn’t earlier appreciated. For one thing, the BioNetGen model suggested exactly how duration and level of an immune signal reaches a “tipping” point. Slightly higher, and the T cell becomes a helper and encourages the immune system to attack. Slightly lower, and the signal inside the cell doesn’t persist, dying down and letting the T cell become a Treg, calming the immune system’s response. The complex dance of the signals in the computer made the change happen just as it does in real T cells.

Possibly the most interesting finding of the simulation, Faeder says, is that a previously under-appreciated molecule plays a much more important role than researchers had realized. Phosphatase and tensin homolog (PTEN), a protein responsible for shutting down one of the key signals inside the T cell, proved far more pivotal in the friend-or-foe decision than expected.

“We tend to focus in modeling on things that generate signals, not things that block signals,” Faeder says. “For some reason, kinase molecules [which initiate cell signals] are much more carefully studied in the literature than phosphatase molecules,” which turn them off.

In PTEN’s case, this philosophical blind spot was surprising in retrospect, Faeder notes. This is because the protein also plays a role in tumor development, and is often the second mutation seen when cells start to turn cancerous. The BioNetGen model helped focus the scientists’ attention on the role of this shut-off signal in the immune friend-or-foe decision. The finding may also shed light on PTEN’s role in cancer.

Dialing an Immune Response

The MMBioS team and their collaborators would like to see their work help alter patients’ immune systems to fight a number of diseases. In theory, doctors could dial up the signal to produce Th cells that attack cancerous tissues, or dial up Treg cells to protect healthy tissues from mistaken immune attacks.

“We’re working with immunologists at Pitt studying this process for a way to modulate the outcome,” Faeder says. “For example, in a cancer vaccine you’re trying to boost the immune response to tumor tissues.” On the other hand, “if we could turn up Treg cells, we could turn off inappropriate immune responses, in effect ‘turning off’ type I diabetes.”

Copyright © 2020 National Center for Multiscale Modeling of Biological Systems. All Rights Reserved.