Mathematical model of phenotypic plasticity predicts interconversion is a drug target for neuroblastoma.

Cancer is rarely a single, uniform enemy. In many tumors, particularly the aggressive childhood cancer known as neuroblastoma, the disease is a shifting landscape of different cell types. Some of these cells grow rapidly and resemble normal nerve cells, while others are slower, more primitive, and resemble the embryonic cells from which nerves first develop. Crucially, these different cell types are not stuck in their roles. They can change into one another, a biological flexibility called phenotypic plasticity. This ability to transform is a major reason why treatments often fail; a drug might successfully kill the fast-growing cells, only for the remaining survivors to morph into a different, drug-resistant form that eventually causes the cancer to return. Understanding how these cells switch identities, and how to control those switches, is becoming a central challenge in finding a cure.

Researchers at the University of Hull have taken a new approach to this problem by building a mathematical model to simulate the behavior of these shifting cell populations. Instead of testing drugs in a petri dish or a living patient immediately, they created a virtual environment where they could observe how neuroblastoma cells interact over time. Their model included three distinct cell groups: the fast-growing, nerve-like cells; the slower, primitive cells; and a middle-ground type that acts as a bridge between the two. The team mapped out every possible way these three groups could transform into one another. With six potential pathways for change between the three types, there were 64 different scenarios of how the cells might interact. They then ran thousands of computer simulations to see which of these 64 scenarios would allow a specific drug to shrink the tumor most effectively.

The simulations revealed that not all patterns of cell change are equal when it comes to treatment. The researchers found that only five specific patterns of interconversion allowed a drug to successfully reduce the tumor population to very low levels. Two of these patterns worked best when the goal was to suppress the fast-growing nerve-like cells, while the other three were most effective against the slower, primitive cells. A key discovery was that the speed at which cells changed from one type to another mattered deeply. When trying to eliminate the fast-growing cells, the treatment worked best if the cells were changing types quickly. Conversely, when targeting the slower, primitive cells, the treatment was more successful if the cells changed types slowly. This suggests that the timing of cell transformation is just as important as the drug itself.

To test if these findings could lead to a real cure, the team simulated a treatment schedule that alternated between different strategies. They found that a single round of treatment was rarely enough to keep the cancer away. Instead, the most reliable path to remission involved a multi-stage approach. The most successful strategy involved first using a method to force the fast-growing cells to change into the slower, primitive type, and then immediately applying a drug designed to kill those primitive cells. Once that population was suppressed, the strategy would switch to a different pattern that forced the remaining cells to change back, followed by a drug to kill the fast-growing cells again. This cycle of forcing the cancer to change its shape and then striking it with the right weapon proved far more effective at maintaining a cancer-free state than simply attacking one cell type and hoping for the best.

The study also looked at whether these theoretical patterns actually exist in the real world. Three of the five most promising patterns identified by the computer model correspond to biological processes that scientists have already observed in neuroblastoma. Furthermore, the researchers noted that known compounds, such as retinoic acid, can already be used to induce some of the necessary cell changes. This means the strategies identified in the simulation are not just mathematical fantasies but are grounded in biological reality. The team concluded that by combining these specific cell-transformation patterns with targeted drugs in a carefully timed sequence, it may be possible to outmaneuver the cancer's ability to adapt. While the model is a simulation and not a clinical trial, it provides a clear roadmap for how doctors might one day use drugs not just to kill cancer cells, but to guide them into a form that is easier to destroy.