Irvine, Calif., Aug. 20, 2026 — Immunotherapy has changed the outlook for many people with advanced cancer by turning the immune system into an active weapon against malignant cells. Instead of poisoning rapidly dividing cells or removing tumors directly, these treatments can restore the immune system’s ability to recognize and destroy cancer. In advanced melanoma, one of the most aggressive forms of skin cancer, drugs that block the immune checkpoint protein PD-1 have transformed some once-terminal diagnoses into long-term survival. Yet the apparent success of these therapies can be deceptive. Even among patients whose tumors initially shrink or disappear, relapse remains common. Approximately seven in 10 melanoma patients treated with PD-1 blockade eventually experience disease recurrence, underscoring a central mystery of modern cancer medicine: why does the immune system lose control of a tumor after treatment appears to be working?
A team of researchers at the University of California, Irvine, has now used mathematical modeling and experiments in mice to identify a possible answer. Their study, published in Cancer Research, suggests that the rate at which regulatory T cells, or Tregs, enter a tumor may be a critical determinant of whether PD-1 immunotherapy produces durable control or eventual resistance. Tregs are specialized immune cells that normally prevent excessive or misdirected immune reactions, protecting healthy tissues from autoimmune damage. Inside a tumor, however, their suppressive properties can be exploited by cancer. By limiting the activity of cancer-killing immune cells, Tregs may help malignant cells survive even after immunotherapy has removed one of the tumor’s most important defenses.
The researchers focused on the complex cellular contest taking place within the tumor microenvironment. Effector T cells patrol tissues, identify abnormal cells and destroy them through direct cellular attacks and the release of toxic molecules. Tumors can interfere with this process through several mechanisms, including the display of PD-L1, a surface protein that binds to the PD-1 receptor on effector T cells. This interaction functions as an immune “brake,” reducing T-cell activity and allowing cancer cells to evade destruction. PD-1 blockade drugs interrupt the PD-1–PD-L1 connection, effectively releasing that brake. The treatment can revive exhausted effector T cells and restore their ability to attack. But the same biological intervention may also intensify or preserve Treg-mediated suppression, creating a previously underappreciated route through which the tumor can recover.
Rather than examining possible resistance mechanisms one at a time, the UC Irvine team built a mathematical model that represented the major interactions among tumor cells, effector T cells, Tregs and the PD-1 pathway. The equations were based on findings accumulated over decades of cancer biology and immunology research. They described how immune cells multiply, migrate into tumors, become activated or suppressed, and influence the growth or elimination of malignant cells. The model was then compared with experimental data from mice bearing melanoma tumors. By repeatedly refining the parameters until the simulations reproduced observed biological outcomes, the researchers created a computational framework intended to capture both the average behavior of the disease and the variability found among individual animals.
That variability was essential to the next stage of the investigation. Once the model had been validated, the team generated 342 virtual mice with melanoma. Each simulated animal received a different combination of biological characteristics, such as the growth behavior of tumor cells, the abundance of immune cells, their rates of activation and their ability to migrate through tumor tissue. This approach allowed the researchers to explore a broad range of plausible immune environments without having to perform a separate experiment for every possible combination. The virtual population was then treated with simulated PD-1 blockade, and the researchers compared the characteristics of animals that achieved favorable responses with those that eventually experienced tumor regrowth.
More than 30 biological parameters were included in the analysis, but one variable repeatedly separated the two groups: the speed of Treg infiltration into the tumor. The model indicated that tumors receiving Tregs rapidly were more likely to resist or escape PD-1 blockade, while slower Treg entry was associated with improved treatment responses. This finding does not mean that Tregs are the only cause of resistance, or that every patient with a high level of Treg activity will fail to respond. Instead, it identifies the rate of Treg influx as a potentially powerful control point in the dynamic system that determines whether immune pressure remains strong enough to suppress cancer. “The mathematical analysis pointed directly to one variable,” said Rachel Sousa, the study’s first author. “It indicated that the rate of Treg infiltration into the tumor was the critical factor.”
The team next tested that prediction in living animals. Researchers engineered mice whose Tregs were less efficient at migrating into tumors while leaving the rest of the immune system intact. These animals were then treated with PD-1 blockade immunotherapy. The combination of reduced Treg infiltration and checkpoint inhibition substantially outperformed PD-1 blockade alone. In mice whose tumors were not completely eradicated, the combined intervention slowed tumor growth and nearly doubled survival duration. The experiment provided an important test of the model because it did not merely show that Tregs were present in resistant tumors; it examined whether changing their movement into the tumor could alter the outcome of therapy. The agreement between the simulated prediction and the mouse experiments suggests that Treg trafficking may be a more actionable target than simply measuring the total number of immune cells within a tumor.
The findings also help explain why earlier efforts to suppress Tregs have been difficult to translate into effective treatments. Tregs are not inherently harmful: throughout the body, they prevent uncontrolled inflammation and protect healthy organs from immune attack. Broadly eliminating them could therefore produce dangerous autoimmune or inflammatory side effects, while also damaging beneficial immune responses. The UC Irvine study points instead toward a more selective strategy, in which the movement or activity of tumor-protective Tregs is disrupted specifically within the cancer microenvironment. Such an approach could potentially be paired with PD-1 blockade, preserving the immune system’s protective functions elsewhere while preventing Tregs from rebuilding the suppressive conditions that allow a tumor to return.
The researchers emphasize that the work is not an immediately available treatment for patients, and the results in mice must be tested through further preclinical studies and, eventually, carefully designed clinical trials. Nevertheless, the study illustrates how mathematical oncology can accelerate the search for therapeutic targets. Conventional research often evaluates one proposed mechanism after another, with each experiment requiring substantial time, biological material and funding. A validated computational model can screen many mechanisms and treatment combinations before laboratory teams commit to large-scale experiments. Francesco Marangoni, one of the study’s senior investigators, said the project brought mathematics and biology together so that each discipline could inform the other. John Lowengrub, the other senior investigator, said the model not only forecast biological outcomes but also identified a potentially overlooked target for improving cancer therapy.
The model may ultimately prove useful beyond melanoma and beyond PD-1 blockade. Because it represents the relationships among tumor growth, immune-cell recruitment, immune suppression and treatment response, researchers can adapt it to examine other immunotherapies or combinations of drugs. It could also help determine which patients are most likely to benefit from interventions aimed at Treg migration, provided that equivalent biological markers can be identified in human tumors. The broader message is that resistance to cancer therapy may not arise from a single mutation or a single immune defect, but from the changing balance of cells moving through the tumor over time. By revealing how one rate of cellular movement can influence that balance, the UC Irvine study offers a potential roadmap for making immunotherapy more durable—and demonstrates how computer-generated disease models can help turn the enormous complexity of cancer biology into testable treatment strategies.
Subject of Research: Regulatory T-cell infiltration as a determinant of acquired resistance to PD-1 immunotherapy in melanoma.
Article Title: Mathematical and Mouse Models Identify Regulatory T Cell Influx as A Key Determinant of Acquired Resistance to PD-1 Immunotherapy
News Publication Date: Aug. 20, 2026
Web References: https://news.uci.edu/ ; https://aacrjournals.org/cancerres/article/doi/10.1158/0008-5472.CAN-25-5784
References: Cancer Research, “Mathematical and Mouse Models Identify Regulatory T Cell Influx as A Key Determinant of Acquired Resistance to PD-1 Immunotherapy.”
Keywords: cancer immunotherapy, melanoma, PD-1 blockade, PD-L1, regulatory T cells, Tregs, tumor microenvironment, immunotherapy resistance, mathematical modeling, computational oncology, effector T cells, cancer research
Tags: advanced melanoma survival ratescancer immunotherapycancer recurrence predictionimmune checkpoint inhibitorsimmune system mechanisms in cancerimmunotherapy resistance mechanismsmathematical modeling in cancer researchmelanoma treatment and relapsemice model studies in cancer researchPD-1 blockade efficacy and challengesrole of regulatory T cells in tumor resistancetumor immune evasion strategies
