Stopping Cancer Before It Begins

A long standing Question

For more than twenty years, the concept of cancer immunoediting has shaped how scientists understand the relationship between tumors and the immune system. Seminal studies by Shankaran et al., Nature 2001 and Koebel et al., Nature 2007 showed that the immune system doesn’t just eliminate cancer—it also sculpts which tumor cells survive. Later work from Matsushita et al., Nature 2012 strengthened this model by identifying immune pressure as a selective force that favors tumor “escape” through antigen loss. DuPage et al., Nature 2012 further emphasized that neoantigens themselves can define which cancers the immune system can recognize and suppress.

Together, these foundational studies led to a widely held view: that T cells exert a Darwinian pressure on tumors, removing the most visible cells and leaving behind antigen-negative variants that thrive in their wake.

But despite the power of this framework, one critical question remained unresolved: Does immune pressure always act only after tumors are established, or can it prevent tumor formation altogether? The models that defined immunoediting were largely retrospective—they revealed what remained after immune selection, but not what happened in the first moments when a transformed cell first appeared.

Our lab set out to bridge this gap. By combining genetically engineered mouse models with trackable fluorescent neoantigens, we developed a way to observe T-cell–tumor interactions from the instant of tumor initiation. This approach allowed us to ask when and how immunoediting acts in vivo—not just how it sculpts surviving cancers, but whether it can stop tumors from existing in the first place.

What we found reframes the concept of immunoediting. Rather than acting only as a sculptor that shapes established tumors, the immune system often acts as a gatekeeper—eliminating nascent tumor clones before they ever become visible. This early immune control means that much of the immune system’s influence on cancer happens before a tumor even exists, fundamentally shifting how we think about cancer surveillance and evolution.

Discovering Immunoediting in Real Time

To move beyond inference, we needed a model where the earliest moments of tumor–immune interaction could be observed directly. Traditional transplant systems—where tumors are already formed before being placed into a mouse—are powerful but limited: they start the clock long after the immune system’s first opportunity to act. Even spontaneous models, while closer to physiological reality, rarely allow one to track which specific transformed cells are recognized by T cells, and when.

Our solution was to build a reportable tumor initiation system that made each emerging cancer cell traceable from the moment of transformation. In our sarcoma model, we introduced a defined neoantigen together with oncogenic drivers, allowing us to visualize how T cells respond as tumors first arise within their native tissue. The fluorescent neoantigen served as a molecular signature, marking the cells that were potential immune targets and revealing when those targets were lost.

What we saw was unexpected. T cells did not wait for a tumor to form—they moved early and decisively. Within days of transformation, CD8⁺ and CD4⁺ T cells converged on emerging tumor sites, eliminating the majority of nascent tumor cells. Roughly half of all initiated sarcomas were destroyed before they ever became detectable. This was immunoediting in real time, but it looked different from the classic model: it wasn’t just about selecting for antigen-negative escape variants. It was about preventing tumors from existing at all.

Even more striking, we found that immune control extended beyond the antigen-positive cells. The T cells triggered a local cytokine environment dominated by interferon-γ, which altered the tumor niche itself. Neighboring antigen-negative cells—those that should have been invisible to the immune system—were also eliminated or growth-suppressed. This meant that T cells could act both specifically, through antigen recognition, and non-specifically, through microenvironmental reshaping. The net result was that immune surveillance constrained not only which clones survived, but whether a tumor could take hold in the first place.

Implications

These findings expand the concept of cancer immunoediting from a process that merely sculpts existing tumors to one that prevents tumors from forming. They suggest that immune pressure is not only a selective force but also a suppressive one — acting before malignancy becomes clinically visible.

By showing that T cells can eliminate emerging tumor cells, including some that never display recognizable antigens, this work reframes immunosurveillance as an active, ongoing process that shapes tissue integrity over time. It highlights the importance of understanding the earliest immune–tumor encounters — when intervention could be most effective — and sets the stage for future work connecting tumor differentiation, immune recognition, and cancer prevention.