Back to Beyond the Research Lab
  • Therapeutics

Rewriting Tumor Identity to Make Cancer Detectable

Many cancers evade immune attack not by overpowering T cells, but by hiding—failing to display the peptide–HLA signals the immune system needs to recognize them.

Some cancers survive not because they overpower the immune system, but because they avoid detection by it altogether. Instead of displaying peptide-HLA complexes—molecular “ID badges” made of protein fragments at the cell surface—these tumors present too few signals to stand out. As a result, T cells (the immune cells responsible for identifying and killing cancer) have no way to recognize the threat. Ellis L. Reinherz, MD, an immunologist at the Dana‑Farber Cancer Institute and professor of medicine at Harvard Medical School, is working to reverse that invisibility by changing what tumors show the immune system in the first place. 

It’s a problem he has been circling for decades, beginning with his foundational discoveries in T‑cell biology including the T-cell receptor, its CD3 signaling subunits and antigen recognition, as well CD4 and CD8 subsets and building as the rise of cancer immunotherapy revealed just how often those signals are missing altogether. 

These peptide-HLA complexes are essentially the language the immune system uses to recognize cancer. If that language is missing or erased, immune surveillance breaks down.  

Ellis L. Reinherz, MD

Identifying a Gap in Current Treatment Modalities 

Reinherz points out that large‑scale analyses across many tumor types now point to a sobering reality: immune failure in cancer is frequently not due to weak immune responses, but to missing information. In particular, cytolytic T lymphocytes (CTLs)—CD8 T cells responsible for directly killing cancer cells—depend on peptide-HLA complexes on the cell surface to distinguish tumors from healthy tissue. When those signals are absent, even fully functional CTLs cannot engage. 

“These peptide-HLA complexes are essentially the language the immune system uses to recognize cancer,” Reinherz says. “If that language is missing or erased, immune surveillance breaks down.” 

Using advanced mass spectrometry to directly examine peptides bound to HLA molecules, Reinherz and his colleagues analyzed data from tumors and cell lines in conjunction with information amassed on more than 5,000 patient tumor samples. The results were striking: across many cancers, few—if any—predicted tumor‑associated antigens or neoantigens were actually displayed on the cell surface. Some tumors suppress antigen display from the outset, while others adaptively restrict peptide presentation once CTLs begin applying immune pressure. 

This form of immune evasion helps explain why cancers (such as breast, prostate, lung, colon, bladder, kidney, non-Hodgkins lymphoma, pancreatic, uterine, thyroid and brain) have remained largely resistant to immunotherapy. Treatments that rely on immune activation alone cannot succeed if there is nothing for immune cells to recognize. 

Decades of Evidence Reveal a Pattern of Immune Escape 

Over time, a series of discoveries by Reinherz and his collaborators have clarified just how deeply this problem is embedded in cancer biology. They found that even when these neoantigens could be recognized by high‑quality T‑cell receptors, they appeared on tumor cells in vanishingly small and unstable quantities. The failure, the researchers showed, was not target selection — it was target display. 

The most comprehensive evidence emerged in 2025, when Reinherz, Barbie, and a large, multidisciplinary team reported findings from ultra‑sensitive mass spectrometry studies examining thousands of predicted p53 neoantigens encoded by mutant TP53 genes. Those genetic abnormalities are found in 50% of human cancers and long regarded as ideal immunotherapy targets. Despite being present in tumor DNA, however, nearly all p53 neoantigens were physically absent from the tumor cell surface. In addition, the research revealed multiple tumor‑driven mechanisms—including loss of specific HLA molecules and destruction of peptides during processing—that systematically erase the relatively rare immune targets before CTLs ever have a chance to effectively respond. 

Crucially, the same work showed that altering peptide processing could restore antigen display and re‑enable CTL recognition in tumor models. If tumors could actively control what they show the immune system, Reinherz reasoned that visibility itself might also be engineered. 

Reprogramming the Immunopeptidome 

That realization prompted a shift in strategy. Rather than continuing to search for rare, naturally occurring antigens that tumors routinely suppress, Reinherz’s lab began developing a way to modify the immunopeptidome itself — the complete set of peptides presented by HLA molecules on cancer cells. 

Drawing on the lab’s long‑standing strengths in structural biology, mass spectrometry and  cellular and molecular immunology, the team identified small, drug‑like molecules that modulate antigen presentation pathways, resulting in dramatic changes in the array of peptides displayed by HLA molecules, using an artificial intelligence (AI)-based discovery pipeline. Preclinical studies show that treatment can induce the display of hundreds of new peptides on tumor cells, creating a far richer set of immune signals for CTLs to recognize. 

By doing so, the approach is designed to generate a polyclonal CTL response—the activation of many different tumor‑killing T cells at once—reducing the likelihood that tumors can escape by eliminating a single target. Proof of concept has already been established for one HLA context, demonstrating that antigen presentation can be reliably reprogrammed in cancer cells. 

“If you give the immune system clear, actionable information,” Reinherz stresses, “CTLs already know how to do the rest.” 

Precision Delivery to Tumor Cells 

Because antigen presentation is a fundamental cellular process, precision is essential. To restrict these changes to tumor cells, the team is pairing peptide‑modifying compounds with antibody–drug conjugates (ADCs) engineered to recognize cancer‑cell surface markers. Once internalized, the ADC releases its payload selectively inside malignant cells, altering their peptide display while sparing healthy tissue. 

Using additional AI‑guided drug design, Reinherz and his colleagues aim to extend this approach to approximately 40 HLA molecules — a range that would cover nearly all racial and ethnic populations worldwide. By developing ADCs targeting markers shared across the twelve most common cancers, the team is building a broadly applicable therapeutic platform rather than a disease‑specific solution. 

A Collaborative Effort 

This work builds on long‑standing collaborations across Dana‑Farber and partner institutions, involving experts in oncology, immunology, structural biology, and mass spectrometry. Key collaborators from Dana-Farber include David A. Barbie, MD, PhD, Director of the Lowe Center for Thoracic Oncology; Koji Haratani, MD, PhD, researcher and thoracic oncologist, Bruce Reinhold, PhD, Principal Research Scientist; Jonathan Duke-Cohan, PhD, Principal Research Scientist; Robert J. Mallis, PhD, Instructor in Medicine; and Andrew Parkins, PhD, Postdoctoral fellow. In addition, other contributors include Matthew J. Lang, PhD, Professor of Chemical and Biomolecular Engineering at the Vanderbilt University Medical Center and Kemin Tan, PhD, Research Scientist at Argonne National Laboratory. 

This research has been supported by multiple grants from the National Institutes of Health, including the National Cancer Institute and NIAID, as well as the Ludwig Center at Harvard Medical School, the Mark Foundation, the Novartis Global Scholars Program, and additional philanthropic and foundation funding. 

Engineering Visibility, Expanding Possibility 

Reinherz emphasizes that this work represents a fundamental shift in immuno‑oncology. Rather than pushing the immune system to respond more forcefully with existing CTL, his approach focuses on engineering a broader and more immunogenic set of peptide-HLA signals on the surface of cancer cells, enabling immune recognition to occur against many new peptide-HLA targets on the surface of cancer cells simultaneously. For tumors long labeled immunologically “cold,” resistance may not reflect a failure of immunity, but a lack of information. 

If successful, this strategy could broaden who benefits from immunotherapy by turning previously undetectable cancers into ones the immune system can finally engage. That shift could translate into meaningful outcomes: opening immunotherapy to patients who currently have few options, enabling more durable tumor control, and reducing the likelihood of immune escape. By reframing cancer as a problem of missing signals, Reinherz’s work points toward a future in which inducing tumor visibility reshapes both treatment possibilities and patient outcomes. 

Team Members