Scientists reveals how cancer-causing p53 mutations remain invisible to the immune system

One of the biggest challenges in cancer immunotherapy may not be that immune cells are too weak, but that some tumors give them too little to see. New research from Dana-Farber Cancer Institute scientists reveals how cancers carrying mutations in the p53 tumor-suppressor protein can hide those mutations from T cells and suggests a therapeutic strategy as a countermeasure: Change what the cancer displays on its surface so the immune system can recognize the tumor.

According to the study, the idea is particularly relevant to so-called "cold" tumors, which contain few effective tumor-fighting T cells or actively exclude and suppress them. Pancreatic cancer, prostate cancer, ovarian cancer, many breast cancers and glioblastoma are among the major tumor types commonly described as immunologically cold. Not every tumor within these categories is cold, but as groups, they have generally been much harder to treat with immune-checkpoint drugs than highly inflamed tumors, such as many melanomas.

The paper was published in Immunity.

The new work identifies a key reason that a tumor can remain immunologically quiet: The cancer may fail to display useful molecular targets in the first place. T cells do not read a tumor's DNA directly. Instead, they inspect tiny protein fragments, so-called peptides, held on the cell surface by HLA molecules. The complete collection of these displayed fragments is called the immunopeptidome-effectively, the cell's molecular "window display." If a cancer-specific fragment never reaches that window, or disappears too quickly, even a highly capable T cell may have nothing durable to attack.

That problem is especially important for p53. Mutations in TP53, the gene encoding p53, occur in roughly half of all human cancers. Because many TP53 mutations arise early in tumor development, they can be inherited by nearly every cancer cell as the tumor grows. Such "truncal" mutations have, therefore, been viewed as unusually attractive immunotherapy targets. In principle, hitting one could strike the entire tumor, rather than only one branch of it.

The study shows why that promise is often not realized. Using an ultrasensitive mass-spectrometry platform that physically measures peptides actually displayed on tumor cells, the investigators found that p53 is far less visible to T cells than DNA sequence or computer prediction alone would suggest. Of 175 predicted wild-type p53 peptide candidates examined across relevant HLA molecules, for example, only five were robustly detected. Many common cancer-causing p53 "hotspot" mutations occurred in regions that were poorly processed and, therefore, failed to generate any detectable surface targets.

The researchers then found several additional escape routes. Some patient tumors carrying potentially powerful p53 targets lacked the HLA molecule needed to display them. In the case of the p53 I195F mutation, high activity in cancer cells of an enzyme called ERAP1 acted like an overzealous molecular trimming machine, destroying an otherwise strongly immunogenic p53 fragment before it could be shown further to T cells as a counter-strategy. Deleting ERAP1, or blocking it with an inhibitor, restored recognition of those cancer cells by p53-specific T cells in laboratory experiments.

A different weakness emerged for the common p53 R175H mutation, which is already being pursued in engineered T-cell therapies in the clinic. The investigators identified highly sensitive T-cell receptors capable of responding to extremely sparse amounts of this target. But the mutant p53 fragment bound its HLA molecule so weakly that the complex was unstable and short-lived. The immune receptor was capable; the target itself was poor. T-cell activation and cancer-cell killing were, therefore, much weaker than against a more stable p53 target.

"Immunotherapy cannot attack what the immune system cannot see," said Ellis L. Reinherz, MD, chief of the Laboratory of Immunotherapy at Dana-Farber and professor of medicine at Harvard Medical School, a co-supervising author of the study alongside David A. Barbie, MD, director of the Lowe Center for Thoracic Oncology at Dana-Farber and a professor of medicine at Harvard Medical School.

First author Koji Haratani, MD, PhD, a postdoctoral fellow, was a primary contributor to this work, along with scientists Bruce Reinhold, PhD, and Jonathan S. Duke-Cohan, PhD, all of Dana-Farber.

A tumor may carry an ideal mutation in every cancer cell, but if that mutation is not processed into a stable surface target, making a stronger T cell may not be enough."

David A. Barbie, MD, Director, Lowe Center for Thoracic Oncology, Dana-Farber

From strengthening T cells to changing the tumor

Most cancer immunotherapies focus on the immune side of the battle. Checkpoint inhibitors release molecular brakes on T cells; engineered TCR-T and CAR-T therapies supply large numbers of highly active immune cells. The new study suggests that, for solid tumors, the other side of the equation may be equally important: First, make the cancer display better targets.

The authors call this approach an "immunopeptidome shift." Rather than relying only on the peptides which a tumor naturally chooses to display, drugs could deliberately alter antigen window display so that cancer cells reveal new peptide targets. The study points to several possible routes, including ERAP1 inhibitors, small molecules that bind to and change which peptides can fit into HLA molecules, and drugs that alter RNA splicing. If delivered selectively to tumors, such agents could potentially create a broader set of new targets that the immune system has not learned to ignore being switched on by designer drugs.

The goal would be to convert an immunologically inconspicuous tumor into one that attracts and activates many different cytotoxic T-cell populations recognizing myriad peptides in the window. In practical terms, this drug-induced switch would help turn a "cold" tumor "hot." That concept could be relevant to difficult cancers such as pancreatic, prostate, ovarian and brain tumors, as well as immunologically cold forms of breast cancer and other solid malignancies. The strategy is broader than p53: p53 provides a particularly important test case because TP53 is found in so many human cancers, but the underlying principle is to rewrite any tumor's surface antigen display, regardless of which mutation created the cancer.

A broader antigen display should also make immune escape harder. Therapies directed against a single target can fail if cancer cells stop displaying that target. In this respect, cancers have developed a strategy to shift their own immunopeptidome by editing out those HLA-bound peptide that make them vulnerable to T-cell attack. By contrast, shifting the immunopeptidome through drugs could potentially expose multiple new targets at once and recruit a polyclonal T-cell response. It might, therefore, be combined with checkpoint inhibitors, therapeutic vaccines or engineered T-cell therapies, rather than replacing them.

The findings also carry an immediate message for neoantigen drug development: A mutation predicted from tumor sequencing should not automatically be assumed to be a useful immune target. The relevant peptide must be processed, displayed by the patient's HLA molecules, remain sufficiently stable on the cell surface and be recognized by a T-cell receptor capable of responding at the very low target densities typical of cancer cells.

"The conventional question has been: How do we make the immune system attack harder?" Reinherz said. "Our results point to a complementary question: How do we make the tumor reveal more? If we can pharmacologically change the peptides a cancer displays, we may be able to create the targets that endogenous T cells, checkpoint therapies and engineered T cells need to work."

What the study does-and does not-show

The study does not show that a drug has already converted cold human tumors into hot tumors in patients. Rather, it establishes mechanisms by which tumors can suppress highly desirable immune targets and shows experimentally that one of those mechanisms, ERAP1-dependent peptide destruction, can be reversed. The work, therefore, provides a mechanistic foundation and a set of testable approaches for future therapies aimed at reshaping tumor antigen display.

The research integrates ultrasensitive immunopeptidomics, human tumor genomics, X-ray crystallography, engineered human T cells, single-molecule T-cell receptor measurements and functional tumor-cell killing assays.

In addition to Dana-Farber and Harvard Medical School, investigators from Vanderbilt University, Argonne National Laboratory, Boston Children's Hospital and collaborating institutions contributed to the study.