Scientists have found that only about 1% of presented tumor antigens come from oncogenic mutations. The remaining 99%, previously overlooked, may offer better treatment targets [1].
Going after mutations
Cancer arises from a series of mutations that drive abnormal cellular behaviors, such as runaway proliferation. Cancerous cells also normally present abnormal peptides on their surfaces that the immune system detects and targets. This presentation is carried out by major histocompatibility complex (MHC) class I proteins, which are not to be confused with MHC class II proteins that present external peptides ingested by the cell, such as from bacteria or viruses.
Targeting these mutated tumor-specific antigens (mTSAs) has become a strategy behind many experimental treatments, most notably anti-cancer vaccines, which prime the immune system to go after cells that express a particular mTSA. However, since mutations are often tumor-specific, those vaccines must be tailored to each patient, which makes them finicky and prohibitively expensive. Despite all the effort, mTSA-targeting vaccines often do not produce the expected result.
To create a vaccine, scientists sequence the tumor’s genome, identify tumor-specific mutations, and then use algorithms to predict what an antigen derived from the mutated gene would look like. The vaccine is then built based on that predicted sequence.
Looking for what’s actually there
Novel techniques have made it easier to look for antigens that are actually present on the cell’s surface rather than rely on predictions. In a new study published in Nature Cancer, scientists from the Institute for Research in Immunology and Cancer (IRIC) at the University of Montreal catalogued tumor antigens in 505 melanoma and 90 lung cancer samples by using a state-of-the-art multi-omic approach, and they arrived at striking results.
They found that only about 1% of the tumor antigens came from mutations. The remaining 99% came from non-mutated parts of the genome, often from sequences that are normally silent or minimally active in healthy tissues.
These fell into three groups. First were aberrantly expressed tumor-specific antigens (aeTSAs), peptides from unmutated genomic regions that are normally silent in healthy adult tissues but activated in cancer. Their origins varied and often included “non-canonical regions,” such as gene parts that are usually spliced out of the final RNA used to make the protein (introns), intergenic regions, and non-coding RNAs.
The second group consisted of tumor-associated antigens (TAAs): unmutated proteins that are overproduced in cancer but also found in some healthy tissues. Finally, the third group included lineage-specific antigens (LSAs): proteins that are typical of the tissue the cancer came from, such as melanocyte markers in melanoma.
The researchers found that mutated antigens are rare because many mutations don’t get transcribed into RNA, which means the corresponding proteins don’t get made. If there’s no protein, there’s nothing for the immune system to see. Plus, the few mutations that are transcribed often occur in genomic regions that don’t generate peptides that are well suited for presentation to immune cells.
Meanwhile, many aeTSAs are made into proteins, get presented on the cell surface, and can trigger immune responses. The study showed that immune cells from healthy donors can recognize and kill cancer cells presenting these aeTSAs. This suggests that they could be powerful targets for cancer immunotherapy.
Importantly, these aeTSAs were often shared across patients, unlike mutated antigens, which tend to be unique to each person. Therefore, aeTSAs could be used to develop generalized, off-the-shelf cancer vaccines or T cell therapies.
The upshot of this study is that while mutations often initiate cancer, that doesn’t mean the mutated proteins are abundant or relevant as immune targets once the tumor is established. Other aberrant proteins, not derived from the initiating mutations but essential for maintaining cancerous behavior, might be better targets for immunotherapies due to their relative abundance or widespread presentation.
“This is an intriguing study that explores a novel approach to cancer immunotherapy,” said Dr. Anna Barkovskaya, a researcher at Lifespan Research Institute who was not involved in this study. “Previous developments in the field, most notably immune checkpoint blockade in mutation-heavy melanoma and non-small cell lung cancer, were dramatic but were only successful for a subset of patients and not applicable to cancers where mutational load is smaller. Multiple studies that followed focused on trying to identify new mutated tumor-specific antigens, but by their nature, those are patient-specific, rare, and are expressed at a low level.”
“By contrast,” she explained, “the authors of this study used genome-wide sequencing that isn’t limited only to protein-coding regions, to identify non-mutated tumor antigens that were either aberrantly expressed or specific to the lineage of origin of the cancer cells. They found that such antigens are much more common than the mutant ones, tend to be more abundant, and are highly immunogenic, inducing a potent and specific cytotoxic CD8+ T cell-mediated response.”
More evidence from a different study
A complementary study, published almost simultaneously in Science, reinforces these insights by showing that most tumor-presented antigens in human pancreatic cancer arise not from mutations but from noncanonical sources, such as long non-coding RNAs and untranslated regions (UTR) [2]. As in the IRIC study, the team found that these aberrantly expressed peptides, which are absent from normal tissues, can elicit T cell responses and serve as potent immunotherapy targets.
Just like in the first study, some of the cancer-derived peptides were shared by multiple patients, which means they can be used to create off-the-shelf treatments. Some were highly immunogenic and enabled the creation of genetically modified T cells that “could exert robust killing of patient-derived pancreatic cancer organoids both ex vivo and in vivo.” Pancreatic cancer remains one of the most treatment-resistant cancers, making this research particularly important.
Literature
[1] Apavaloaei, A., Zhao, Q., Hesnard, L., Cahuzac, M., Durette, C., Larouche, J. D., … & Perreault, C. (2025). Tumor antigens preferentially derive from unmutated genomic sequences in melanoma and non-small cell lung cancer. Nature Cancer, 1-19.
[2] Ely, Z. A., Kulstad, Z. J., Gunaydin, G., Addepalli, S., Verzani, E. K., Casarrubios, M., … & Freed-Pastor, W. A. (2025). Pancreatic cancer–restricted cryptic antigens are targets for T cell recognition. Science, 388(6747), eadk3487.
