Core Framework & Diagram What Are Tumor Antigens?
July 7, 20261 Min Read The Immune Surveillance Theory
July 7, 2026The 'evidence' of cancer cells: the abnormal proteins that reveal their identity
—— Tumor antigens: what the immune system uses to recognize cancer cells — and the scientific basis of individualized cancer vaccines.
I. Why can the immune system distinguish cancer cells from normal cells?
This question is deeper than it appears. The difference between cancer cells and normal cells is at the genetic level — cancer cells have mutations, normal cells don't. But the immune system can't directly 'read' DNA; what it can see are the protein fragments cells display on their surfaces. Inside cells, proteins are chopped into short peptides by proteasome complexes; these peptide fragments are transported via TAP proteins into the endoplasmic reticulum, bound to MHC-I molecules, then transported to the cell surface for display — this process is called 'antigen processing and presentation.'
Normal cells display peptide fragments all from normal self-proteins; T cells are 'trained' during thymic development to have tolerance to these self-peptides, and won't attack them. But cancer cells with gene mutations produce abnormal proteins; the peptide fragments from these abnormal proteins differ from normal cells — they're 'new name cards' T cells haven't seen before, not in the self-tolerance list. When these peptide fragments are displayed on cancer cell surfaces via MHC-I, specific T cells can recognize them and mark the cell as an attack target.
2. Neoantigens: each person's cancer is unique
Neoantigens are currently the most studied class of tumor antigens in immunology, because they possess a quality other tumor antigen types lack: they only exist in cancer cells — there's simply no corresponding protein sequence in normal cells. This means: immune attacks targeting neoantigens theoretically only hit cancer cells, completely harmless to normal cells.
Importantly: each patient's cancer cell mutation combination is unique, so their neoantigen profiles are also unique. This is why the 'individualized tumor vaccine' concept became a research hotspot over the past decade: by genetically sequencing a patient's tumor, finding unique mutations, predicting which mutations will produce neoantigens that can be recognized by the immune system, then synthesizing these specific peptide fragments as vaccines to inject, activating the patient's own T cells to attack cancer cells containing these neoantigens.
In 2023, Moderna and BioNTech respectively reported their individualized mRNA cancer vaccine clinical trial results, showing encouraging early efficacy signals in difficult-to-treat cancer types like melanoma and pancreatic cancer.
3. Cancer-testis antigens (CTAs): accidental revival of sleeping genes
One category of tumor antigens is a fascinating 'evolutionary accident' — Cancer-Testis Antigens (CTAs). These proteins, under normal circumstances, are expressed in only two places: the fetal development period, and adults' testes (males) or placenta. In other adult tissues, these genes are silenced. But in some cancers, these 'should be silent' genes are reactivated, beginning to express these embryonic/germ cell-specific proteins.
Since other normal tissues don't express these proteins, they're relatively 'new' targets for the immune system — immune tolerance toward them is much weaker than for ordinary self-proteins. The most famous CTAs include NY-ESO-1 (highly expressed in melanoma, ovarian cancer, and multiple other cancers) and MAGE family proteins (expressed in lung cancer, melanoma, etc.). Both are important targets for cancer vaccines and T cell therapies.
Why doesn't the testis suffer immune attack? Because the testis is an 'immune privilege' organ with special barriers and regulatory mechanisms preventing immune cells from entering. Cancer cells 'borrow' testis-specific proteins but don't have the testis's immune protection — result: these proteins become relatively exposed targets.
4. HER2: the story from 'overexpression' to precision treatment target
HER2 (Human Epidermal Growth Factor Receptor 2) is the most successful treatment target example among TAAs, worth discussing separately. Normal cells express small amounts of HER2 — a receptor protein involved in cell growth signals. In approximately twenty to twenty-five percent of breast cancer patients, the HER2 gene is amplified or overexpressed, making HER2 protein density on cell surfaces far higher than normal cells — these cancer cells display HER2 'name cards' dozens to hundreds of times as densely as normal cells.
In 1998, Trastuzumab (Herceptin) was FDA-approved targeting HER2 overexpression, becoming one of the most important milestones in breast cancer treatment history. Notably, Trastuzumab's efficacy isn't only from blocking HER2 signaling pathways — it also recruits NK cells to tumor sites through ADCC (antibody-dependent cellular cytotoxicity), letting NK cells recognize antibody-'labeled' cancer cells and kill them. This is a good example: antibody treatment and NK cells' natural killing function can work synergistically.
5. Common clinical tumor markers mostly come from tumor-associated antigens
You may have seen these names on health check reports: CEA (carcinoembryonic antigen), AFP (alpha-fetoprotein), PSA (prostate-specific antigen), CA125 (ovarian cancer marker), CA19-9 (pancreatic cancer related)... These tumor markers mostly belong to TAAs or carcinoembryonic antigens — they're also expressed in small amounts in normal tissues, but in corresponding cancers they're secreted in large amounts into blood, causing blood concentrations to significantly rise, so they can be used as screening and follow-up reference indicators.
But this also reveals the limitation of TAAs as treatment targets: they're not tumor-exclusive; normal tissues also express them (just less), so immunotherapy targeting them carries risks of harming normal tissue. This 'on-target toxicity' is a challenge many CAR-T and vaccine therapies based on TAAs face in clinical development. In contrast, therapies targeting neoantigens theoretically don't have this problem — neoantigens only exist in cancer cells, normal tissues simply don't have corresponding sequences.
6. From tumor antigens to individualized immunotherapy: a technology roadmap
Understanding tumor antigens unlocks understanding of how the most cutting-edge individualized cancer treatments are designed. Here's a simplified technology roadmap: whole exome sequencing (WES) of the patient's tumor tissue and normal tissue to find tumor-specific somatic mutations; bioinformatics algorithms predicting which mutations will produce neoantigen peptides that can bind to that patient's HLA type; synthesizing these neoantigen peptides (as vaccines), or using them to expand and activate the patient's own T cells in vitro (adoptive T cell therapy, ACT); giving the patient individualized vaccines or reinfusing expanded tumor-specific T cells, activating the patient's immune system to directionally attack cancer cells containing these neoantigens; simultaneously using immune checkpoint inhibitors to remove the tumor microenvironment's blockade of activated T cells.
This entire process would have seemed like science fiction ten years ago. Now, with dramatically falling sequencing costs and matured mRNA technology, it has entered clinical trial stages. From sequencing the tumor's mutations, to synthesizing immune weapons exclusive to your tumor, to activating your own immune system for precision attack — this is the closest attempt in medical history to 'medicine tailored specifically for you.'
7. The next step in tumor antigen research: beyond neoantigens
Neoantigen vaccine's rise is exciting, but the frontier of tumor antigen research is extending in deeper directions. Shared neoantigens are a concept receiving increasing attention. While most neoantigens are unique to each patient ('private neoantigens'), certain specific driver mutations (like KRAS G12D, TP53 R175H) appear commonly across many patients, producing identical neoantigens. 'Off-the-shelf' vaccines targeting these 'public neoantigens' don't need individualized customization, can cover patient groups carrying these specific driver mutations, greatly reducing cost and preparation time. mRNA vaccine platform flexibility makes this direction especially attractive.
Non-coding region neoantigens are another emerging direction. Traditional neoantigen prediction focused on exome (protein-coding region) mutations. But recent research found that non-coding region mutations (like introns, long non-coding RNAs) in tumors can also produce neoantigen peptides recognizable by T cells — and in some tumors, the number of non-coding region neoantigens even exceeds coding region neoantigens. Whole genome sequencing (WGS) replacing whole exome sequencing will be the key technical step for next-generation individualized vaccine precision upgrades.
Frequently Asked Questions