Population HLA differences had minimal impact on cancer-testis antigen vaccine targets but significantly affected neoantigen-based personalized vaccine design.
Inverse relationshipmore common cancer mutations have fewer predicted immune targets — complicating mutation-based vaccine design
What the researchers found
The study analyzed 6 cancer-testis antigens (CTAs) and 95 common cancer mutations for their ability to produce peptides that bind HLA class I molecules in two populations.
For CTAs, the overall epitope prediction difference between white and East Asian populations was small. There was a linear relationship between peptide length and epitope occurrence, meaning longer fragments produced more potential targets.
For mutation-derived neoantigens (from missense mutations), population differences were larger, reflecting different HLA allele distributions.
The most striking finding: mutations with the highest incidence in cancer patients had the lowest predicted epitope occurrence. Conversely, rare mutations had the most immune-visible peptides. This inverse relationship suggests immunosurveillance: the immune system kills tumors with easily detectable mutations, so only mutations that evade immunity become common.
Frameshift/indel mutations fell between CTAs and point mutations in the peptide length-epitope relationship.
Why it matters
Cancer vaccines need to target peptides that the patient's immune system can actually recognize. This study shows population-specific HLA differences matter more for neoantigen vaccines than for CTA vaccines. The finding that common mutations are immunologically invisible explains a major challenge in cancer immunotherapy and guides target selection.
The numbers in context
6 CTAs; 95 mutations; inverse mutation incidence/epitope relationship; small CTA population differences; larger neoantigen differences
How the study worked
Computational immunology study. Six CTAs commonly used in immunotherapy and 95 hotspot mutations from The Cancer Genome Atlas were analyzed. HLA class I binding affinity was predicted for the most common alleles in white and East Asian populations. Epitope occurrence was correlated with mutation incidence.
Who was studied
Computational analysis: 6 CTAs + 95 mutations across white and East Asian HLA databases
What this study cannot tell us
Entirely computational. Predicted binding does not guarantee immunogenicity (actual immune response). HLA binding is necessary but not sufficient for T cell activation. The analysis used common HLA alleles, which do not cover all patients. Real-world cancer vaccine efficacy depends on many factors beyond HLA binding.
How to read the evidence
Preliminary evidence — entirely computational predictions. Binding predictions do not guarantee actual immune responses.
When this study was published
Published in 2020. Personalized cancer vaccine clinical trials have expanded significantly since.
The bigger picture
Personalized cancer vaccines need to match each patient's HLA type. This study shows that shared cancer antigens (CTAs) may work across populations, but mutation-based vaccines need population-specific optimization — important for global cancer vaccine equity.
Questions still open
- Does the inverse mutation-immunogenicity relationship mean common mutations are harder to vaccinate against?
- How do other populations (African, South Asian) compare?
- Can multi-epitope vaccines overcome HLA diversity challenges?
Common questions
Why do population differences matter for cancer vaccines?
Are universal cancer vaccines possible?
Read the original research
HLA class I restricted epitopes prediction of common tumor antigens in white and East Asian populations: Implication on antigen selection for cancer vaccine design.
PloS one, 15(2), e0229327
Citation
Hu, Wei; He, Meifang; Li, Liangping. (2020). HLA class I restricted epitopes prediction of common tumor antigens in white and East Asian populations: Implication on antigen selection for cancer vaccine design.. PloS one, 15(2), e0229327. https://doi.org/10.1371/journal.pone.0229327