Background: Why high expression alone is not enough to select a vaccine antigen

A systematic review of 16 antigen candidates for pancreatic cancer mRNA vaccines concluded that the strength of evidence and priorities for further validation differ across candidates. The paper compares antigens proposed in earlier studies and outlines criteria for considering multi-antigen designs; it does not demonstrate the therapeutic efficacy of a new vaccine. It was published in Frontiers in Immunology.

In pancreatic cancer, dense stroma, an immunosuppressive tumor microenvironment, and inefficient antigen presentation hinder the antitumor activity of vaccines. Approaches targeting patient-specific neoantigens or KRAS mutations are also under study, but the burden of individualized production and limited patient eligibility remain challenges. Identifying antigens applicable to multiple patients requires assessing not only high tumor expression but also whether immune cells actually recognize them and whether they pose a risk of attacks on normal tissue.

Key findings: 16 candidates grouped into three evidence tiers

The authors synthesized 16 antigen candidates described in four core studies selected after deduplication and screening of 109 PubMed records and 166 Web of Science records. The included studies comprised retrospective analyses of public databases and basic experimental research. The review compared tumor expression, associations with prognosis, correlations with antigen-presenting cell infiltration, and immunogenicity data. It also examined links to ferroptosis, a form of cell death associated with iron-dependent lipid peroxidation, and pyroptosis, an inflammatory form of cell death. However, a correlation with infiltration does not establish that a candidate antigen can recruit immune cells.

The candidates the authors classified as having relatively substantial evidence were MET, SCP-1, GAGE, PAK2, and TPX2. Some had data on antigenic regions recognized by T cells or on immune responses, but these findings cannot be taken as proof of clinical efficacy or safety for pancreatic cancer mRNA vaccines. SCP-1 and GAGE, which are cancer-testis antigens, have also been reported to be expressed in chronic pancreatitis tissue. They should not be described as safe targets found only in cancer cells.

ADAM9, EFNB2, KDM5A, ANO6, RAB5A, and TMOD3 were grouped as candidates functionally associated with tumor progression or cell-death pathways. Direct evidence of immunogenicity is limited, so the effects of suppressing a gene’s function cannot be interpreted as the effects of a vaccine targeting that antigen. OSBPL9, NRAS, CHMP2B, WNT7A, and AGPS were exploratory candidates supported mainly by bioinformatic analyses or indirect associations. The review judged the evidence insufficient to treat them as immediate clinical development targets.

Significance and outlook: Multi-antigen combinations remain design hypotheses to test

The paper proposed examples such as adding PAK2 and TPX2 to MET, or combining MET with ADAM9 and KDM5A. The rationale is to broaden immune recognition and address tumor heterogeneity. It did not present results directly comparing the superiority of the proposed combinations or demonstrating that a vaccine containing all 16 candidates prevented immune escape and changed the immunosuppressive tumor environment.

Key next steps are to determine whether peptides derived from candidate antigens bind to and are presented by human leukocyte antigen (HLA) molecules, whether they induce T-cell responses that recognize pancreatic cancer, and whether those responses also target normal tissue. Coverage across patients’ HLA types, mRNA delivery efficiency, dense stroma, and the immunosuppressive environment also remain challenges. The review’s contribution is to clarify the relative strength of evidence for each candidate and suggest priorities for future experiments.

The study types, candidate groupings, and limitations are described in the full paper and the bibliographic record and abstract.