Peptide Proof — Evidence-Based Peptide Intelligence. Independent analysis. No paywall.
Here’s What Happened
Neoantigen peptide vaccines — individualized cancer vaccines built from tumor-specific mutant peptides — have progressed from a scientific curiosity to a clinical reality in under a decade. In 2025, Moderna and Merck reported that their mRNA-4157 neoantigen vaccine combined with pembrolizumab reduced the risk of recurrence or death by forty-nine percent in high-risk melanoma. Peptide-based neoantigen vaccines, while lagging mRNA approaches in speed, offer advantages in manufacturing simplicity, regulatory precedent, and cost that position them for a distinct and potentially larger role in the immunotherapy landscape.
How Neoantigen Vaccines Work
The premise is elegant: sequence a patient’s tumor DNA, identify mutations that generate novel peptide sequences not present in normal tissue (neoantigens), synthesize peptides corresponding to those mutations, and administer them to train the patient’s immune system to recognize and destroy tumor cells bearing those mutations.
The workflow has four stages: tumor biopsy and sequencing (1–2 weeks) → bioinformatic neoantigen prediction — identifying which mutations generate peptides that bind the patient’s MHC molecules and are likely to be immunogenic (1 week) → peptide synthesis — typically 10–20 peptides, 15–30 amino acids each, synthesized by solid-phase peptide synthesis → formulation and administration with an immune adjuvant such as poly-ICLC — 1 week —. Total turnaround time from biopsy to first dose is 5–8 weeks — longer than mRNA vaccines (4–6 weeks) but well within the window for adjuvant therapy in most solid tumor settings.
Clinical Evidence: 2020–2026
Three randomized trials have established proof-of-concept for neoantigen peptide vaccines:
NeoVax — Dana-Farber/Broad Institute —. A Phase I trial of a personalized neoantigen peptide vaccine (up to 20 long peptides with poly-ICLC) in 8 patients with high-risk melanoma, published in Nature (2017), showed that all patients generated neoantigen-specific CD4+ and CD8+ T-cell responses. At 4-year follow-up (2021), 6 of 8 patients remained disease-free — a result that launched the field.
NEO-PV-01 — Neon Therapeutics/BioNTech —. A Phase Ib trial combining a neoantigen peptide vaccine with nivolumab in advanced melanoma, NSCLC, and bladder cancer, reported in Cell (2020), demonstrated that vaccination induced de novo neoantigen-specific T-cell responses in all three tumor types. The epitope spreading observed — T-cell responses against neoantigens not included in the vaccine — suggested that peptide vaccination can broaden the anti-tumor immune response beyond the initially targeted mutations.
EVX-01 — Evaxion Biotech —. A Phase IIa trial combining an AI-designed neoantigen peptide vaccine with pembrolizumab in metastatic melanoma reported a sixty-seven percent objective response rate at ASCO 2025, compared to approximately forty percent for pembrolizumab alone in historical controls. The trial was small (n=28) but provided the strongest efficacy signal to date for a peptide-based neoantigen approach.
Peptide vs. mRNA: A Strategic Comparison
Manufacturing | solid-phase peptide synthesis | In vitro transcription (newer).
Regulatory precedent | Extensive — peptide drugs — | Limited — COVID vaccines —.
The comparison reveals complementary strengths. mRNA vaccines offer faster turnaround and the ability to encode more neoantigens, but they require complex cold-chain logistics and carry higher manufacturing costs. Peptide vaccines are slower but benefit from a century of peptide chemistry experience, simpler regulatory pathways,. lyophilized formulations that can be shipped and stored at ambient temperature — a critical advantage for global deployment outside major academic medical centers.
Expert Insight: The Bioinformatic Bottleneck
The single greatest challenge in neoantigen vaccine development is not peptide synthesis or clinical trial design — it is neoantigen prediction. Current algorithms identify 100–500 candidate neoantigens per patient, but only 1–five percent of predicted neoantigens are actually immunogenic. Every false-positive neoantigen included in a vaccine competes for immune attention with true immunogenic neoantigens, potentially diluting the therapeutic effect.
The prediction problem has three dimensions: MHC binding prediction (how well does the mutant peptide bind the patient’s MHC molecules?), TCR recognition prediction (will the patient’s T-cell repertoire recognize the MHC-peptide complex?), and tumor presentation prediction (is the mutant peptide actually presented on the tumor cell surface at sufficient density?). Current algorithms handle MHC binding well, also known as AUC > 0.9 but perform poorly on TCR recognition — AUC ~ 0.6–0.7 — and tumor presentation (AUC ~ 0.5–0.6). Companies that crack the TCR recognition problem — likely through AI trained on paired TCR–peptide–MHC data — will define the next generation of neoantigen vaccines.
[Natural sign-off — one sentence summary of why this matters.]
Further Reading
- Peptide-Drug Conjugates — another peptide-based approach to oncology
- AI-Designed Peptides — the algorithms that could improve neoantigen prediction
Last reviewed: June 2026. Peptide Proof Editorial Team.
