Peptide-Pulsed Dendritic Cells: Why They Beat mRNA on CD8+ Response

The 2024 mRNA-4157 readout in melanoma reframed how the oncology field thinks about personalized neoantigen vaccines, and much of the subsequent attention has flowed to mRNA platforms. A parallel 2025 analysis led by investigators at Mount Sinai reintroduced peptide-pulsed autologous dendritic cells into the conversation with a specific claim: for CD8+ T-cell priming against HLA class I-restricted neoantigens, direct peptide loading onto the DC surface often outperforms mRNA-encoded antigen on a per-epitope basis [1].

This article unpacks that claim mechanistically, summarizes the comparative per-epitope CD8 response data, and explains why manufacturing and indication biology will likely keep both modalities relevant through the late 2020s.

In a peptide-pulsed DC protocol, autologous monocyte-derived or natural DC-enriched cell preparations are incubated ex vivo with short (8-11 amino acid) synthetic peptides corresponding to predicted HLA class I-binding neoantigens. At the relevant peptide concentrations, empty or peptide-receptive MHC class I molecules at the DC surface can directly bind the exogenous peptides without requiring internal processing [2].

• Step 1: Leukapheresis of patient peripheral blood
• Step 2: Monocyte isolation and differentiation into DCs
• Step 3: Maturation cocktail to drive CCR7 and CD80/CD86 upregulation
• Step 4: Pulsing with patient-specific short peptides predicted to bind the patients HLA class I alleles
• Step 5: Reinfusion — DCs traffic to lymph nodes and prime CD8+ T cells

The net effect is that CD8+ T cells in the lymph node encounter the neoantigen on a professional APC already expressing high levels of co-stimulatory molecules, without a cross-presentation step that can introduce patient-to-patient variability.

mRNA neoantigen vaccines such as mRNA-4157 encode a concatenated string of neoantigen sequences on a single lipid-nanoparticle-delivered mRNA. After intramuscular injection, APCs at the injection site and draining lymph node translate the mRNA, produce the polyprotein, process it through the proteasome, and present the resulting peptides on MHC-I via TAP-mediated loading in the endoplasmic reticulum [3].

This pathway is biologically elegant and broadly scalable, but it introduces two variables: (1) the efficiency of cytoplasmic processing of the encoded polyprotein, and (2) the efficiency of TAP transport and ER loading, both of which can vary across patients and cell states.

Direct head-to-head randomized comparisons between peptide-pulsed DCs and mRNA neoantigen vaccines are rare, but per-epitope CD8 response rates can be compared across separately conducted studies by pooling ELISpot, tetramer, and intracellular cytokine staining readouts [4].

These ranges are illustrative rather than definitive, since patient populations, epitope prediction pipelines, and assay sensitivities vary across studies. The consistent observation is that direct peptide loading onto DC MHC-I is often at or above mRNA on a per-epitope basis when the delivery is executed well.

Cross-presentation is the pathway by which exogenous antigens are delivered into the MHC-I presentation pipeline. It is a highly specialized capability concentrated in specific DC subsets (notably cDC1 cells expressing XCR1 and CLEC9A) and is not a given across every APC that encounters an antigen [5]. Both SLP vaccines and mRNA vaccines depend on cytoplasmic access of the antigen, whereas short-peptide-pulsed DCs bypass this step entirely by loading MHC-I from the outside.

The practical consequence is that peptide-pulsed DC protocols are less sensitive to patient-to-patient variability in cross-presentation capacity, which may be particularly relevant for patients whose cDC1 compartment is compromised by disease biology or prior therapy.

The key structural difference is that peptide-pulsed DC products are autologous cell therapies, which brings cell-therapy-grade logistics (apheresis centers, GMP cell manufacturing, patient-matched release), while mRNA products are individualized biologics that leverage biologic manufacturing infrastructure. Neither is obviously easier at scale — each imposes a different operational footprint [6].

Several patient-level factors may favor one modality over the other. Peptide-pulsed DCs are well-suited to indications where cDC1-dependent cross-presentation may be compromised, where co-stimulation needs to be tightly controlled, or where adjuvant-related reactogenicity limits other formats. mRNA platforms are well-suited to indications that benefit from rapid re-dosing, broader epitope coverage per dose, and integration with checkpoint inhibitor backbone regimens [7].

• HLA class I-focused CD8+ priming: direct peptide loading on DCs has mechanistic advantages
• Broad MHC coverage including class II: mRNA and SLP platforms have advantages
• Rapid redosing: mRNA logistics favor repeated dosing schedules
• Cell-therapy infrastructure available: DC platforms feasible in specialized centers
• Apheresis not feasible: mRNA platforms more accessible

The 2026 picture of personalized cancer vaccines is not one of single-platform dominance. Peptide-pulsed DCs, synthetic long peptides, short-peptide HLA-I-focused formats, and mRNA neoantigen vaccines address overlapping but distinct corners of the immunological design space. Different combinations — DC priming followed by mRNA boost, SLP for CD4 help combined with DC or mRNA for CD8 priming — are under active investigation [8]. The Mount Sinai-led analysis supports a multi-modal future rather than a winner-takes-all outcome.

Several caveats apply to any cross-study peptide-DC vs mRNA comparison: epitope prediction pipelines are not standardized across programs, assay platforms differ in sensitivity, patient populations vary in tumor mutational burden and prior therapy, and adjuvant / maturation conditions are not uniform. The claim that peptide-pulsed DCs outperform on CD8+ response rates holds in matched comparisons but should not be over-extrapolated to clinical efficacy outcomes, which depend on many additional variables beyond CD8 priming.