The Need for Cancer Vaccines Non-small cell lung cancer (NSCLC) accounts for approximately 85% of lung cancer cases and carries a poor 5-year survival rate. Conventional therapies including surgery, chemotherapy, and checkpoint inhibitors have significant limitations including resistance and restricted efficacy to specific patient subgroups. Neoantigen-based vaccines represent a promising personalized immunotherapy approach.
Multi-Neoepitope Vaccine Concept A multi-neoepitope vaccine (MNEV) targets multiple tumor-specific antigens arising from cancer cell mutations simultaneously. By incorporating B-cell, cytotoxic T lymphocyte (CTL), and helper T lymphocyte (HTL) epitopes into a single construct, the vaccine aims to activate both humoral and cellular arms of the immune system against the tumor.
Reverse Vaccinology Approach Reverse vaccinology uses genomics and bioinformatics to identify vaccine targets computationally before any laboratory work, dramatically reducing development time and cost. This study applied whole exome sequencing, RNA sequencing, and a suite of epitope prediction tools to design an MNEV targeting two NSCLC cell lines: murine LL/2 (LLC1) and human A549.
Validation Strategy The designed MNEV was validated through in silico structural and physicochemical analysis, molecular docking against MHC molecules and toll-like receptors, and then tested in vivo in C57BL/6 mice, measuring IgG antibody levels, T cell activation, and cytotoxic granzyme B production.
Whole Exome and RNA Sequencing Publicly available WES and RNA-seq data from LL/2 (murine) and A549 (human) NSCLC cell lines were downloaded from NCBI. Reads were aligned to mouse (GRCm38) and human (GRCh38) reference genomes using BWA-MEM for WES and STAR for RNA-seq data.
Somatic Mutation Calling MuTect and MuTect2 were used to identify somatic single nucleotide variants in tumor vs. normal comparisons. Initial analysis revealed 962 missense mutations in LL/2 cells and 2,003 in A549 cells. Variant annotation was performed using VEP (Variant Effect Predictor).
Neoantigen Selection with Isovar The isovar tool integrated RNA-seq data to identify mutation-supporting reads and determine the actual expressed mutant coding sequences, accounting for splicing diversity and tumor heterogeneity. Only mutations with FPKM greater than 1 - indicating active expression - were retained as neoantigen candidates.
Shared Mutations A Venn diagram analysis of LL/2 data across seven datasets revealed that 40% of substitution mutations are recurrent across datasets, supporting the biological relevance of these neoantigens as stable tumor-associated targets rather than sequencing artifacts.
CTL Epitope Prediction MHC class I binding epitopes were predicted using four algorithms: NetMHCpan4.1, NetCTLpan, NetMHCcons, and SYFPEITHI. Epitopes were filtered by binding affinity (IC50 less than 500 nM), TAP transport scores, proteasomal cleavage scores, and antigenicity via VaxiJen v2.0. Five CTL epitopes were selected for LL/2 and eight for A549.
HTL Epitope Prediction MHC class II binding epitopes for CD4+ helper T cells were predicted using RANKPEP, NetMHCIIpan-4.0, IEDB, MHC2pred, and NetMHCII-1.1. Epitopes were also assessed for cytokine induction potential (IL-4, IFN-gamma, IL-10) using dedicated prediction servers.
B-Cell Epitope Prediction Linear B-cell epitopes were predicted using ABCpred and BepiPred-2.0. Conformational (discontinuous) B-cell epitopes were predicted using ElliPro, with IgPred confirming their capacity to generate IgG antibodies. All selected epitopes were verified to be non-allergenic (AllerTOP), non-toxic (ToxiPred), and antigenic (VaxiJen).
Vaccine Assembly The final MNEV construct assembled all epitopes using specific linkers: KK for B-cell epitopes, GPGPG for HTL epitopes, and ADARY for CTL epitopes. The 50S ribosomal protein L7/L12 was incorporated as an adjuvant via an EAAAK linker. The immunoglobulin kappa (Igk) signal peptide improved secretion, and a His6 tag enabled purification. The final construct contains 472 amino acids.
3D Structure Prediction The MNEV's 3D structure was predicted using I-TASSER, the leading server for protein structure prediction, achieving a confidence C-score of -1.02 with a TM score indicating reliable fold prediction. The structure was refined using GalaxyRefine and validated with ProSA-web and Ramachandran plot analysis.
Physicochemical Properties ProtParam analysis of the 472 amino acid vaccine showed a molecular weight of approximately 51,447 g/mol, theoretical pI of 6.46, aliphatic index of 74.53 (indicating thermostability), GRAVY score of -0.546 (hydrophilic), and an instability index below 40, confirming the construct's stability.
TLR Docking Molecular docking with ClusPro assessed interactions between the MNEV and toll-like receptors TLR3, TLR4, and TLR9. The best cluster for each showed binding energies of -924.8 kcal/mol (TLR3), -1043.8 kcal/mol (TLR4), and -958.5 kcal/mol (TLR9). MM-GBSA free energy calculations confirmed binding free energies of -86.4, -313.26, and -89.38 kcal/mol respectively, with TLR4 showing the strongest binding.
In Silico Codon Optimization The MNEV coding sequence was codon-optimized using JCat for high expression in eukaryotic systems, achieving a codon adaptation index (CAI) of 0.99 and GC content of 53.4%, and was cloned in silico into the pLenti-GIII-CMV-GFP-2A-Puro vector.
Antibody Response Female C57BL/6 mice immunized with the lentiviral MNEV (MNEV-LV, 2x10^7 infectious units subcutaneously on days 0, 14, and 28) showed significantly elevated total serum IgG levels compared to empty lentiviral vector (empty LV) and PBS mock controls two weeks after the final injection. No significant difference was found between the two control groups.
IFN-gamma Secretion Splenocytes from MNEV-immunized mice secreted significantly elevated IFN-gamma upon restimulation with MNEV-conditioned medium from transduced mesenchymal stem cells (MSC-MNEV-CM). This confirms Th1 immune activation with the cytokine profile needed for effective anti-tumor cytotoxic responses.
T and B Cell Expansion Flow cytometry revealed that MNEV immunization significantly increased the proportions of CD3+CD4+ helper T cells, CD3+CD8+ cytotoxic T cells, and CD19+ B cells in splenocytes compared to controls. This comprehensive immune cell activation demonstrates engagement of both adaptive cellular and humoral immunity.
Granzyme B Secretion MNEV-immunized mice showed substantially higher granzyme B levels in splenocyte supernatants compared to both control groups. Granzyme B is a key cytotoxic molecule released by activated CD8+ T cells to kill target cells, confirming that the vaccine induced functional cytotoxic T lymphocyte responses with direct tumor-killing potential.
Preclinical Stage All results are from cell line analysis and a murine model, which does not guarantee efficacy in human NSCLC patients. The vaccine must be tested in established syngeneic mouse tumor models (such as subcutaneous LL/2 implantation) to confirm tumor growth inhibition before clinical translation.
Tumor Challenge Studies Needed While the vaccine induced immune responses in healthy mice, tumor protection experiments directly testing whether MNEV-immunized mice resist lung tumor growth are an essential next step. The immunogenicity data shown here does not by itself prove tumor-killing efficacy in vivo.
Human-Specific Vaccine Refinement The MNEV includes both mouse (LL/2) and human (A549) epitopes. A fully clinical-grade human MNEV would require personalization based on individual patient tumor mutational profiles, as neoantigens are by definition patient-specific. The computational pipeline developed here is well suited for this future personalized application.
Combination with Checkpoint Blockade Combining the MNEV with anti-PD-1 or anti-PD-L1 checkpoint inhibitors could synergize by both generating tumor-specific T cells (vaccine) and releasing their inhibitory brakes (checkpoint blockade). This combination strategy, already validated for other neoantigen vaccines in early clinical trials, is a logical next step for this MNEV design.