CELL INTELLIGENCE · 肿瘤细胞治疗研究

Background/Objectives: Glioblastoma multiforme (GBM) is an aggressive primary brain tumor with limited therapeutic options. Neoantigen-based immunotherapy offers a promising avenue, but its efficacy primarily depends on the ability of somatic mutations to generate immunogenic peptides effectively presented by HLA class I molecules and recognized by cytotoxic T cells, in concert with innate immune mechanisms such as NK-cell activation and DAMP/PAMP signaling.

This study aimed to characterize the MHC-I binding diversity of peptides derived from GBM-associated somatic variants, with a particular focus on interactions involving HLA-A68:01 and HLA-B15:01 alleles. These alleles were selected based on their ethnic prevalence and potential structural compatibility with neoepitopes. Methods: Somatic missense variants from TCGA-GBM were filtered using high-confidence genomic databases, including dbSNP, COSMIC, and MANE. Neoepitope prediction was performed across multiple HLA class I alleles using binding affinity algorithms (MHCflurry2). Peptide-HLA interactions were characterized through motif analysis and anchor residue enrichment. Structural modeling of peptide-HLA complexes was conducted using ColabFold (AlphaFold2-multimer v3) to evaluate conformational stability. The population frequency of selected HLA alleles was examined through epidemiological comparisons.

Results: Canonical GBM driver mutations (e. g. , EGFR, TP53, PIK3R1) are recurrent and biologically relevant, although pharmacological inhibition of EGFR alone has not consistently improved patient outcomes, underscoring the complex signaling redundancy in glioblastoma. HLA-A68:01 exhibited high binding affinity and favorable motif compatibility, supporting its potential for effective neoantigen presentation. HLA-B15:01 was identified as a viable presenter for the EGFR p.

Arg108Lys variant. Structural modeling confirmed stable peptide insertion into the MHC-I binding groove, with high-confidence folding and preserved interface integrity. Ethnic distribution analysis revealed varying GBM incidence across populations expressing these alleles. Conclusions: This integrative analysis identified structurally validated, immunogenically promising neoantigens derived from GBM mutations, particularly for HLA-A68:01 and HLA-B15:01.

These findings support allele-informed neoepitope prioritization in personalized immunotherapy, especially for patient populations with corresponding HLA genotypes and MHC-I presentation capacity.