Integrative Neoepitope Discovery in Glioblastoma via HLA Class I Profiling and AlphaFold2-Multimer

Area of Science:

• Oncology
• Immunology
• Genomics

Background:

• Glioblastoma multiforme (GBM) is an aggressive brain tumor with limited treatment options.
• Neoantigen-based immunotherapy shows promise but relies on effective presentation of mutated peptides by Human Leukocyte Antigen (HLA) class I molecules.
• Understanding peptide-HLA interactions is crucial for developing effective immunotherapies.

Purpose of the Study:

• To characterize the binding diversity of peptides derived from GBM somatic variants.
• To focus on interactions with HLA-A68:01 and HLA-B15:01 alleles for neoantigen presentation.
• To identify structurally validated neoantigens for personalized glioblastoma immunotherapy.

Main Methods:

• Filtered somatic missense variants from TCGA-GBM using genomic databases (dbSNP, COSMIC, MANE).
• Predicted neoantigens across HLA class I alleles using binding affinity algorithms (MHCflurry2).
• Performed structural modeling of peptide-HLA complexes using ColabFold (AlphaFold2-multimer v3) and analyzed population frequencies of HLA alleles.

Main Results:

• Identified recurrent GBM driver mutations (e.g., EGFR, TP53) with complex signaling.
• HLA-A68:01 demonstrated high binding affinity and motif compatibility for neoantigen presentation.
• HLA-B15:01 effectively presented the EGFR p.Arg108Lys variant, with stable structural modeling of peptide-HLA complexes.

Conclusions:

• Structurally validated, immunogenic neoantigens from GBM mutations were identified for HLA-A68:01 and HLA-B15:01.
• Findings support prioritizing neoepitopes based on HLA genotype for personalized immunotherapy.
• This approach can enhance T-cell recognition and treatment efficacy in specific patient populations.