Predicted impact of the viral mutational landscape on the cytotoxic response against SARS-CoV-2

Anna Foix-Romero, Anna Foix, Daniel López, Francisco Diez-Fuertes, Michael J. McConnell, Antonio J. Martin-GalianoPLOS Computational Biology 2022 · DOI

Introduction

While much of the public discourse surrounding SARS-CoV-2 variants has focused on "antibody escape," the cytotoxic T-cell (CD8+) response remains a critical pillar of long-term immunity and viral clearance. T-cells recognize infected cells by identifying viral fragments, or epitopes, presented by Human Leukocyte Antigen (HLA) molecules. Given the extreme diversity of HLA genes across the human population, the virus faces a significant challenge in evading the T-cell response globally. This study utilizes computational modeling to quantify the extent of T-cell escape across ~118,000 viral genomes and identifies specific human genetic backgrounds that may be disproportionately vulnerable.

Methodology

The researchers employed an immunoinformatic pipeline to bridge viral genomics with human immunogenetics. Using the netMHCpan 4.1 EL algorithm, they predicted potential HLA class I epitopes across the entire SARS-CoV-2 proteome for 2,915 human alleles. These alleles were grouped into 12 "supertypes" to represent 90% of the global population. The team then screened ~118,000 viral isolates from the NCBI database for mutations that significantly reduced predicted binding affinities, classifying these as "escape mutations."

Overview of the immunoinformatic pipeline used to identify T-cell escape mutations.

Results: The Epitope Landscape and Vulnerable Subpopulations

The analysis identified a total of 5,224 independent epitopes within the reference SARS-CoV-2 proteome. A key finding was the identification of "poor presenter" HLA families. While the average individual presents hundreds of viral fragments to their immune system, eight specific allele families (including A74, B46, and C*18) were found to recognize very few epitopes. For individuals carrying these alleles, even a single viral mutation in a key epitope could significantly impair their T-cell-mediated defense.

Average number of epitopes recognized across different HLA allele families.

Key Insights into Viral Escape

The study revealed that escape mutations have already affected approximately 47% of common "supertype" epitopes. Mutations are most disruptive when they occur at "anchor" residues (the 2nd or C-terminal positions), which are essential for locking the viral fragment into the HLA molecule. Furthermore, the researchers noted that large recurrent deletions, particularly in ORF1ab, serve as a "master key" for the virus, allowing it to eliminate multiple epitopes simultaneously.

Despite these findings, the overall risk to the global population remains relatively low. In any single viral isolate, mutations rarely bypassed more than 15% of the epitopes for a given supertype, suggesting that the inherent redundancy of the human immune system provides a robust buffer against total viral evasion.

Discussion and Limitations

The study found evidence of localized selective pressure, with specific escape mutations co-localizing in regions where the corresponding vulnerable HLA alleles are most prevalent. This suggests that the virus may be adapting to the specific genetic backgrounds of different ethnic groups.

However, several limitations must be considered. The results are based on in silico predictions which, while validated against experimental data from the Immune Epitope Database (IEDB), may not perfectly capture in vivo T-cell activation. Additionally, the study focused solely on HLA Class I (CD8+ T-cells) and did not account for the contributions of HLA Class II (CD4+ T-cells) or the humoral response. Future research should model the combined HLA repertoire of individuals to provide more precise risk assessments.

Summary updated

BibTeX
@article{foixromero2022predicted,
  author = {Anna Foix-Romero and Anna Foix and Daniel López and Francisco Diez-Fuertes and Michael J. McConnell and Antonio J. Martin-Galiano},
  title = {{Predicted impact of the viral mutational landscape on the cytotoxic response against SARS-CoV-2}},
  journal = {PLOS Computational Biology},
  year = {2022},
  doi = {10.1371/journal.pcbi.1009726},
}
Figure