Human antibody immune responses are personalized by selective removal of MHC-II peptide epitopes

Abstract

SummaryHuman antibody responses are established by the generation of combinatorial sequence diversity in antibody variable domains, followed by iterative rounds of mutation and selection via T cell recognition of antigen peptides presented on MHC-II. Here, we report that MHC-II peptide epitope deletion from B cell receptors (BCRs) correlates with antibody development in vivo. Large-scale antibody sequence analysis and experimental validation of peptide binding revealed that MHC-II epitope removal from BCRs is linked to genetic signatures of T cell help, and donor-specific antibody repertoire modeling demonstrated that somatic hypermutation selectively targets the personalized MHC-II epitopes in antibody variable regions. Mining of class-switched sequences and serum proteomic data revealed that MHC-II epitope deletion is associated with antibody class switching and long-term secretion into serum. These data suggest that the MHC-II peptide epitope content of a BCR is an important determinant of antibody maturation that shapes the composition and durability of humoral immunity.HighlightsAntibody somatic hypermutation selectively removes MHC-II peptide epitopes from B cell receptors.Antibodies with lower MHC-II epitope content show evidence of greater T cell help, including class-switching.MHC-II peptide epitope removal from a BCR is linked to long-term antibody secretion in serum.MHC-II genotype provides a personalized selection pressure on human antibody development.

Regulatory approved monoclonal antibodies contain framework mutations predicted from human antibody repertoires

“…Interestingly, peptides with > 2 K D fold-change from their unmutated germline peptides showed higher PSSM scores than mutations with < 0.5 K D fold-change, and these data were statistically significant. This analysis suggests that prevalent FR mutations are more likely to decrease the affinity of mAb peptide fragments for the MHC-II groove than to increase the peptide:MHC-II affinity at a given mutation site; these data were consistent with prior reports of natural repertoire antibodies using paired heavy and light chain human sequence datasets (20) ( Figure 5 ). As expected, mutations with moderate or no change in binding (0.5 ≤ K D fold change ≤ 2) showed the highest PSSM scores across all groups because the majority of mutations do not affect the ∼5-8 heavy chain variable region MHC-II peptide epitopes for a given MHC-II gene.…”

“…MHC-II peptide epitope affinity prediction was carried out for a set of 38 representative HLA-DRB1 alleles (20) using netMHCIIpan 4.0 (28). Predictions were performed using 15-mer peptide scans for each peptide containing mutations with an associated PSSM, and the netMHCIIpan flag -BA was selected to obtain predicted MHC-II binding affinities.…”

“…However, uncovering the rationale for every FR mutation is challenging as any individual mutation is pleiotropic, thus impacting many qualities in vivo including aggregation propensity, proteolytic stability, B cell receptor expression, antigen-binding affinity, and polyspecificity. Beyond these traits, it has also been hypothesized that engineering antibodies with FR regions derived from GL gene-preferred substitution patterns would result in lower immunogenicity as these antibodies would more closely reflect native antibody responses, which are readily recognized by the immune system (6,20). As an initial effort, we decided to focus on two properties which could reasonably explain the molecular basis behind these high scores: stability and MHC-II peptide epitope content.…”

“…Previous research has shown that unmutated antibody sequences have germline-encoded peptides that bind to and are presented by MHC-II proteins. These peptide epitopes are targeted for removal by SHM during B cell development in vivo (20). We sought to determine whether individual FR PSSM scores calculated here correlate with differential peptide:MHC-II binding affinities.…”

“…We restricted our analysis to the framework (FR) regions of the variable heavy (V H ) domain as antibody FRs impact in vivo stability, solubility, and immunogenicity (6) while also contributing significantly less than complementarity determining regions (CDRs) for binding antigen. We also explored some of the dynamics of peptide-MHC-II interactions using computational binding predictions (20), as the MHC-II peptide epitope contained within antibodies and other protein drugs has been recognized as an important component of clinical success (21,22). As a result, sequence information for FR regions can be applied to a broad array of antibodies with varying applications.…”