Immunopeptidomic Data Integration to Artificial Neural Networks Enhances Protein-Drug Immunogenicity Prediction.

Carolina Barra,Jean-Philippe Goulet,Birkir Reynisson,Simon Comtois-Marotte,Morten Nielsen,Leon Eyrich Jessen,Jana Schockaert,Mark Watson,Sofie Pattijn,Anne Jang,Chloe Ackaert,Eustache Paramithiotis

doi:10.3389/fimmu.2020.01304

Abstract

Recombinant DNA technology has, in the last decades, contributed to a vast expansion of the use of protein drugs as pharmaceutical agents. However, such biological drugs can lead to the formation of anti-drug antibodies (ADAs) that may result in adverse effects, including allergic reactions and compromised therapeutic efficacy. Production of ADAs is most often associated with activation of CD4 T cell responses resulting from proteolysis of the biotherapeutic and loading of drug-specific peptides into major histocompatibility complex (MHC) class II on professional antigen-presenting cells. Recently, readouts from MHC-associated peptide proteomics (MAPPs) assays have been shown to correlate with the presence of CD4 T cell epitopes. However, the limited sensitivity of MAPPs challenges its use as an immunogenicity biomarker. In this work, MAPPs data was used to construct an artificial neural network (ANN) model for MHC class II antigen presentation. Using Infliximab and Rituximab as showcase stories, the model demonstrated an unprecedented performance for predicting MAPPs and CD4 T cell epitopes in the context of protein-drug immunogenicity, complementing results from MAPPs assays and outperforming conventional prediction models trained on binding affinity data.

Highlights

The advent of recombinant DNA technology in the last decades has boosted the use of protein drugs as pharmaceutical agents
After removing ligands with a common motif to Infliximab and Rituximab protein sequences, the remaining dataset was combined with single-allele binding affinity (BA) and mass spectrometry (MS) data collected from IEDB, to construct a dataset for training a model for HLA-DR antigen presentation prediction (Figure 1)
Our findings strongly suggest that the use of such prediction methods could effectively serve as a complement to MHC-associated peptide proteomics (MAPPs) assays to improve the sensitivity for identification of hotspot regions enriched in major histocompatibility complex (MHC) ligands and T cell epitopes

Summary

Introduction

The advent of recombinant DNA technology in the last decades has boosted the use of protein drugs as pharmaceutical agents. A major potential problem of these—compared to lower molecular weight pharmaceutical counterparts—is adverse effects associated with protein immunogenicity. Immunogenicity is generated because the drug is recognized as non-self, involving an unwanted activation of CD4 T cells, and the formation of anti-drug antibodies (ADAs), potentially producing a hypersensitivity reaction in treated patients. A small proportion of those peptides are loaded into major histocompatibility complex class II (MHC-II) molecules. Stable peptide-MHC-II complexes are exported to APCs’ surface for presentation to CD4 T cells, which can initiate, maintain, and regulate immune responses, including the production of ADAs [2]. Finely characterizing the rules of MHC-II binding and antigen presentation is of high interest to promote a general understanding of T cell immunogenicity and for the development of biotherapeutics

Objectives

Methods

Results

Conclusion