Screening for Potential HLA-DRB1-Restricted T-Cell Epitopes in Factor VIII Using Peptide Microarrays

Therapeutic factor VIII (FVIII) infusions provoke CD4 T-cell proliferation and cytokine secretion leading to neutralizing anti-FVIII antibodies (inhibitors) in a substantial fraction of hemophilia A patients, leading to bleeding that is difficult to control. T-cell stimulation occurs when FVIII peptides bind to HLA-DRB1 proteins and are presented on the surface of antigen presenting cells, followed by T-cell recognition of these peptide-HLA-DRB1 complexes. Our laboratory is interested in identifying immunodominant, HLA-restricted T-cell epitopes in FVIII, in order to identify and then modify these immunological “hotspots” that cause inhibitors to develop. We have previously mapped several immunodominant HLA-restricted T-cell epitopes in FVIII recognized by the immune systems of inhibitor subjects, using HLA-DRB1 tetramers loaded with 20-mer FVIII peptides. However, because FVIII is such a large protein (>2,000 amino acid residues), methods more efficient than systematic mapping using overlapping peptides spanning its sequence would be useful. This study focuses on the binding of FVIII peptides to recombinant extracellular domains of 10 HLA-DRB1 proteins, which together represent HLA-DRB1 alleles found in >50% of the U.S population. PepStar^TM (JPT Peptide Technologies) microarrays containing triplicate sets of FVIII peptides immobilized on glass slides via a flexible linker were incubated with HLA-DRB1 proteins at several dilutions, followed by incubation with a fluorescent-labeled antibody that recognizes the peptide-bound conformations of HLA-DRB1 proteins, and the slides were washed and read on a microarray scanner. The intensities of fluorescent signals from the triplicate sets of spots were quantified and also displayed as heat maps. Fluorescent signals ranged from background levels (indicating no binding) up to the detector saturation range. Between 20-40% of the FVIII peptides produced fluorescent signals indicating peptide binding to each of the HLA-DRB1 proteins tested, including positive control peptides containing FVIII and non-FVIII epitopes with known HLA-DRB1 restriction. The percentage of apparently high-affinity binding interactions was much lower. T-cell epitopes are generally 10-14 amino acid residues long. Therefore, strong signals from pairs of 20-mer peptides with a 12-residue overlapping sequence localized the FVIII sequence fitting into the HLA-DRB1 binding groove to the overlapping region, while disparate signals from other pairs of overlapping sequences localized the HLA-DRB1 binding sequence to the non-overlapping regions. For subsets of the FVIII peptides, the signal strengths were compared to (a) affinities (IC50 values) derived from quantitative, competition peptide-HLA-DRB1 binding assays, and (b) results of computer prediction algorithms MHCPanII and Propred. Peptides that bound strongly to one or more HLA-DRB1 protein were prioritized for further testing. This microarray-based screening procedure is an efficient method to generate manageable lists of potential T-cell epitopes in large antigenic proteins such as FVIII, where systematic epitope mapping using HLA-DRB1 tetramers loaded with peptides spanning the entire sequence is not feasible, due to both the high cost of such experiments and the blood volumes that can reasonably be requested from patient volunteers.