Potential Immunogenicity of Amino Acid Sequences Encoded by Ns-SNPs in Factor VIII,

Abstract 3329

Development of inhibitory antibodies that block factor VIII (FVIII) pro-coagulant function requires T-cell help. In order for T-cell stimulation to occur, one or more FVIII-derived peptides must first bind effectively to MHC Class II (e.g. HLA-DRB1) receptors on the surface of antigen presenting cells. If this Class II-peptide complex is recognized by a receptor on a helper T cell, then high-avidity Class II-peptide-T-cell receptor interactions can (but may not) cause proliferation and cytokine secretion leading to anti-FVIII antibody production. Thus the initial binding of a FVIII peptide to a Class II receptor can initiate an HLA-restricted cascade of events that leads to inhibitor formation in hemophilia A (HA) patients. Non-hemophilic, non-synonymous single nucleotide polymorphisms (ns-SNPs) in the human F8 gene were identified previously (Viel et al., Blood 109:3713–24, 2007) that result in the FVIII sequence variations R484H, R776G, D1241E and M2238V. Recombinant FVIII proteins with R484, R776, either D1241 or E1241, and M2238 are currently used therapeutically. The potential immunogenicity of FVIII peptides corresponding to these sequence variations is now being analyzed directly using in vitro assays employing cellular fractions of HA blood, recombinant proteins and synthetic peptides. These studies are motivated by a concern that some HA patients who express even very low levels of their endogenous, hemophilic FVIII, which should tend to promote immune tolerance to exogenous FVIII, could in principle develop inhibitors provoked solely by amino acid sequence mismatch(es) at sites in FVIII encoded by ns-SNPs. This could happen in patients with specific HLA types if their corresponding Class II receptors presented a FVIII peptide that then stimulated a helper T-cell response.

Synthetic FVIII peptides containing both alternate sequences encoded by these 4 ns-SNPs were tested for binding to the monomeric extracellular domains of 10 common DRB1 alleles using competitive binding assays. The results are as follows: (1) 484: R484 peptides had strong affinity (IC₅₀ < 1 mM) for DRB1*15:01 and DRB1*03:01; H484 peptides had moderate affinity (IC₅₀ = 1–10 mM) for DRB1*03:01 and DRB1*15:01; both R484 and H484 peptides bound with moderate affinity to DRB1*11:04 and DRB1*04:04; (2) 776: R776 and G776 peptides had moderate affinity for DRB1*09:01; (3) 1241: E1241 peptides had strong affinity for DRB1*09:01 and moderate affinity for DRB1*01:01, D1241 peptides bound weakly (IC₅₀ = 30–50 mM) to DRB1*09:01 and DRB1*01:01; (4) 2238: Both M2238 and V2238 peptides had strong affinity for DRB1*03:01 and moderate affinity binding to DRB1*01:01, DRB1*11:04, DRB1*11:01 and DRB1*04:04. Results of the binding assays compared reasonably well with computer-predicted affinities of the peptides for some of the better-characterized HLA-DRB1 alleles.

T-cell stimulation and MHC Class II tetramer staining experiments (which can unambiguously identify antigen-specific T cells) were next carried out using CD4 T cells from (1) 4 HA subjects (1 with a current inhibitor, 1 with a previous inhibitor and 2 without inhibitors) whose F8 gene encoded the V2238 variant and therefore had been treated with a therapeutic FVIII produce that was mismatched at this site with respect to the endogenous sequence encoded by their hemophilic F8 gene; and (2) from 4 HA subjects with the M2238 variant. All 8 patients expressed either the DRB1*03:01 or DRB1*11:01 alleles (one expressed both). No peptides corresponding to regions encoded by either version of this ns-SNP were recognized as T-cell epitopes by CD4 cells from these subjects, despite moderate to strong binding affinity of the peptides for DR0301 and DR1101 proteins. This suggests that the immunogenicity of these FVIII regions in HA patients who express these F8 and HLA genotype combinations may be less than might have been anticipated based on the binding data.

HA patients with HLA-DRB1 receptors that do not effectively bind peptides encoded by ns-SNPs in the F8 gene would be expected to have a low risk of developing helper T-cell responses to FVIII sequences encoded by these ns-SNPs. Thus, the negative as well as positive peptide-binding assay results have significant clinical and pharmaceutical relevance. Results of assays such as those described above could be useful in developing improved methods to assess the relative risk of individual HA patients developing an inhibitor.