Mitigation of T-cell dependent immunogenicity by reengineering factor VIIa analogue

Wojciech Jankowski, Joseph McGill, H. A. Daniel Lagassé, Stepan Surov, Gary Bembridge, Campbell Bunce, Edward Cloake, Mark H. Fogg, Katarzyna I. Jankowska, Abdul Khan, Joseph Marcotrigiano, Mikhail V. Ovanesov and Zuben E. Sauna

Data supplements

Article Figures & Data


  • Figure 1.

    Overview of the computational and experimental strategies used in this study to design and evaluate an rFVIIa variant that has improved functional activity and lower immunogenicity risk.

  • Figure 2.

    In silico design of deimmunized variants. Deimmunization strategy based on algorithm described in supplemental Figure 1. Deimmunization was applied to the neo-epitope identified in VA generated by introducing neo-sequences during protein engineering. (A) Positions in green indicate a predicted decrease in promiscuity scores for the given amino acid substitution. (B) Amino acid substitutions with a decreased promiscuity score were evaluated for conservation score; substitutions depicted in orange indicate substitutions with a conservation score ≥0.6. (C) The amino acid substitutions with a decrease in promiscuity score at nonconserved regions were evaluated using EVmutation; substitutions with a score ≥−0.25 are shown in blue. (D) The Log2 fold change of percent rank of predicted HLA-DRB1 binding compared with VA for each HLA allele. (E) Surface representation of FVII mutant (PDB ID: 3ELA). The potential mutation sites for deimmunization identified using RID are represented as sticks.

  • Figure 3.

    T-cell responses. T-cell proliferation assay. (A) Cells from 50 donors were subjected to a 3H-incorporation T-cell proliferation assay. PBMCs from each donor were stimulated with peptides derived from WT FVIIa, VA, and the mutants N301D and N301E. Day 8 SI values are shown. Comparisons between the groups were made using 1-sided Fisher's exact test on the hypotheses that there were (1) more responders to VA than WT and (2) fewer responders to either N301D or N301E than VA. The test compared the number of measurements that would be considered responders using a SI threshold of 1.9. Each dot represents a single measurement per donor. (B) The mean SI for each donor (3 replicates) was calculated. A mean SI > 1.9 (dotted line) is considered a positive responder. (C-D) IL-2 ELISpot assay. PBMCs from the same donors used in the T-cell proliferation assay were stimulated with the same peptides and IL-2 secreting cells were assessed by ELISpot assay. The analysis from panels A and B were repeated for the ELISpot assay.

  • Figure 4.

    T-cell responses summary. A summary of donor responses, including high-resolution HLA-typing of the donors DRB1 allele, after restimulation with WT FVIIa, VA, and the mutants N301D and N301E peptides. Keyhole limpet hemocyanin (KLH) is included as a positive control. Only responders with SI > 1.9 (using 3 replicates) are shown.

  • Figure 5.

    The functional activity of FVIIa variants. (A) Thrombin generation assays carried out using the following purified proteins: WT FVIIa, VA, and the mutants N301E and N301D using hemophilia A plasma (ie, plasma without functional FVIII). (B) The dose effect of FVIIa concentrations (from 0.03 to 1600 nM) for variants described in panel A. (A-B) The assay was carried out in the presence of phospholipid vesicles (4 µM) either in the presence (left) or absence (right) of TF (0.5 pM). (C) Effect of PL concentration (left, PL range: 0, 1-16 µM) and TF concentrations (right, TF range: 0, 0.1-25 pM) on thrombin generation by the FVIIa variants (FVIIa concentration: 50 nM) described in panel A.

  • Figure 6.

    Molecules with desired functional activity and lower immunogenicity risk. (A) Correlation of preclinical data with clinical outcomes. (B) The relation between activity and T-cell response in all evaluated FVII variants. N301E and N301D exhibit desired characteristics for a therapeutic-protein (ie, low T-cell response and activity comparable to the VA). T-cell responses were measured using 2 independent experiments of 50 donors each. In experiment 1, PBMCs from 50 donors were exposed to WT-rFVII and VA sequences. In experiment 2, PBMCs from 50 donors were exposed to WT-rFVIIa, VA, and deimmunized sequences (N301E and N301D). Activities for each FVII variant are represented as a mean calculated from 2 independent protein preparations using VA as a standard (100%).