Hemophilia B Mouse Lines Doubly Humanized for Human F9 and MHC Class II Genes Advance Specificity of Inhibitor Study

Abstract 200

The development of neutralizing alloantibodies (inhibitors) against factor IX (FIX) in humans with hemophilia B is a potentially devastating complication; clinical investigation and evidence-based intervention for FIX inhibitors is complicated by their rarity. Current hemophilia B mouse models are not ideal for investigating the potential immunogenicity of hemophilia B therapies due to differences in mouse and human MHC class II structure and FIX epitope recognition. This recognition depends on CD4+ T cells and their direct interaction with surface-expressed MHC class II molecules complexed with antigen. We have developed a series of mouse lines that do not express mouse FIX or mouse MHC II, but instead express human FIX genes and humanized MHC II. The potential immunogenicity of human FIX and specific hFIX epitopes for CD4+ T cell recognition have been characterized in each of these “double humanized” (2H) mouse lines.

Our derivation of the 2H mouse lines, all having first been bred through a complete knockout of mouse MHCII (mMHC II^−/−), resulted in four lines of human HLA-DRB1*1501 mice. These four lines include 1) mouse FIX^−/− (complete deletion; Cross-reacting material negative (CRM-)); 2) human R29X (expresses human FIX gene with early nonsense mutation; CRM-); 3) human R333Q (expresses human factor IX defective circulating protein with missense mutation; CRM+), and 4) human WT FIX (WT2 expresses hemostatic human FIX; CRM+). All lines were first shown to have similar numbers of MHCII and CD4+ expressing T cells in peripheral blood based on flow cytometry when compared with conventional (mouse MHC H2^b) hemophilia B mice. To determine whether (hu-) mice have the same incidence and degree of immune response to hFIX compared to conventional hemophilia B mice, we immunized subcutaneously with hFIX emulsified in CFA. Importantly, control mMHC II^−/− mice did not develop anti-FIX antibodies. Both strains of CRM- 2H hu-mice developed inhibitors and anti-hFIX IgG antibodies with similar intensity. CD4+-enriched splenocytes proliferated in vitro in response to hFIX challenge and secreted high amount of Th1-derived IFN-γ and moderate Th2 (IL-4 and IL-10) cytokines. However, when compared to the inhibitor-prone strains, conventional (mouse MHC H2^b) and hu-DRB1*1501 CRM+ mice (both R333Q and WT2) tolerated hFIX, had larger total Treg populations, increased TGF-β1 cytokine, and demonstrated regulatory apoptosis of CD4+ Teffector cells. Additional lymphocyte response studies mapped four peptides derived from FIX catalytic domain as responsible for anti-FIX antibodies in the CRM- lines. (CD4+ T cells from CRM+ animals (hu-R333Q or hu-WT2 mice) did not proliferate to these four or any other hFIX peptides). The peptides identified in 2H hu- mice differed from immunostimulatory peptides identified in reported epitope mapping studies using FIX^−/− mice having mouse MHC H2^b, H2^d, and H2^kbackgrounds. These differences may reflect the fundamental structural and epitope recognition differences between murine and human MHC class II complexes.

Finally, we have examined systemic and muscle-directed AAV-mediated gene therapy expressing either WT hFIX or a hFIX with a single amino acid gain of function variant (FIX Padua, R338LFIX). We have confirmed that expression of this single amino acid change does not increase inhibitor risk, as modeled in this double humanized hemophilia B model. This observation appears consistent with the extreme rarity of clinical inhibitor development in patients having underlying FIX missense mutations. In summary, the series of double humanized hemophilia B mouse models have hFIX recognition profiles that make them important tools for preclinical evaluation of novel hemophilia B therapeutics and to advance understanding of FIX inhibitors.