In our previous studies of genetic factors that influence the immune response (particularly inhibitor antibodies) to therapeutic coagulation factors we have shown that H2 and cytokine genes control the antibody response to human FIX protein in mice (

Lozier et al.,
). We have extended our investigation to the murine resonse to human FVIII. Based on the observation that FVIII K/O mice in Balb/cByJ and C57BL/6ByJ backgrounds have significantly different antibody responses to human FVIII after gene transfer (
Brown et al.,
J Thromb Haemost
) we studied the antibody response to human FVIII in these mice and 13 CXB recombinant inbred mouse lines derived from these parental strains. Balb/cByJ mice were significantly higher responders to human FVIII than C57BL/6ByJ mice after a single injection with an adenovirus vector (AVPGKf8mg) expressing human FVIII (P < 0.001) or serial intravenous injections of human FVIII protein sans adjuvant (P < 0.01). For both strains the difference between male and female antibody responses following adenovirus or FVIII protein injections were not significantly different. Phage display mapping with 7-mer looped and 12-mer linear libraries showed reactivity against epitopes in the A1, A2, B, and C1 domains of human FVIII as well as the acidic region between the A1/A2 domains, after administration of FVIII protein. Analysis of antibody titers to human FVIII in male CXB recombinant inbred mice after weight-adjusted doses of AVPGKF8mg vector revealed suggestive linkage-relatedness scores for a series of polymorphic markers on mouse chromosome 8 that are 40–45 centimorgans from the centromere. These include D8mit 240 (LOD score 3.1), D8mit45 (LOD score 3.1), and D8mit84 (LOD score 3.0). Each marker had an equal additive effect to the antibody response. In female CXB recombinant inbred mice different markers were associated with the antibody response to FVIII than in males. For female CXB recombinant inbred mice there were suggestive linkage relatedness scores for polymorphic markers on mouse chromosome 17 (D17mit16, LOD score 3.0) and mouse chromosome 5 (D5mit31, LOD score 2.4). Both markers had equal additive effects on the antibody response to human FVIII. D17mit16 is adjacent to the mouse H2 gene complex, an obvious candidate for genetic control of the immune response. D5mit31 is a polymorphic marker associated with the zinc alpha-2-glycoprotein I (AZGP) gene on mouse chromosome 5. AZGP is a 307 amino acid plasma glycoprotein whose physiologic function involves regulation of lipolysis. It has homology to class I protein in its amino terminus and to the immunoglobulin constant domain at its carboxy terminus. AZGP binds lipids and zinc, which raises the possibility that it may relate to FVIII protein by mutual interaction with lipids and/or transition metals (e.g., copper/zinc). Our FVIII antibody linkage results differs from our previous work with FIX in that the linkage scores are not as high for the anti-FVIII response as for the anti-FIX response. Further, in this study there are different polymorphic markers associated with the antibody response in male and female mice. We continue to investigate genes in mice that control the antibody response to factors VIII and IX in order to identify and predict candidate genes that should be investigated in higher animal models of hemophilia (e.g., hemophilia A and B dogs) as well as in humans with hemophilia.

Author notes

Corresponding author