Education and Training:

B.A., Amherst College, Amherst, Massachusetts (1957-61)

M.D., University of Washington, Seattle, WA (1961-66)

Ph.D., University of Washington, Graduate School, Biochemistry (1963-64; 69-71)

Intern, Medicine, New York Hospital, Cornell Medical Center, NY (1966-67)

Assistant Resident, Medicine, University of Washington, Seattle, WA (1971-72)

Senior Fellow, Hematology, University of Washington, Seattle (1972-73)

Current Clinical Interests
Congenital bleeding disorders, inherited thrombotic risks, transfusion medicine & general hematology

Current Research Interests
Immune responses to factor VIII; structure-function relationships of clotting factors VIII and IX; gene therapy for the hemophilias

Research Description
Our laboratory has focused on structure-function relationships of normal and mutant factors VIII and IX to identify interactive sites important for hemostasis in vivo. Dysfunctional hemophilia A & B mutations are substituted into molecular models and mutants based on homologous proteins expressed to identify functionally important surface sites on factors VIII or IX. Alloimmune and autoimmune responses are being characterized in terms of changes in B-cell and T-cell epitopes across time. A second major focus has been support for basic and clinical studies on gene therapy for the hemophilias.

Factor IX and hemophilia B. Factor IX circulates as a single chain 415 aa zymogen at 90 nM. Its 34 kb gene in Xq26 has 8 exons, coding for signal and propeptides, the Gla domain, 2 EGF- domains, connecting and activation peptides and the catalytic (serine protease) domain (Bajaj & Thompson, 2006). The crystal structure of an inhibited porcine factor IX is known. Nearly 700 distinct mutations have been reported in about 2000 families. In the Seattle series, over 170 distinct genotypes have been characterized in 99% of nearly 300 families with hemophilia B studied.

Factor VIII and hemophilia A.  Factor VIII is a 2332 aa cofactor that circulates as a double chain precursor at 0.5-1 nM. Its 186 kb gene in Xq28 has 26 exons, coding for a signal peptide and a mature protein with A1-A2-B-A3-C1-C2 domains (Thompson, 2003). Cleavages by thrombin dissociate a von Willebrand factor (VWF) binding site and activate the cofactor. They crystal structure of human factor VIIIa became available in the Fall of 2007 from a group at FHCRC. Over 800 distinct mutations have been reported (half, missense). The two C domains provide a surface for lipid and a secondary VWF binding site; the crystal structure of a recombinant C2 domain has been solved and from that, a model of the homologous C1 domain constructed (Liu et al, 2000). C1C2 has been expressed and has enhanced binding to activated platelets compared to C2 alone (Hsu et al, 2008). Allo- and auto antibodies have common epitopes, most frequently in the A2 and C domains and are under active investigation.

Nearly half of severe hemophilia A is due to recurrent gene inversions and over 200 families have been identified in the Seattle series (Liu et al, 1999); ~20% are associated with allo-antibody inhibitors in at least one affected member (Thompson, Nakaya & Johnson, 2006). In nearly 500 other families with hemophilia A, a variety of other genotypes have been identified (Liu et al, 1998 & 2000). Baseline plasma testing for factor VIII antigen by a monoclonal ELISA allows distinction of the ~1/4th of patients with missense mutations that circulate excess, dysfunctional factor VIII antigen. These have provided a focus on the relationship between structural changes and their functional consequences. Selected mutations have been expressed in mammalian cells or in fragments. Partial gene deletions have undergone breakpoint analysis (Nakaya et al, 2004). We are scaling up expression of human C1C2 fragment (Hsu et al, 2008) including point mutations in each domain that, with data on C2 and the light chain of factor VIIIa will allow definition of phospholipid and von Willebrand factor binding sites, interactions with vitamin K-dependent factors (Liu & Thompson, 2000 & 2001, abstracts) and definition of inhibitory epitopes (Pratt et al, 2003, 2006).

Immune responses to factor VIII.  Little is known about which residues are involved as immunodominant epitopes for either B or T lymphocytes. We are preparing isolated domains of factor VIII with normal and surface mutant sequences to identify B cell epitope clusters (Pratt et al, 2005). Serial patient samples will be used to characterize the extent of epitope spreading and/or maturation as the response develops and may be predictive of responsiveness to therapy. We are studying T-cell epitopes in patient samples, including tetramer analysis to identify and isolate specific T-cell clones and determine the degree to which T-cell epitopes vary during an immune response to factor VIII. A DRb-dependent T-cell epitope was present in one individual with a C2 domain missense genotype (James et al, 2007) and different patients including some with A2 domain missense genotypes are being studied. Additional A2 domain epitopes have been defined and HLA-b-specific T cells have been cloned to study their cytokine responses (James et al, 2007, abstract).

Gene therapy for the hemophilias.  A clinical trial of intramuscular delivery of an AAV-factor IX vector into hemophilia B patients was conducted in Philadelphia and Palo Alto (Manno et al, 2000) and included 2 of 8 subjects that have been followed in Seattle; safety was demonstrated but efficacy, if any, was marginal. A trial with liver delivery of a more potent construct with tissue-specific promoter developed by our collaborators in Seattle showed the first evidence of a significant response. Unfortunately, there was a delayed inflammatory response to the capsid protein that was not predicted in animal models leading to termination of the protocol. As alternatives to viral gene transfer, we have found prolonged expression of therapeutic levels of factor IX in deficient mice due to episomal liver expression of therapeutic levels of factor IX in deficient mice due to episomal liver expression (Miao et al, 2000 & 2001) although the rapid, hydrodynamic delivery method used would not be suitable for humans. This technique also works to deliver factor VIII to hemophilia A mice and that provides an animal model for evaluating the immune response to factor VIII (Ye et al, 2004; Miao et al, 2006). Our current role in these studies is supportive and consultative.

Selected Publications

Liu M-L, Shen BW, Nakaya S, Pratt KP, Fujikawa K, Davie EW, Stoddard BL and Thompson AR: Hemophilic factor VIII C1 and C2 domain missense mutations and their modeling to the 1.5 Å human C2 domain crystal structure. Blood 96:979-987, 2000.

Thompson AR: Structure and function of the factor VIII gene and protein. Sem. Thrombos. Haemostas. 29:11-22, 2003.

Bajaj SP, Thompson AR: Molecular and structural biology of factor IX.  In: Hemostasis and Thrombosis: Basic Principles and Clinical Practice (5th ed.), Colman RW, Marder VJ, Clowes AW, George JN and Goldhaber SZ (eds).  Lippincott-Williams & Wilkins, Philadelphia, PA, Chapter 7, pgs 131-150, 2006.

Thompson AR: Congenital bleeding disorders from other coagulation protein deficiencies.  In: Clinical Hematology, Young NS, Gerson SL and High KA (eds).  Mosby Elsevier, Philadelphia, Chapter 66:855-866, 2006.

James EA, Kwok WW, Ettinger RA, Thompson AR, Pratt KP.  T-cell responses over time in a mild hemophilia A inhibitor subject: Epitope identification and transient immunogenicity of the corresponding self-peptide. J. Thrombos. Haemostas. 2007; 5:2399-2407.

Hsu TC, Pratt KP, Thompson AR. The factor VIII C1 domain contributes to platelet binding. Blood 111:200-208, 2008.