Presenting ADAMTS13 on a TTP-associated MHC

In this issue of Blood, Sorvillo et al investigate possible molecular triggers leading to idiopathic, autoimmune thrombotic thrombocytopenic purpura (TTP) by identifying naturally processed ADisintegrin And Metalloprotease with ThromboSpondin type 1 motif 13 (ADAMTS13)-derived peptides presented on human dendritic cells.¹ 

TP is a rare but serious pathological disorder in which the ultra large form of von Willebrand factor (ULVWF) released by activated endothelium into the circulation cross-links platelets, causing intravascular agglutination under high shear stress, ie, primarily in the microvasculature.²  Most patients with idiopathic TTP have acquired autoantibodies to ADAMTS13, a large circulating metalloprotease that cleaves ULVWF at a specific site, generating VWF multimers that are significantly less adhesive than ULVWF.^3,4  Anti-ADAMTS13 antibodies in plasma from TTP patients consist primarily of IgG₁ and IgG₄ subclasses, indicating the involvement of effector T cells in the etiology of this disease,⁵  and the major histocompatibility complex (MHC) allele HLA-DRB1*11 has recently been identified as a risk factor for the development of TTP. The study by Sorvillo et al presented in this issue¹  systematically investigated the HLA restriction of ADAMTS13 peptide presentation on human MHC Class II by isolating and expanding dendritic cells from 17 healthy blood donors in culture, “feeding” them different concentrations of recombinant ADAMTS13 protein, and then recovering the peptides presented on their surface and identifying them using mass spectrometry. Interestingly, dendritic cells exposed to ADAMTS13, but not control cells treated with phosphate-buffered saline alone, presented peptides derived from several ADAMTS13 domains, and peptides derived from its C-terminal CUB2 domain were presented with the highest efficiency. Dendritic cells from donors with an HLA-DRB1*11 allele exposed to a higher ADAMTS13 concentration presented only differentially processed versions of the same CUB2 peptide, which contains the predicted DRB1*11-binding sequence FINVAPHAR. The binding of naturally processed ADAMTS13 peptides to MHC Class II on human dendritic cells indicates that these peptides may contain clinically relevant T-cell epitopes, although future experiments showing stimulation of human effector T cells by similar peptides will be required to confirm this hypothesis.

Autoimmune TTP is caused by antibodies that bind to ADAMTS13 and neutralize its proteolytic activity. Most acquired TTP patients circulate antibodies that bind to the spacer domain of ADAMTS13, although antibodies with specificity for other regions, including the CUB domains, have also been identified in subsets of TTP patients.^6-8  Although T-cell and B-cell epitopes often occur in close proximity or even overlap, they can also be derived from spatially distant regions of a protein antigen. This is because the presentation of T-cell and B-cell epitopes to the immune system is fundamentally different. CD4 T-cell epitopes consist of peptides at least 9 to 16 residues in length derived from antigens such as ADAMTS13, which bind to the MHC Class II binding groove on antigen-presenting cells. The MHC Class II-peptide complex is presented to T-cell receptors on, presumably in the case of TTP, autoreactive T cells that escaped thymic deletion and subsequently became stimulated by ADAMTS13. B-cell epitopes are 3-dimensional surfaces, eg, of properly folded protein antigens, recognized by antibodies and by receptors on the surface of memory B cells. Unlike T-cell epitopes, they are often comprised of noncontiguous amino acid sequences in the protein antigen.

A major contribution of this study is that it identifies ADAMTS13 peptides that have been naturally processed in the MHC compartment of human dendritic cells and then presented on their surface, which is required for subsequent effector T-cell stimulation leading to antibody formation. Both epitope prediction algorithms and studies of synthetic peptides binding to HLA proteins, although extremely useful in many immunologic investigations, tend to overpredict T-cell epitopes. The elution of a predominant ADAMTS13 core peptide from dendritic cells isolated from donors with the risk-associated allele HLA-DRB1*11 points the way to additional future studies that will determine whether this is indeed an immunodominant epitope involved in the etiology of autoimmune/idiopathic TTP. Clearly, the vast majority of individuals with an HLA-DRB1*11 allele will never develop the rare autoimmune disorder TTP. The identification of T-cell epitopes involved in the development and progression of TTP, however, will lead to the isolation and characterization of T cells responsible for the production of neutralizing anti-ADAMTS13 antibodies. In addition to the immediate clinical relevance of such studies, this will present an opportunity to study the processes by which T cells that evaded thymic deletion become autoreactive and thus pathogenic. For example, approximately two-thirds of TTP cases occur in women, and pregnancy can be the initiating event.⁹  Idiopathic TTP has also been associated with HIV, various drugs, bone marrow transplantation, malignancies, infections, and inflammatory disorders.¹⁰  The common factors among these precipitating events and conditions are poorly understood. The identification of specific epitopes in ADAMTS13 will facilitate studies in which the relevant pathogenic T cells may be isolated from peripheral blood donated by human subjects with TTP and will also better define HLA-associated risk factors for TTP. Finally, many autoimmune responses, eg, rheumatoid arthritis, are provoked by epitopes in multiple, often poorly-defined, antigens. The developing mechanistic story of human autoimmune responses to a single well-defined protein antigen, ADAMTS13, will likely be of significant interest to researchers carrying out both basic and clinical studies of other autoimmune disorders.