Chronic graft-versus-host disease is difficult to model in mice. Zhang and colleagues present a new mouse model with many features in common with the human syndrome, along with insights into the pathophysiology of this syndrome.
Chronic graft-versus-host disease (cGVHD) is a major cause of long-term nonrelapse morbidity and mortality in patients who undergo an allogeneic hematopoietic cell transplantation (HCT).1 cGVHD classically arises more than 100 days and rarely later than 500 days after transplantation. It presents as a progressive clinical syndrome that can involve multiple organs. Features of cGVHD resemble certain autoimmune diseases such as progressive systemic sclerosis, systemic lupus erythematosus, and others. Among the target organs commonly involved are the skin, liver, eyes, and/or oral pharynx. cGVHD can be treated with immunosuppression, and the mainstay of therapy is corticosteroids. However, despite immunosuppressive treatment, approximately 20% to 50% of patients who develop cGVHD die from complications associated with this disease.1
The pathophysiology of cGVHD is poorly understood. Donor T cells are thought to be the principal cause of both forms of GVHD—acute and chronic. However, different target antigens and effector pathways seem to mediate these clinical syndromes. Whereas acute GVHD is triggered by alloantigens recognized by T cells that mature in the thymus of the donor, the pathologic process of cGVHD is less certain. It has not yet been determined if donor postthymic or de novo graft-derived T cells predominate in this aberrant immune response. Whether the inciting antigens represent different allelic forms of donor and recipient molecules (alloantigens) or nonallelic endogenous molecules (autoantigens) is also not known. Recent clinical studies have shown that antibody responses against documented allogeneic targets correlate with the development and severity of cGVHD.2 Since alloantibodies are not thought to play a role in acute GVHD, these data further suggest that distinct pathologies underlie these syndromes.
Few animal models exist for the study of cGVHD. Some of the best-studied models use strain combinations that differ at the mouse major histocompatibility complex (MHC) equivalent of the human HLA, called H-2.3,4 In humans, HLA disparities certainly increase the likelihood of developing cGVHD. However, approximately 40% of recipients of HLA-matched sibling grafts develop the syndrome. Thus, in the search for ways to accurately model the cGVHD syndrome, investigators are now more frequently turning to the use of donor/recipient strain combinations that are matched at H-2 but differ at other genetic loci (minor alloantigens).5
In this issue of Blood, Zhang and colleagues report on the development of a new mouse model of cGVHD. The novelty of the model is based on the investigators' use of a donor/recipient pair that has not been previously reported, and disease is induced in recipients given the equivalent of a mouse nonmyeloablative transplant. DBA/2 and BALB/c mice are matched at H-2 but have multiple minor antigenic differences. When sublethally irradiated BALB/c mice receive transplants of large numbers of DBA/2 spleen cells (which contain progenitor and mature immune cells), the mice develop a syndrome with features resembling human cGVHD within 20 to 50 days after HCT. These features include sclerodermatous skin changes (more pronounced than in an existing model) and lupuslike abnormalities including serum antibodies to double-stranded DNA and glomerulonephritis. Removal of the thymus in some recipients did not change disease outcome, suggesting that donor postthymic T cells cause this syndrome. More importantly, Zhang et al show that mature donor B lymphocytes and not de novo B cells that arise after transplantation are associated with disease development. The investigators also address the role of donor T-cell subsets in mediating disease, including the activities of regulatory CD4+CD25+ cells. Development of cGVHD in this model is dependent upon CD4+ and not CD8+ T cells, and CD4+CD25– cells are required for cGVHD induction, where as CD4+CD25+ cells suppress the disease.
This newly established mouse model, which shares important features with the human syndrome, is certainly a welcomed addition to the armamentarium against cGVHD. The work of Zhang et al reinforces the notion that it will be possible to unravel the complex pathophysiology of cGVHD with the help of our furry friends. ▪
Comment on Zhang et al, page 2993