Int“Dll”igent control of T-cell pathology in GVHD

In this issue of Blood, Radojcic et al achieved almost total prevention of chronic graft-versus-host-disease (cGVHD) in 2 complementary mouse models of bone marrow transplantation by inhibition of Notch Delta-like ligands 1 and 4 (Dll1 and Dll4) signaling, raising hope for a new therapeutic intervention in humans.¹ 

GVHD remains a major complication after allogeneic hematopoietic bone marrow transplantation. Acute GVHD (aGVHD), usually treated with glucocorticoids, is frequently followed by the devastating chronic form of this disease for which treatment options are limited. What if a limited number of antibody injections following bone marrow transplantation could mitigate or even prevent both aGVHD and cGVHD? This vision has gained support by data presented in 3 studies that identified Notch signaling as a highly promising target for dampening excessive T- and B-cell activity.^1-3 

Notch signaling provides a cell–cell communication pathway. The 4 members of the Notch receptor family in mammals are activated by binding to 1 of 5 different ligands, named Dll1, Dll3, Dl4, Jagged1, and Jagged2. Originally, Notch proteins were recognized as arbiters of cell fate decisions. In the hematopoietic system, activation of Notch receptors guides the early differentiation of T cells.⁴  Constitutive activation of Notch signaling can be oncogenic; activating Notch1 mutations, for instance, are well-known drivers of T-cell acute lymphoblastic leukemias.⁵  Small molecule inhibitors of Notch signaling have been tested in clinical trials; however, they lack specificity for single receptors and treatment-related toxicity limits their widespread use. The use of blocking antibodies overcomes these limitations.

During aGVHD, an inflammatory immune reaction takes place that is mediated by mature donor T cells directed against the host alloantigens. Various approaches to deplete donor T cells from hematopoietic transplantation products have been tried to minimize the subsequent incidence and severity of GVHD.⁶  Maillard’s group identified in previous studies a critical role for Notch1/2 in these host-reactive T cells in the pathology of aGVHD.^2,7  Suppressing Notch signaling by the expression of a dominant negative variant of Mastermind-like, an essential signaling component of the Notch pathway, decreased the production of inflammatory cytokines and increased the number of regulatory T cells, leading to reduced organ damage. Restoring antigen-specific tolerance by the induction of regulatory T cells is an elegant way to suppress dangerous immune reactions while leaving the general immune competence intact.⁸  Furthermore, after receiving allogeneic bone marrow transplantation (allo-BMT), mice survived for a long period free of severe GVHD when the Notch1–Dll4 interaction was blocked by neutralizing antibodies. In contrast to Notch-directed antibodies, which were poorly tolerated, targeting Dll proteins was devoid of systemic side effects.

Often following aGVHD, cGVHD displays a wider spectrum of clinical manifestations.⁹  The underlying complex immune reactions engage T cells, B cells, and other antigen-presenting cells. The pathological events include destruction of thymic tissue, the emergence of allo- and autoreactive T cells with a bias toward Th17 polarization, and reduced numbers of regulatory T cells. Macrophages become polarized toward the M2 phenotype and stimulate the production of transforming growth factor-β, which finally promotes fibrosis in multiple tissues and compromises organ function.¹⁰  Potentially pathologic B cells are increased and characterized by constitutive B-cell receptor signaling and BAFF-regulated pathways. Notch2 signaling contributes to the cGVHD-associated B-cell pathology, augmenting B-cell receptor responses by lowering the threshold for responses to alloantigens, and Notch2 blockade reversed the hyperactivated state of the pathogenic B cells.³ 

T-cell pathology of cGVHD is in the focus of the current study, Radojcic et al confirmed that targeting Notch signaling in donor T cells achieved beneficial immunomodulation and inhibited GVHD after allo-BMT. In humans, cGVHD typically manifests in multiple organs and commonly includes skin pathology and bronchiolitis obliterans, the inflammatory obstruction of the lung bronchioles. However, the manifestations are diverse and cannot be replicated in a single mouse model. Radojcic et al used 2 models to mimic 2 findings in human cGVHD: the first is characterized by systemic fibrosis including scleroderma (Scl-cGVHD), and the second is characterized by prominent lung involvement with bronchiolitis obliterans. Using antibody-mediated blockage of specific members of the Notch family receptors and their ligands, the major candidates responsible for pathological alterations were identified. In the Scl-cGVHD model, inhibition of Dll4-stimulated Notch activation was sufficient to prevent cGVHD manifestation, whereas blocking Dll1 had almost no beneficial effect. Most interestingly, a single application of the blocking antibody was sufficient for long-term protection, but it had to be given within the first 48 hours after transplantation. In this model, Notch blockade also resulted in the expansion of alloantigen-specific regulatory T cells. In the bronchiolitis obliterans-cGVHD model, antibody-mediated neutralization of Dll1 or Dll4 provided similar disease protection, and the treatment was also effective after disease manifestations were present, thus mitigating even established cGVHD.

The use of mouse models raises the question about the relevance of the findings for the human system. Murine models mimic certain aspects of the immunopathogenesis of cGVHD, but interspecies differences regarding immune cell signaling, details of disease pathogenesis, and efficacy of treatment remain concerning. Despite this concern, Notch signaling targets are an attractive potential therapeutic approach. A short-term treatment blocking Notch signaling during the first days after transplantation might be sufficient to prevent aGVHD and at least some manifestations of cGVHD.

Although the heterogeneous nature of cGVHD likely prevents a simple treatment fitting all manifestations, the complexity of the molecular interactions also offers a variety of specific targets to interrupt pathways critical for the emergence and maintenance of the disease. Therefore, identification of additional details of the Notch–Dll signaling interaction, such as the cell types expressing the critical Dll ligands or a possible cross talk of Notch with other signaling pathways, is warranted to discover additional therapeutic targets.

In summary, Notch signaling in the early posttransplant environment was identified as crucial for the occurrence of subsequent cGVHD. Modulating Notch signals using genetic or pharmacological approaches with antibodies to Dll1/Dll4 produced therapeutic effects. The near future will hopefully show whether targeting Notch signaling is a feasible and curative approach to treat human GVHD.