A Novel Function of CD26-Mediated Costimulation in the Cytotoxic Activity of Human CD8+ T Cells in Xenogeneic Chronic Graft-Versus-Host Disease (cGVHD) and GVL Mice Model.

Abstract 3548

Poster Board III-485

CD26 is a 110-kDa cell surface glycoprotein with known dipeptidyl peptidase IV (DPPIV) activity in its extracellular domain, capable of cleaving amino-terminal dipeptides with either L-proline or L-alanine at the penultimate position. CD26 activity is dependent on cell type and the microenvironment, which influence its multiple biological roles. CD26 plays an important role in immunology, autoimmunity, diabetes and cancer. The role of CD26 in immune regulation has been extensively characterized, with our recent findings elucidating its linkage with signaling pathways and structures involved in T cells activation as well as antigen presenting cell (APC)-T-cell interaction. CD26 is preferentially expressed on human CD4+ memory T cells and CD4+CD45RO+CD26+ T cells exhibit response maximally to the recall antigen tetanus toxoid. Human T-helper 1 (Th1) cells display a higher expression of CD26 and are much more sensitive to CD26-mediated costimulation than human Th2 cells. In contrast, the role of CD26 in human CD8+ T cells still remains to be elucidated, while CD8+ T cells express CD26 as well as CD4+ T cells. In this study, we studied the effector function of CD26-mediated costimulation of CD8+ T cells. In comparison with CD28-medaited costimulation, which is one of representative T cell costimulatory pathway, CD26-mediated costimulation in CD8+ T cells showed delayed proliferation than that of CD28 stimuatlion, but finally expanded to a similar extent. The secretion of inflammatory cytokines, TNF-alpha and IFN-gamma, was strongly induced after CD26-mediated costimulation. In contrast, the secretion of IL-2 and IL-5 was significantly less as compared with CD28-mediated costimulation. Finally, we showed that the expression of Granzyme B, one of the major effector molecules in the cytotoxic activity of CD8+ T cells, was markedly enhanced by CD26-mediated costimulation in dose dependent manner of anti-CD26 monoclonal antibody stimulation. Moreover, with CD26-mediated costimulation, CD8+ T cells were observed to kill target cells in a Granzyme B-dependent manner. Next, we examined CD26-dependent organ injury by human T cells using xenogeneic chronic GVHD (x-cGVHD) mice model. About 6 weeks after transplantation of human peripheral blood mononuclear cells (PBMC) into NOG-Scid mice, mice develop loss pf hair, loss of weight, and human CD8+ lymphocytes infiltration into liver, lung, and salivary glands, which exhibit pathological changes of cGVHD. Administration of anti-human CD26 mAb decreased onset of x-cGVHD, while human lymphocytes were grafted successfully. Finally, we examined whether GVL effect is exerted via CD26-mediated costimulation, using x-cGVHD mice model in which P388 leukemic cells were grafted subcutaneously. In this tumor-graft model, P388 cells were not grafted in mice with onset of x-cGVHD, which indicated that GVL effect was exerted by human lymphocytes. On the other hand, P388 cells were also diminished in mice with decreased x-cGVHD by administration of anti-CD26 mAb. Taken together, our data strongly suggest that CD26-mediated costimulation in CD8+ T cells deeply involves in the pathology of cGVHD. The present study offers a novel notion that CD26 may become one of therapeutic targets of GVHD without impairment of GVL effect.