In vitro separation of host specific graft-versus-host and graft-versus-leukemia cytotoxic T cell activities

Summary: The association of graft-versus- host disease (GVHD) with lower relapse rates following allogeneic bone mar- row transplantation (BMT) in humans led us to analyse post HLA-identical BMT derived anti-host cytotoxic T cells (CTL) for their putative anti-Ieukemic activity. To establish whether graft-versus-host (GVH) and graft- versus-leukemia (GVL) activities are separate, CTL lines were generated at different time points post-BMT from three patients suffering from acute GVHD. These CTL lines, which exhibited lysis of host normal lym- phocytes and neoplastic cells, were analysed at the clonal level. Three functionally different types of clones were characterized: clones directed at host specific minor Histocompatibility (mH) antigens which are shared by patient's periphera! blood lymphocytes (PBL) and leukemic cells; clones recognizing only host PBL but not host leukemic cells; and putative GVL clones directed at patient's neoplastic cells only. These data could explain the long controversies on dissection of GVH and GVL activities. Our results demonstrate that GVH and GVL activities can be dissected, while non- separable effector cells which exhibit both activities do exist as well. One of the major obstacies in allogeneic bone marrow transplantation (BMT) is the oecurrence of graft- versus-host disease (GVHD).' T cell depletion of the donor bone marrow inoculum shows a reduetion in the ineidence and severity of GVHD but an increase in relapse rate. Mature T cells in this donor bone marrow inoculum vital for graft aeeeptance and responsible for GVHD are probably also needed to eliminate residual leukemic cells: the graft-versus-leukemia (GVL) effect.2·3 A substantial number of experimental animal modeis indicate a GVL effect of allogeneic BMT.4·5 Although controversial, mouse data support the notion that the GVL reaction can be distinguished from the