Quantitative studies on graft-versus-leukemia after allogeneic bone marrow transplantation in rat models for acute myelocytic and lymphocytic leukemia.

The major shortcoming of present day treatment of leukemia by bone marrow transplantation (BMT) remains leukemia relapse. It has become clear that a graft-versus-host reaction (GVHR) is accompanied by a graft-versus-leukemia reaction (GVLR) which may prevent leukemia relapse. In two non-immunogenic rat leukemia models, the Brown Norway acute myelocytic leukemia (BNML) and the WAG/Rij acute lymphocytic leukemia L4415, total body irradiation (TBI) was given to induce 'minimal residual disease' (MRD). Subsequently, it was attempted to evoke a GVLR by using syngeneic or allogeneic BMT, with or without addition of graded numbers of lymphocytes. In both leukemia models the addition of high numbers of syngeneic lymphocytes to the syngeneic graft had no antileukemic effect. Allogeneic marrow grafts, which contain at the most 8% lymphocytes, only resulted in a GVLR when splenocytes were added. The therapeutic window was found to be narrow, i.e. in fully mismatched BMT the number of allogeneic splenocytes resulting in a significant GVLR (2-3 log leukemic cell kill) without inducing (lethal) acute GVHD was critical. Increasing the number of allogeneic spleen cells added to the allogeneic BM graft induced lethal acute GVHD. To date, our data indicate that the GVLR is an allogeneic effect, inseparable from GVHD.