An expanded pool of unrelated donors (URD) is now being utilized for clinical allogeneic bone marrow transplantation. Because URD transplants can involve a mismatch at least at one genotypic MHC locus, we developed a C57BL/6 congenic mouse model to better understand graft rejection based exclusively on MHC class I or class II disparities. T cell-depleted (TCD) BM from class II-disparate mutant bm12 mice was transplanted into irradiated C57BL/6-Ly5.2 congenic hosts. These mice express a different allelic form of the Ly5 (CD45) marker than bm12 and thus permit definitive typing of reconstituted mice by flow cytometry. Peripheral blood typing indicated that host-mediated graft rejection was restricted to class II-reactive CD4+ cells since an anti-CD4 monoclonal antibody significantly inhibited rejection by enhancing the level of mean donor cell engraftment from 13% to 53%. The administration of anti-CD8 had no inhibitory effect on the ability of the recipient to reject the donor graft. Hematologic reconstitution studies revealed that there was a direct relationship between the level of donor cell engraftment and the extent of lymphoid, myeloid, and erythroid recovery. Mice that did not engraft had sustained reductions in in hematologic recovery. In other studies, TCD BM from bm1 mice mutated only in the MHC class I region was rejected by C57BL/6-Ly5.2 recipients. Anti-CD8 mAb infusion prevented the class I response resulting in an increase in the level of mean donor cell engraftment from 1% to 78%. Anti-CD4 alone had no effect. A CD4/CD8 T cell interaction could be important since the combination of anti-CD4 and anti-CD8 mAb resulted in significantly better donor engraftment than with the individual antibodies. Since the role of NK cells in these models has not been previously established and there is a reported association between NK cells and the rejection of BM cells that lack the expression of self-MHC antigens, we tested the role of NK cells. The elimination of NK cells using anti-NK mAb had no effect on graft rejection in either the bm1 or bm12 model. In addition to their value in exploring mechanisms of graft rejection, these models may prove useful for evaluating the efficacy of anti-BM rejection agents exclusively against class I- or class II-restricted disparities.