Toxicology and Pharmacokinetic Studies in Mice and Nonhuman Primates of the Nontoxic, Efficient, Targeted Hexameric FasL: CTLA4-FasL.

Cytotoxic T-lymphocyte antigen 4 (CTLA4)-FasL, a homo-hexameric signal converter protein, is capable of inducing robust apoptosis in malignant cells of the B-cell lineage expressing its cognate B7 and Fas targets, while sparing nonmalignant ones. This fusion protein's striking proapoptotic efficacy stems from its complementary abilities to coordinately activate apoptotic signals and abrogate antiapoptotic ones. A limiting factor in translating FasL or Fas receptor agonists into the clinic has been lethal hepatotoxicity. Here, we establish CTLA4-FasL's in vivo efficacy in multiple murine and xenograft models, for both systemic and subcutaneous tumors. Significantly, good laboratory practice (GLP) toxicology studies in mice indicate that CTLA4-FasL given repeatedly at doses up to five times the effective dose was well-tolerated and resulted in no significant adverse events. An equivalent single dose of CTLA4-FasL administered to nonhuman primates was also well-tolerated, albeit with a moderate dose-dependent leukopenia that was completely reversible. Interestingly, monkey peripheral blood mononuclear cells were more sensitive to CTLA4-FasL-induced apoptosis when tested in vitro. In both species, there was short-term elevation in serum levels of IL6, IL2, and IFNγ, although this was not associated with clinical signs of proinflammatory cytokine release, and further, this cytokine elevation could be completely prevented by dexamethasone premedication. Liver toxicity was not observed in either species, as confirmed by serum liver enzyme levels and histopathologic assessment. In conclusion, CTLA4-FasL emerges from animal model studies as an effective and safe agent for targeted FasL-mediated treatment of B7-expressing aggressive B-cell lymphomas.

Fn14·TRAIL fusion protein is oligomerized by TWEAK into a superefficient TRAIL analog.

Tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) demonstrates specific anti-cancer activity, but insufficient efficacy in patients. A fusion protein Fn14·TRAIL, that combines soluble TRAIL molecule with a specific TWEAK receptor Fn14, is a better apoptosis-inducer for hepatocellular carcinomas than soluble TRAIL. However, Fn14·TRAIL does not effectively induce apoptosis in tumors of the lymphoid origin. As malignant cell apoptosis is strongly enhanced by secondary oligomerization of TRAIL, we tested the hypothesis that soluble Fn14·TRAIL can be oligomerized and become more active by adding TWEAK, a cytokine secreted in the tumor environment. We revealed that TWEAK and Fn14·TRAIL spontaneously formed a stable complex that induced apoptosis of malignant lymphoblasts earlier and more efficiently than TRAIL. The TWEAK-modified Fn14·TRAIL oligomer bound to target cells and delivered apoptotic signaling via TRAIL receptors. The oligomer induced faster and stronger cleavage of procaspase-8, -9, and -3; BID; poly-ADP ribose polymerase; and RIP compared to TRAIL. The oligomer also reduced expression of the anti-apoptotic proteins c-FLIP short and cIAP-1. Our data indicate that Fn14·TRAIL can be converted into a highly effective TRAIL oligomer upon binding to TWEAK.

Highly efficient, in-vivo Fas-mediated apoptosis of B-cell lymphoma by hexameric CTLA4-FasL.

Non-Hodgkin lymphomas (NHLs) account for 4% of all malignancies. 5-year survival rate increased to 50% with new treatment modalities, however there is need for new effective treatment for the more aggressive, relapsing forms. Recently, CTLA4-FasL, that can bind to B7 and Fas receptor (Fas), was shown to induce robust apoptosis of cell lines originating from B cell lymphomas expressing both B7 and Fas, by activating pro-apoptotic signals in parallel to abrogating anti-apoptotic ones. The present study focuses on the unique properties of CTLA4-FasL as a potent apoptosis inducer of malignant cells in-vitro and in a xenograft model. CTLA4-FasL was found to naturally form a stable homo-hexamer. CTLA4-FasL induces robust apoptosis of a large variety of malignant cells while relatively sparing non-malignant ones, being more efficient when both receptors (B7 and Fas) are expressed on target cells. Even in non-B7 expressing cells, CTLA4-FasL exhibited better apoptotic activity than its parts, alone or in combination, however, only in B7 expressing cells apoptosis occurs at low concentrations and CTLA4-FasL induces activation of apoptotic signals and reduces anti-apoptotic ones. Importantly, CTLA4-FasL efficiently inhibited the growth of human B cell lineage tumors in a xenograft model, by provoking tumor cells' apoptosis. Thus, CTLA4-FasL, a natural homo-hexamer protein, induces robust apoptosis of malignant cells, in-vitro and in-vivo. In B-cell lymphoma, its potency stems from the combination of its synergistic effect of activating the caspases while abrogating the anti-apoptotic signaling, with its unique hexameric structure, making CTLA4-FasL a promising candidate for aggressive B cell lymphomas treatment.

Fn14•TRAIL effectively inhibits hepatocellular carcinoma growth.

BACKGROUND: New strategies for the treatment of hepatocellular carcinoma (HCC) are needed, given that currently available chemotherapeutics are inefficient. Since tumor growth reflects the net balance between pro-proliferative and death signaling, agents shifting the equilibrium toward the latter are of considerable interest. The TWEAK:Fn14 signaling axis promotes tumor cell proliferation and tumor angiogenesis, while TRAIL:TRAIL-receptor (TRAIL-R) interactions selectively induce apoptosis in malignant cells. Fn14•TRAIL, a fusion protein bridging these two pathways, has the potential to inhibit tumor growth, by interfering with TWEAK:Fn14 signaling, while at the same time enforcing TRAIL:TRAIL-R-mediated apoptosis. Consequently, Fn14•TRAIL's capacity to inhibit HCC growth was tested. RESULTS: Fn14•TRAIL induced robust apoptosis of multiple HCC cell lines, while sparing non-malignant hepatocyte cell lines. Differential susceptibility to this agent did not correlate with expression levels of TRAIL, TRAIL-R, TWEAK and Fn14 by these lines. Fn14•TRAIL was more potent than soluble TRAIL, soluble Fn14, or a combination of the two. The requirement of both of Fn14•TRAIL's molecular domains for function was established using blocking antibodies directed against each of them. Subcutaneous injection of Fn14•TRAIL abrogated HCC growth in a xenograft model, and was well tolerated by the mice. CONCLUSIONS: In this study, Fn14•TRAIL, a multifunctional fusion protein originally designed to treat autoimmunity, was shown to inhibit the growth of HCC, both in vitro and in vivo. The demonstration of this fusion protein's potent anti-tumor activity suggests that simultaneous targeting of two signaling axes by a single fusion can serve as a basis for highly effective anti-cancer therapies.

Amelioration of autoimmune neuroinflammation by the fusion molecule Fn14·TRAIL.

BACKGROUND: Multiple sclerosis (MS) is a, T cell-mediated autoimmune disease, the management of which remains challenging. The recently described fusion protein, Fn14·TRAIL, combining the extracellular domain of Fn14 (capable of blocking the pro-inflammatory TWEAK ligand) fused to the extracellular domain of the TRAIL ligand (capable of sending apoptotic signals through its receptors on activated inflammatory cells) was designed to modulate the immune system as an anti-inflammatory agent. The present study explores the efficacy of this purified protein as an anti-inflammatory agent, using the animal model of MS - experimental autoimmune encephalomyelitis (EAE). METHODS: EAE was induced by myelin oligodendrocyte glycoprotein (MOG). Fn14·TRAIL or vehicle were injected daily for 4 to 16 days, at different time points after disease induction. Animals were examined daily and evaluated for EAE clinical signs. Lymphocytes were analyzed for ex vivo re-stimulation, cytokine secretion, transcription factor expression and subtype cell analysis. Spinal cords were checked for inflammatory foci. The Mann- Whitney rank sum test, Student's t-test or ANOVA were used for statistical analysis. RESULTS: Significant improvement of EAE in the group treated with Fn14·TRAIL was noted from day 6 of disease onset and lasted until the end of follow-up (day 40 from disease induction), even in animals treated for 4 days only. Clinical improvement was linked to decreased lymphocyte infiltrates in the central nervous system (CNS) and to decreased Th1 and Th17 responses and to increased number of T- regulatory in the treated mice. No liver or kidney toxicity was evident. In vitro assays established the ability of Fn14·TRAIL to induce apoptosis of T cell lines expressing TRAIL receptors and TWEAK. CONCLUSIONS: In this study we established the potency of Fn14·TRAIL, a unique fusion protein combining two potentially functional domains, in inhibiting the clinical course of EAE, even when given for a short time, without apparent toxicity. These findings make Fn14·TRAIL a highly promising agent to be used for targeted amelioration of neuro-inflammatory processes, as well as other autoimmune pathologies.

IL-2-targeted therapy ameliorates the severity of graft-versus-host disease: ex vivo selective depletion of host-reactive T cells and in vivo therapy.

T cell depletion prevents graft-versus-host disease (GVHD) but also removes T cell-mediated support of hematopoietic cell engraftment. A chimeric molecule composed of IL-2 and caspase-3 (IL2-cas) has been evaluated as a therapeutic modality for GVHD and selective ex vivo depletion of host-reactive T cells. IL2-cas does not affect hematopoietic cell engraftment and significantly reduces the clinical and histological severity of GVHD. Early administration of IL2-cas reduced the lethal outcome of haploidentical transplants, and survivor mice displayed markedly elevated levels of X-linked forkhead/winged helix (FoxP3(+); 50%) and CD25(+)FoxP3(+) T cells (35%) in the lymph nodes. The chimeric molecule induces in vitro apoptosis in both CD4(+)CD25(-) and CD4(+)CD25(+) subsets of lymphocytes from alloimmunized mice, and stimulates proliferation of cells with highest levels of CD25 expression. Adoptive transfer of IL2-cas-pretreated viable splenocytes into sublethally irradiated haploidentical recipients resulted in 60% survival after a lethal challenge with lipopolysaccharide, which is associated with elevated fractions of CD25(high)FoxP3(+) T cells in the lymph nodes of survivors. These data demonstrate that ex vivo purging of host-presensitized lymphocytes is effectively achieved with IL2-cas, and that IL-2-targeted apoptotic therapy reduces GVHD severity in vivo. Both approaches promote survival in lethal models of haploidentical GVHD. The mechanism of protection includes direct killing of GVHD effectors, prevention of transition to effector/memory T cells, and induction of regulatory T cell proliferation, which becomes the dominant subset under conditions of homeostatic expansion.

Visualization of immune response kinetics in full allogeneic chimeras.

BACKGROUND: Donor Lymphocyte Infusion (DLI) is a well-recognized tool for augmentation of the anti-leukemia effect after mismatched bone marrow transplantation. Experimental results show, however, that DLI efficacy is strongly dependent on the number of donor hematopoietic cells persisting in recipient after transplantation. It is strong in mixed chimeras and relatively weak in full chimeras (FC) that replace host antigen-presenting cells by donor antigen-presenting cells. In this study we applied a new in vivo cytotoxicity monitoring method for evaluation of the changes in FC anti-host immunity after co-transplantation of donor and host hematopoietic cells together. METHOD: Full hematopoietic chimeras and naïve control mice were transplanted with a mixture of equivalent numbers of donor and recipient or donor and third party splenocytes labeled by a cell-permeable fluorescent dye CFDA-SE. The animals were sacrificed at various time points, and their splenocyte suspensions were prepared, depleted of red blood cells, stained with allophycocyanin-labeled anti-H2(b) antibodies, and analyzed using fluorescence-activated cell sorting. The immune response was assessed according to the percentage of single positive CFDA-SE(+)/ H2(b-) cells of all CFDA-SE(+) cells. RESULTS: FC grafted with splenocytes from similar FC mixed with splenocytes from naïve host-type or third-party-type mice rejected host cells within 14 days, and third-party cells within 7 days. NK cell depletion in vivo had no influence on host cell rejection kinetics. Co-infusion of host-type splenocytes with splenocytes obtained from naïve donor-type mice resulted in significant acceleration of host cell rejection (10 days). Naïve mice rejected the same amount of allogeneic lymphocytes within 3 days. CONCLUSIONS: Proposed method provides a simple and sensitive tool to evaluate in vivo post-transplant cytotoxicity in different experimental settings. The method demonstrates that FC is specifically deficient in their ability to reject host lymphocytes even when antigen-presenting host cells are provided. DLI improve anti-host immune response in FC but can not restore it to the level observed in naïve donor-type mice.

Prevention of acute graft-vs-host disease by a single low-dose cyclophosphamide injection following allogeneic bone marrow transplantation.

OBJECTIVE: Previously, we documented that conditioning based on donor-specific cell transfusion (DST) and subsequent selective depletion of activated donor-reactive cells by cyclophosphamide (CY) facilitates alloengraftment in a murine transplantation model. Transplantation event represents a strong immunogenic stimulus for host-reactive donor T cells that induce graft-vs-host disease (GVHD). Therefore, in this study, we addressed the question of whether a single posttransplantation CY administration (CY2) can prevent acute GVHD-related mortality without compromising engraftment of allogeneic transplant. MATERIALS AND METHODS: Splenocyte-enriched C57BL/6 bone marrow was transplanted to BALB/c recipients after mild irradiation, and conditioning with DST and 100 mg/kg CY. Following transplantation, recipients were left untreated or given on a specified day a single CY2 injection (50 mg/kg). All animals were monitored for survival, chimerism, and clinical signs of GVHD. Experimental mice that received BCL1 leukemia cells before transplantation were monitored for leukemia-related mortality as well. RESULTS: Animals that received no CY2 after transplantation died of acute GVHD. A single low-dose CY2 treatment within the first 5 days after transplantation prevented mortality in most recipients. However, only CY2 administration on days +1 or +5 preserved chimerism. Most chimeras survived GVHD-free for >200 days. Prolonged persistence of host-reactive T cells in mice (CY2 on day +5) permitted a reduction to be made in engraftment-essential irradiation dose and preserved a strong graft-vs-leukemia effect of transplantation. CONCLUSION: Acute GVHD can be prevented in mice by a single properly timed posttransplantation low-dose CY administration.

Mesenchymal stromal cells lose their immunosuppressive potential after allotransplantation.

OBJECTIVE: The stem cell fraction of mesenchymal stromal cells (MSCs) is capable of self-renewal and under inductive conditions differentiates into bone, cartilage, hematopoietic stroma, and other mesenchymal tissues. Therefore, MSCs represent a promising source for hard tissue repair therapies. MSCs are also immunosuppressive and prevent activation of allogeneic lymphocytes in vitro. Thus it has been suggested that they might be able to engraft in allogeneic recipients and downregulate recipients' immunity. In this study we examined whether MSCs retain their immunomodulating properties in vivo after allotransplantation. MATERIALS AND METHODS: MSCs were propagated from bone marrow (BM), placenta, or umbilical cord tissues. Using a murine parental-into-F1 model of graft-vs-host disease (GVHD) we tried to control GVHD by intravenous transplanting parental or recipient MSCs together with parental lymphocytes (day 0) and on days +7 and +14. MSCs' immunosuppressive potential in vivo was also examined by comparing their ability to construct ectopic bone after local transplantation with osteogenic inductor (demineralized bone matrix) under the kidney capsule of syngeneic and allogeneic recipients. RESULTS: Repeated IV MSC injections failed to reduce GVHD-related recipient mortality. Local implantation of MSCs propagated from BM, placenta or umbilical cord resulted in ectopic bone formation in syngeneic recipients and in transplant rejection by allogeneic mice. CONCLUSION: MSCs lose their immunosuppressive potential in mismatched setting.

Depletion of alloantigen-primed lymphocytes overcomes resistance to allogeneic bone marrow in mildly conditioned recipients.

OBJECTIVE: Successful implantation of allogeneic bone marrow (BM) cells after nonmyeloablative conditioning would allow to compensate for the inadequate supply of compatible grafts and to reduce mortality of graft-vs.-host disease (GVHD). Recently, we proposed to facilitate engraftment of mismatched BM by conditioning for alloantigen-primed lymphocyte depletion (APLD) with cyclophosphamide (CY). Here we summarize the experimental results obtained by this approach. MATERIALS AND METHODS: Naive or mildly irradiated BALB/c mice were primed with C57BL/6 BM cells (day 0), treated with CY (day 1) to deplete alloantigen-primed lymphocytes, and given a second C57BL/6 BM transplant (day 2) for engraftment. Recipients were repeatedly tested for chimerism in the blood and followed for GVHD and survival. The protocol was also tested for inducing tolerance to donor tissue and organ allografts, and for treatment of leukemia, breast cancer, and autoimmune diabetes in NOD mice. RESULTS: APLD by 200 mg/kg CY provided engraftment of allogeneic BM from the same donor in 100% mildly irradiated recipients. Eighty percent chimeras remained GVHD-free more 200 days. All chimeras accepted permanently donor skin grafts and donor hematopoietic stromal progenitors. Allogeneic BM transplantation (BMT) after APLD had a strong therapeutic potential in BALB/c mice harboring malignant cells and in autoimmune NOD recipients. Tolerance-inducing CY dose could be reduced to 100 mg/kg. Conditioning for APLD resulted in engraftment of allogeneic BM after a significantly lower radiation dose than treatment with radiation and CY alone. CONCLUSION: Our results demonstrate that conditioning for APLD has a definite advantage over general immunosuppression with CY and radiation therapy.

Depletion of donor-reactive cells as a new concept for improvement of mismatched bone marrow engraftment using reduced-intensity conditioning.

OBJECTIVE: New nonmyeloablative strategies to improve acceptance of mismatched bone marrow (BM) may compensate for the inadequate supply of compatible grafts. Recently we proposed to facilitate engraftment of mismatched BM by selective depletion of activated donor-reactive host cells with cyclophosphamide (CY). Here we have compared engraftment of allogeneic BM after depletion of antigen-activated host lymphocytes by CY, with BM engraftment following general immunosuppression by the same CY dose. MATERIALS AND METHODS: Naive or mildly irradiated BALB/c mice were primed with C57BL/6 BM cells (day 0), treated with CY in order to deplete activated T cells (day 1), and transplanted with a second C57BL/6 BM inoculum (day 2) in order to achieve BM engraftment. Alternatively, mice received an equal dose of donor BM cells in a single injection one day after the same CY dose. Treated animals were repeatedly tested for persistence of donor cells in the blood. RESULTS: Depletion of alloantigen-primed lymphocytes by 200 mg/kg CY provided stable GVHD-free engraftment of allogeneic BM in nonirradiated mice, while immunosuppressive treatment with the same CY dose alone resulted in BM rejection. Low-dose irradiation before priming with donor BM allowed the tolerance-inducing CY dose to be reduced to 100 mg/kg. Alloantigen-primed lymphocyte depletion (APLD) by a reduced CY dose resulted in engraftment of donor BM after a significantly lower irradiation dose than treatment with irradiation and CY alone. CONCLUSION: Our results demonstrate that conditioning that focuses on APLD has a definite advantage over general immunosuppression with CY and radiation therapy.

Prevention of diabetes in nonobese diabetic mice by nonmyeloablative allogeneic bone marrow transplantation.

OBJECTIVE: Autoimmune diabetes in nonobese diabetic (NOD) mice can be prevented by allogeneic bone marrow transplantation (BMT) from diabetes-resistant murine strains. Donor-specific tolerance can also be induced by BMT; however, clinical application of nonmyeloablative conditioning prior to BMT may be essential for reducing transplant-related toxicity and mortality. In this study, we have attempted to treat autoimmunity using a new nonmyeloablative regimen for BMT. MATERIALS AND METHODS: Naïve NOD were irradiated with 650 cGy and injected intravenously (i.v.) with splenocytes from overtly diabetic NOD mice for induction of diabetes mellitus. Three days later, experimental mice received allogeneic C57BL/6 or (C57BL/6 x BALB/c) F1 bone marrow (BM) cells i.v. for intentional activation of donor-reactive cells, and 24 hours later intraperitoneal injection of cyclophosphamide (CY) for selective depletion of alloreactive cells. In order to induce chimerism, recipients were given a second IV inoculum of donor BM 1 day after CY. RESULTS: Our method of nonmyeloablative BMT converted recipients to full or to mixed chimeras and prevented development of diabetes. Although NOD mice treated with 200 mg/kg CY died of graft-vs-host disease (GVHD), we observed diabetes-free survival for >300 days in 90% of C57BL/6 --> NOD BM chimeras treated with 60 mg/kg CY. CONCLUSION: Our data show that allogeneic BMT after reduced-intensity conditioning based on deletion of activated donor-reactive host cells by means low-dose CY results in prevention of autoimmune diabetes by converting recipients to stable, GVHD-free BM chimeras.

Low-dose gamma-irradiation promotes survival of injured neurons in the central nervous system via homeostasis-driven proliferation of T cells.

Protective autoimmunity was only recently recognized as a mechanism for attenuating the progression of neurodegeneration. Using a rat model of optic nerve crush or contusive spinal cord injury, and a mouse model of neurodegenerative conditions caused by injection of a toxic dose of intraocular glutamate, we show that a single low dose of whole-body or lymphoid-organ gamma-irradiation significantly improved the spontaneous recovery. Animals with severe immune deficiency or deprived of mature T cells were unable to benefit from this treatment, suggesting that the irradiation-induced neuroprotection is immune mediated. This suggestion received further support from the findings that irradiation was accompanied by an increased incidence of activated T cells in the lymphoid organs and peripheral blood and an increase in mRNA encoding for the pro-inflammatory cytokines interleukin-12 and interferon-gamma, and that after irradiation, passive transfer of a subpopulation of suppressive T cells (naturally occurring regulatory CD4(+)CD25(+) T cells) wiped out the irradiation-induced protection. These results suggest that homeostasis-driven proliferation of T cells, induced by a single low-dose irradiation, leads to boosting of T cell-mediated neuroprotection and can be utilized clinically to fight off neurodegeneration and the threat of other diseases in which defense against toxic self-compounds is needed.

CD40 ligand-specific antibodies synergize with cyclophosphamide to promote long-term transplantation tolerance across MHC barriers but inhibit graft-vs-leukemia effects of transplanted cells.

OBJECTIVES: We previously demonstrated that allogeneic bone marrow transplantation (BMT) after low-dose total lymphoid irradiation (200 cGy) and depletion of donor-reactive cells with cyclophosphamide (Cy) converted recipients to graft-vs-host disease (GVHD)-free chimeras tolerant to donor skin grafts. BMT also generated strong graft-vs-leukemia (GVL) response. However, clinical application of the protocol was hampered by the requirement for a relatively high dose of Cy (200 mg/kg). In this study we have tried to minimize the Cy dose by a concomitant blockade of CD40-CD40L interaction. MATERIALS AND METHODS: Mildly irradiated BALB/c mice were primed with C57BL/6 BM cells (BM(1)) and skin graft on day 0, injected with Cy (200 mg/kg or less) on day 1, and transplanted with a second C57BL/6 BM cell inoculum (BM(2)) on day 2. CD40L-specific antibody (MR1) was given with BM(1), BM(2), and 2 days later. Treated animals were monitored for survival, chimerism, and skin allograft rejection. The GVL potential of transplanted cells was examined in mice inoculated with BCL1 leukemia cells before irradiation. RESULTS: Blocking CD40-CD40L interaction with MR1 mAb allowed the reduction of a tolerance-generating Cy dose by 50%. Unfortunately, adding MR1 to the protocol reduced the GVL potential of the transplanted cells. Neither low-dose Cy nor antibodies alone could downregulate donor or recipient immune response. CONCLUSIONS: CD40L-specific antibodies synergize with Cy to induce bilateral transplantation tolerance. Therefore, their use may be beneficial for safer allogeneic BMT for nonmalignant indications. However, due to MR1-associated reduction of GVL effects, MR1 should be considered with caution as conditioning for BMT for leukemia-bearing recipients.

Nonmyeloablative allogeneic bone marrow transplantation as immunotherapy for hematologic malignancies and metastatic solid tumors in preclinical models.

OBJECTIVE: We previously demonstrated that a combination of mild total lymphoid irradiation (TLI) with selective depletion of the host's donor-reactive cells allows for stable and graft-vs-host disease (GVHD)-free engraftment of allogeneic bone marrow (BM). In this study, we investigated the efficacy of this nonmyeloablative strategy for BM transplantation (BMT) as immunotherapy for minimal residual disease. MATERIALS AND METHODS: BALB/c mice inoculated with leukemia (BCL1) or breast carcinoma (4T1) cells were conditioned for BMT with TLI (200 cGy) followed by priming with donor (C57BL/6) BM cells on day 1, and by injection with 200 mg/kg cyclophosphamide on day 2. After conditioning (day 3), recipients were transplanted with BM cells from the same donor. Treated animals were monitored for 230 days for survival, development of leukemia/solid tumor, and GVHD. RESULTS: BMT converted the mice to complete chimeras and prevented development of leukemia in 90% of recipients and locally growing breast carcinoma in 40% of the mice. Immunization of donors of the second BM with 4T1 cells prevented development of breast carcinoma in 80% of 4T1 inoculated mice. Fewer animals treated for malignancy by nonmyeloablative BMT died of GVHD than those treated by myeloablative BMT. However, late GVHD-related mortality in mice treated for leukemia was higher than after nonmyeloablative BMT to naive recipients (p < 0.00001). Infusion of host-type anti-donor immune lymphocytes 8 days after BMT improved the survival of recipients treated for leukemia without affecting engraftment and the graft-vs-leukemia potential of donor BM. CONCLUSIONS: Effective eradication of malignant cells can be achieved following allogeneic BMT after nonmyeloablative conditioning.