Targeting host B-cell immune responses by persistent donor NK-cell alloreactivity following nonmyeloablative allogeneic stem cell transplantation

To the editor:

Allogeneic bone marrow transplantation (BMT) is an established treatment for a variety of hematological and immunological disorders. It is postulated that the beneficial effects after myeloablation and haploidentical BMT are mediated by alloreactive donor natural killer (NK) cells, and host-reactive NK-cell clones have been isolated from human transplant recipients up to 3 months after transplantation.^1,2  However, little is known about the ability of graft-derived NK cells to target host immune cells after nonmyeloablation and haploidentical BMT. Likewise, the persistence of donor NK-cell alloreactivity in chimeric hosts is poorly defined in this setting.³  Here, we report the ability of alloreactive donor NK cells to eradicate host immune cells after haploidentical BMT in mice, without inducing graft-versus-host disease (GVHD).

The Experimental Animal Commission of Leiden University Medical Center (Leiden, The Netherlands) approved the experiments described in this letter.

Recently, we showed the effective treatment of B-cell–mediated autoimmune disease (AID) in mice by allogeneic BMT.⁴  We postulated the rapid eradication of host B cells produced pathogenic autoantibodies by donor NK cells. To further analyze donor NK-cell alloreactivity toward antibody-secreting B cells, we used a model of haploidentical (F1→P) BMT involving host B cells recognizing a model antigen, ovalbumin (OVA). In this model, donor T cells are tolerant for host cells,^5,6  but donor NK cells are activated as they lack inhibitory signals provided by self–major histocompatibility complex (MHC) class I.^7,8  After immunization, high levels of anti-OVA antibodies were detected. Subsequently, the mice were treated with nonmyeloablative conditioning (ie, low-dose total-body irradiation (6 Gy) and anti-CD40L antibodies (0.5 mg), and BMT from syngeneic, haploidentical or fully MHC-mismatched donor mice. Both fully MHC-mismatched and haploidentical BMT resulted in the early elimination of OVA-specific B cells (Figure 1A) and anti-OVA antibodies (Figure 1B), in contrast to syngeneic BMT. Recipient mice had no signs of acute GVHD (ie, no weight loss and/or detectable histologic abnormalities of liver, gut, and skin), supporting the notion that donor NK cells can mediate beneficial anti–host B-cell alloresponses, without causing GVHD.

Next, we examined whether host-reactive NK-cell alloresponses after haploidentical BMT are still present 3 months after BMT using an in vivo cytotoxicity assay.^9,10  In this experiment, differentially carboxyfluoroscein succinimidylester (CFSE)-labeled splenocytes from syngeneic (host) and haploidentical (control) mice were injected into chimeric hosts (4% to 5% host Gr-1⁺ cells). Prior to the challenge with CFSE-labeled splenocytes, NK cells in recipient mice were (or were not) depleted by anti-NK1.1 antibodies. At day +7, peripheral-blood cells were analyzed for the presence of CFSE⁺ cells, and the extent of host-cell elimination was determined. Syngeneic cells were eliminated in control mice, while this elimination was inhibited after depletion of NK1.1⁺ cells (Figure 1C). Although the effect of depletion of NK1.1⁺ cells is probably underestimated, as not all F1-derived NK-cells express NK1.1, these data indicate that host-reactive NK-cells contribute to the eradication of host immune cells, and that this alloreactivity persists up to 3 months after BMT.

In conclusion, our data suggest that donor NK-cell alloreactivity toward host immune cells may be exploited in the setting of allogeneic BMT to treat AID in which B cells play a pathogenic role.