Transient NKG2D Blockade Attenuates Graft-Versus-Host Disease While Preserving Graft-Versus-Leukemia Effects

In allogeneic hematopoietic stem cell transplantation (HSCT), identification of mechanisms to control GVHD yet maintain GVL responses is of critical importance. One key effector cell that mediates both GVHD and GVL is the CD8⁺ T cell, which expands in response to T cell receptor (TCR) stimulation by allogeneic MHC class I molecules during allogeneic HSCT. In addition, co-stimulatory molecules facilitate the TCR-mediated activation process and the effector function of CD8⁺ T cells. Recent data suggest that NKG2D may play a co-stimulatory role in activation and in augmenting anti-tumor cytotoxic responses of CD8⁺ T cells. NKG2D is an NK cell-associated receptor that is also expressed on all human CD8⁺ T cells and on activated/memory mouse CD8⁺ T cells. NKG2D recognizes a diverse array of MHC-related ligands that are expressed by many tumors and induced on cells under stress such as myeloablative conditioning during HSCT. As the role of NKG2D in allogeneic HSCT is unknown, we hereby investigated the role of NKG2D on CD8⁺ T cells in a mouse model of GVHD and GVL.

Our results show that a large fraction (40-50%) of mouse CD8⁺ T cells inducibly express NKG2D upon activation by allogeneic MHC in vitro and in vivo. To test the role of NKG2D in GVHD pathogenesis, we employed a major MHC-mismatched mouse model of GVHD involving the transplantation of C57BL/6-derived CD8⁺ T cells and bone marrow (BM) into lethally irradiated Balb/c mice (B6→Balb/c). Using 3 different approaches to block NKG2D on CD8⁺ T cells (shRNA-mediated silencing, germline NKG2D deficiency, and antibody blockade), we found that weight loss, clinical score, and survival were significantly improved in transplanted mice with NKG2D blockade. The attenuation in GVHD correlated with a significant reduction in TNFα and IFNγ production, cytotoxicity, and proliferation (BrdU incorporation) by CD8⁺ T cells. Although CD4⁺ T cells did not express NKG2D, a protective effect of NKG2D blockade was still observed in GVHD induced by a mixture of CD8⁺ and CD4⁺ T cells, albeit to a lesser extent.

We next tested the effects of NKG2D on CD8⁺ T cell-mediated GVL. To this end, irradiated Balb/c mice were transplanted with C57BL/6-derived CD8⁺ T cells and BM, challenged intravenously with luciferase-positive A20 leukemia cells, and followed by total body imaging of luciferase-expressing cells. Given that NKG2D ligands are constitutively expressed on many tumor cells and plays an important role in their eradication, we predicted that continuous NKG2D blockade would inhibit GVL effects. However, as NKGD ligands are upregulated only transiently on stressed normal tissue, we reasoned that transient NKG2D blockade might be sufficient to attenuate GVHD and allow CD8⁺ T cells to regain their GVL function. To test this hypothesis, we compared the effect of anti-NKG2D antibody as continuous treatment or as 5-day transient treatment to mice receiving isotype control antibody. As expected, mice that received isotype control antibody cleared the A20 cells but developed severe GVHD. Continuous anti-NKG2D antibody-mediated blockade improved GVHD but also blunted the GVL response leading to increased A20 growth. In contrast, a large proportion of mice transiently treated with anti-NKG2D antibody cleared the A20 cells, while maintaining the attenuated GVHD state. Together, these data support a positive role of NKG2D on CD8⁺ T cells in mediating GVHD and GVL. Given the transient nature of NKG2D ligand upregulation on stressed tissues, a window of opportunity may exist where transient NKG2D blockade could provide a novel therapeutic strategy for treatment of acute GVHD while preserving the GVL function of CD8⁺ T cells after allogeneic HSCT.