Natural killer cells in graftversus-host disease and graft-versus-leukemia - Immune Biology of Allogeneic Hematopoietic Stem Cell Transplantation - page 333

identification of the various inhibitory receptors that regulate self-toler-
ance and provided an explanation for “hybrid resistance,” in which
parental BM allografts are resisted by their F1 hybrids [2] . The concept
of NK licensing was also postulated after a study of mouse NK cells that
displayed differential inhibitory receptors for MHC [14] .

The promise of using NK cells in cancer immunotherapy was initially
described using mouse models of allogeneic HSCT. Mouse NK studies
demonstrated the efficacy of using NK cells to promote graft-versus-tumor
effects while inhibiting graft-versus-host disease (GVHD) in allogeneic
HSCT [50] . As NK cells do not initiate the rejection of solid tissue allografts,
it was reasonable to hypothesize that the solid organ tissues often targeted
in GVHD (skin, gut, liver) would not be targets of donor NK cells. However,
NK cells can be found in GVHD lesions, suggesting that they may contribute
to the pathology once initiated. In nonmyeloablative conditioning, donor
NK cells may suppress host hematopoiesis. Indeed, it has been suggested
that donor NK cells may actually allow for less conditioning to be given to
the recipient and promote donor myeloid engraftment [51] . Mouse mod-
els have also demonstrated that donor-type NK cells can act as “veto” cells
and inhibit host effector cells (T and NK cells) capable of mediating graft
rejection [52] . Thus, the use of donor-type NK cells can promote donor
engraftment not only by suppressing host hematopoietic cells but also by
suppressing host resistance to the donor graft.

337

Another potential advantage of using NK cells is the anti-tumor capability
of these cells, particularly in hematopoietic malignancies. NK cells have also
been demonstrated to mediate resistance to metastatic spread in numer-
ous mouse models, suggesting that solid tumors can also be potentially tar-
geted. The receptors that NK cells use in recognition of tumors are focused
on NKG2D and the Ly49 receptor systems. Inhibitory Ly49 molecules have
been thought to inhibit NK cell anti-tumor responses. Blockade of these
receptors using Fab fragments has been shown to promote anti-tumor effects
without inducing myelosuppression in mouse models [53] , suggesting that
these approaches may be of use in the clinic. The implication of activating
NK receptors, such as NKG2D or DNAM-1, in tumor surveillance has also
been demonstrated in mouse models. For example, an earlier and more
aggressive generation of spontaneous prostate tumors in the transgenic ade-
nocarcinoma of mouse prostate model of autochthonous prostate cancer
development was observed in mice defective for NKG2D [54] . Furthermore,
downregulation of NKG2D expression has been detected in cancer patients
and correlated with poor cytotoxic NK function. Tumor cells can promote
NKG2D downregulation by multiple means, such as release of soluble NKG2D
ligands, demonstrating the importance of this activating receptor in control-
ling tumor growth. Accelerated tumor growth was also observed in mice defi-
cient for DNAM-1 [55] . As NKG2D ligands also are expressed on proliferating
normal hematopoietic cells, these may also be a cause of myelosuppression.

Finally, there is increasing evidence that NK cells play important roles in
immunoregulation. There is ample evidence in mouse models that NK cells
can both promote and inhibit T-cell responses, either directly or indirectly.
NK cells have been demonstrated to inhibit primary immune responses
through their ability to kill dendritic cells [56] . Indeed, when activated,