NK cell adoptive immunotherapy

Comment on Miller et al, page 3051

In this issue of Blood, Miller and colleagues present data on the administration of haploidentical NK cells to cancer patients. They identify a feasible and safe method for in vivo NK cell expansion and clinical efficacy.

Natural killer (NK) cells are negatively regulated by inhibitory receptors that are specific for self–major histocompatibility complex (MHC) class I molecules.¹  In humans, inhibitory cell killer immunoglobulin (Ig) receptors (KIRs) recognize groups of HLA-C and HLA-B molecules (“KIR ligands”).²  Consequently, when faced with KIR ligand–mismatched allogeneic targets, KIR-bearing NK cells sense the missing expression of self–class I alleles and mediate alloreactions. In HLAhaplo-type–mismatched hematopoietic transplantation, donor versus recipient NK cell alloreactions are associated with enhanced control of acute myeloid leukemia (AML) relapse and no risk of graft-versus-host disease (GvHD).³ 

Miller and colleagues recruited patients with metastatic melanoma and renal cell carcinoma, refractory Hodgkin disease, and poor-prognosis AML. Patients received 3 different preparative regimens before they were given an infusion of allogeneic NK cells. All NK cell donors were haploidentical family members; few were KIR ligand–mismatched in the GvH direction.

This is the first study to use allogeneic NK cells as a form of adoptive immunotherapy in a nontransplantation setting. The exciting information that emerges is that adoptively transferred NK cells can be expanded in vivo. Expansion was induced only by a high-dose cyclophosphamide plus fludarabine (Hi-Cy/Flu) preparative regimen, which caused lymphopenia and high endogenous concentrations of interleukin 15 (IL-15; which is essential for the in vivo expansion and survival of NK cells).⁴  Interestingly, as the authors hypothesized on the basis of results from murine models and haploidentical human transplantations,³  the study shows that in vivo NK cell infusion/expansion does not cause GvHD.

Although tumor response was not a primary goal, 5 of 19 poor-prognosis AML patients achieved complete remission after Hi-Cy/Flu and haploidentical NK cell infusion. As the authors point out, the precise role of the cells versus the high-intensity chemotherapy regimen in responding patients cannot be definitively determined in this current study. However, preferential expansion of NK cells in responding patients suggests NK cells were implicated in disease remission. Intriguingly, a significantly higher complete remission rate was observed when KIR ligand–mismatched (ie, potentially NK cell alloreactive) donors were used.

Larger clinical trials using KIR ligand–mismatched donors are warranted. In order to optimize selection of donors with potential to exert NK cell alloreactions, such studies should focus on HLA-C KIR ligand mismatches as, besides being more common, they are associated with higher frequencies of alloreactive NK cells.⁵  KIR genotyping should be used to rule out the (rather rare) donor who lacks a KIR gene involved in HLA-C recognition.⁶  Furthermore, as frequencies of alloreactive NK cells vary at least 10-fold among donors (eg, from approximately 1 cell in 4 to approximately 1 cell in 50),⁵  frequency analyses of the alloreactive NK clone repertoire in every donor will be essential in establishing the actual dosing of the “drug”(ie, the absolute number of effector cells in each bulk NK cell preparation and, eventually, in identifying the minimum effective number of alloreactive NK cell effectors for achieving clinical objectives). It is hoped that such studies will confirm the promising results by Miller et al and that allogeneic NK cell infusions will become a safe and effective form of adoptive immunotherapy of cancer. ▪FIG1