Abstract
Abstract 4150
Natural killer (NK) cell alloreactivity has been implicated in graft-versus-host (GVH), graft-versus-leukemia (GVL), and host-versus-graft reactions following hematopoietic stem cell transplantation (HSCT). Though the basic mechanisms of NK cell alloreactivity are well described, comparatively little is known about its clinical effects. In umbilical cord blood (UCB) transplantation, both stem cells and donor-derived T-cell numbers are limited compared to other transplant modalities, so NK cell effects may be more pronounced. NK cell-mediated cytotoxicity is mediated by a balance of activating and inhibitory cell-cell contacts, the most important of which entails engagement of MHC-I on target cells by inhibitory killer immunoglobulin-like receptors (KIR) on NK cells. This interaction inhibits NK cell cytotoxicity, whereas the absence of MHC-I on target cells stimulates cytotoxicity, thus providing a mechanism for immune surveillance against pathologic states in which MHC-I is down-regulated such as cancer and viral infection. Applied to the scenario of HSCT, based on the specificity of various KIR for MHC-I encoded by particular HLA-B and HLA-C alleles, GVH NK cell alloreactivity can be predicted from the HLA genotypes of donors and recipients. The clinical impact of predicted GVH NK cell alloreactivity, termed KIR ligand incompatibility, has been investigated retrospectively in various transplant settings (Verheyden 2008), including UCB transplantation (Brunstein 2009, Willemze 2010), with inconsistent results regarding its association with relapse, GVH disease incidence, and survival. Methods: We stratified a cohort of 80 patients who underwent double umbilical cord blood (UCB) transplantation according to the presence or absence of KIR ligand incompatibility in the graft-versus-host direction. The median age was 48 years. Patients had a mix of myeloid and lymphoid malignancies, the most common of which were AML (31%) and non-Hodgkin lymphoma (18%). 59% of patients received non-myeloablative conditioning. Results: We found no association between KIR ligand incompatibility and neutrophil (p=0.30) or platelet (p=0.95) engraftment, incidence of acute (p=0.70) or chronic (p=0.44) GVHD, relapse (p=0.67), or progression-free (p=0.73) and overall (p=0.71) survival. Graft rejection, defined as neutrophil engraftment with donor-predominant chimerism followed by loss of donor chimerism, occurred in six patients in the cohort, all but one of whom received KIR-ligand-incompatible UCB units (p=0.08). Of these six patients, the only one who received a KIR-ligand-compatible transplant received a far less immunosuppressive conditioning regimen than the rest of the cohort due to comorbidities. When this patient was retrospectively excluded from the analysis, the association between graft rejection and KIR ligand incompatibility was statistically significant (p=0.01). An analysis accounting for the dominance of one cord unit and its associated KIR incompatibility was not performed due to the concordance of KIR ligand compatibility/incompatibility among transplanted cord unit pairs. Conclusions: These results suggest a role for NK cell alloreactivity in mediating graft rejection in double UCB transplantation but no effect on relapse or survival. These findings warrant further evaluation in larger cohorts.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.