Purpose and methods: The clinical significance of donor NK cell engraftment after nonmyeloablative conditioning for allogeneic HCT is not defined. We prospectively investigated the relationship between kinetics of donor engraftment of T-cells and NK cells and outcomes in 229 pts (130 with HLA-matched related donors and 99 with unrelated donors) conditioned with 2 Gy TBI +/− fludarabine. Diagnoses were hematological malignancies (n=217), solid tumors (n=8), or primary immunodeficiency diseases (n=4). T-cells (CD3+CD56) and NK cells (CD3CD14CD56+) were isolated from PB by flow cytometry on days 14, 28 and 42 after HCT. The proportions of cells of donor and host origin were assessed by VNTR-PCR and quantification with phosphorimaging.

Results: Median T-cell and NK cell chimerism levels were 66 (range, 1–100)% and 78 (range, 3–100)% on day 14, 77 (range, 1–100)% and 88 (range, 2–100)% on day 28, and 81 (range, 1–100)% and 88 (range, 1–100)% on day 42, respectively. The Spearman correlation coefficient between day-14 T-cell and NK cell chimerism levels was 0.59. Day-14 T-cell (P=.0005) and NK cell (P=.0003) chimerism levels < 50% were each associated with increased risks of graft rejection. Modeling chimerism levels as a continuous linear variable, high T-cell (P=.01) but not NK cell (P=.21) chimerism levels on day 14 were associated with increased probability of acute GVHD. Further, both high levels of T-cell (P=.008) and NK cell (P=.005) donor chimerism levels on days 14–42 were associated with decreased risk of relapse in time-dependent analyses adjusted for disease risk, while high levels of donor NK cell (but not T-cell) chimerism were associated with better overall (OS; P=.003) and progression-free (PFS; P=.003) survivals when adjusting for disease risk (Table 1). To further characterize the role of NK cells on relapse, we investigated the impact of missing donor KIR ligand on HCT outcome in 150 of the 229 pts (

Blood
2005
, vol
150
, p
4878
). Although no statistically significant association between missing any donor KIR ligand(s) and PFS was seen, there was a suggestion for a lower risk of relapse in patients missing 1-donor KIR ligand (HR=0.61, p=.15).

Conclusions: High levels of donor T-cell chimerism were associated with a lower risk of relapse, a benefit which was offset by higher incidence of acute GVHD leading to a trend for increased nonrelapse mortality. Conversely, earlier establishment of donor NK-cell chimerism was associated with a lower risk of relapse without increasing acute GVHD, leading to improved OS. Further studies are needed to define the role of NK alloreactivity after nonmyeloablative conditioning.

Association between donor T-cell and NK cell chimerism levels and HCT outcomes (p values).

T-cellNK cell
Chimerism levels were modeled as a continuous linear variable. 
Grade II–IV acute GVHD p=.01 (higher risk with increasing chimerism level) p=.21 (higher risk with increasing chimerism level) 
Relapse p=.008 (lower risk with increasing chimerism level) p=.005 (lower risk with increasing chimerism level) 
Nonrelapse mortality p=.13 (higher risk with increasing chimerism level) p=.26 (lower risk with increasing chimerism level) 
PFS p=.16 (better PFS with increasing chimerism level) p=.003 (better PFS with increasing chimerism level) 
OS p=.77 (better OS with increasing chimerism level) p=.003 (better OS with increasing chimerism level) 
T-cellNK cell
Chimerism levels were modeled as a continuous linear variable. 
Grade II–IV acute GVHD p=.01 (higher risk with increasing chimerism level) p=.21 (higher risk with increasing chimerism level) 
Relapse p=.008 (lower risk with increasing chimerism level) p=.005 (lower risk with increasing chimerism level) 
Nonrelapse mortality p=.13 (higher risk with increasing chimerism level) p=.26 (lower risk with increasing chimerism level) 
PFS p=.16 (better PFS with increasing chimerism level) p=.003 (better PFS with increasing chimerism level) 
OS p=.77 (better OS with increasing chimerism level) p=.003 (better OS with increasing chimerism level) 

Author notes

Corresponding author

Sign in via your Institution