Selecting Haploidentical Transplant Donors for Deploying Optimal NK Cell Allotherapy.

Abstract 511

Donor-versus-recipient NK cell alloreactivity has been established as a key therapeutic element in HLA haplotype mismatched hematopoietic transplants in adult AML (Ruggeri et al. Blood 1999; Science 2002; Blood 2007; Stern et al., Blood 2008) and paediatric ALL (Pende et al. Blood 2009). It is effected by donor NK cells which express inhibitory KIRs for self class I KIR ligands (HLA-C1, C2, Bw4). In KIR ligand mismatched recipients, they sense missing expression of donor KIR ligand(s) and mediate alloreactions. Recent studies have proposed refinements to this mechanism and have emphasized the role of donor activating KIR variants in the control of leukemia relapse after haploidentical (Pende et al. Blood 2008) and unrelated (Cooley et al. Blood 2009) hematopoietic cell transplantation. ∼25% of Caucasians are homozygous for A haplotypes which contain inhibitory KIR genes and the KIR2DS4 activating KIR (non-functional in 2/3 of individuals). 75% of Caucasians are either heterozygous or homozygous for B haplotypes which carry not only inhibitory KIRs but also various combinations of activating KIRs (KIR2DS1-2-3-5 and KIR3DS1). Our extensive functional assessments of donor NK clone repertoires from representative sets of A (homozygous) or B (homozygous and heterozygous) haplotype donors, involving functional analyses of >5,000 NK clones from 70 donors, revealed that frequencies of alloreactive NK cell clones did not differ significantly in A vs B haplotype donors, that NK alloreactive repertoires in B haplotype positive individuals were largely composed of clones which did not express activating KIR(s) and, finally, that activating KIR-negative clones killed as efficiently as the positive. In contrast, dramatic differences emerged when inflammatory cytokine production was assessed in alloreactive NK clones with possessed or did not possess activating KIR variants. Upon incubation with KIR ligand-mismatched LPS or Aspergillus or CMV antigen-treated DCs, NK clones expressing activating KIR receptors produced several-fold more IFN-gamma and TNF-alfa than NK clones not expressing activating KIRs. The clinical role of donor activating KIR genetics was evaluated in 86 haploidentical transplants for AML. 49 recipients were transplanted from NK alloreactive (KIR ligand-mismatched) donors (12 with group A KIR gene haplotypes vs 37 with B haplotypes) and 37 recipients from non-NK alloreactive (KIR ligand-matched) donors (8 with group A KIR gene haplotypes vs 29 with B haplotypes). The remarkable GvL effect of NK alloreactive transplants (Ruggeri et al. Blood 2007) was unaffected by donor A vs B KIR gene haplotypes. In contrast, in transplants from NK alloreactive donors, presence of group B haplotype KIR genes in the donors was associated with reduced incidence of TRM (largely infection-related) (B vs A haplotypes: 20% vs 67% TRM, p<0.01). In multivariate analyses it was the only significant variable predicting protection from TRM (RR: 0.24; 95% CI:0.14-0.42; p<0.01). When the number of activating KIR genes in the donor was taken into account, donors carrying ≥ 3 activating KIR genes provided significant protection from TRM and significantly better EFS compared with A haplotype donors (TRM: 12% vs 67%, p<0.003) (EFS: 71% vs 33%, p=0.02). In multivariate analysis, transplantation from alloreactive donors carrying ≥3 group B haplotype activating KIR genes was the only variable predicting protection from TRM (RR: 0.36; 95% CI: 0.25-0.54; p<0.01) and tended to improve EFS (RR: 0.64; 95% CI: 0.38-1.07; p<0.1). We conclude that NK alloreactive, KIR ligand-mismatched donors who possess activating KIRs may confer protection against infections through their enhanced NK cell cytokine secretion upon interaction with recipient pathogen-infected DCs.