Reduced Risk of Relapse In Pediatric ALL After Haploidentical Transplantation of T-Cell Depleted Grafts From KIR Haplotype B Donors,

Abstract 4133

The Killer immunoglobulin-like receptor (KIR) gene family consists of 17 genes (15 KIR genes and 2 pseudogenes) encoding for the different receptors and is a key regulator of Natural Killer (NK) cells. Their diversity is similar to the HLA-system, however they segregate independently from the HLA-system and every human being possesses an individual pattern of KIR genes which can be assigned to either KIR haplotype A or B. Haplotype A is characterized by a relatively fixed gene content of inhibitory receptors. The presence of at least one activating KIR-receptor is necessary for defining haplotype B.

The importance of the KIR system has been demonstrated in haploidentical as well as in HLA-matched transplantation in adult patients with AML, but not with ALL. Patients who received a haploidentical transplant from a KIR ligand-ligand mismatched donor had a significant lower risk of relapse. In more recent studies in adult patients with AML but not with ALL, a significant influence of the donor KIR haplotype was demonstrated after allogeneic transplantation from HLA-matched sibling or unrelated donors, and patients who received grafts from haplotype B donors had a significant lower risk of relapse. In contrast to the results in adult patients, it has been shown in pediatric ALL that children who received an haploidentical transplant from KIR receptor-ligand mismatched donors had a better event-free survival.

The influence of the donor KIR haplotype in pediatric haploidentical transplantation in patients with ALL has not yet been investigated sofar. We have therefore analyzed the influence of the donor KIR haplotype on the risk of relapse in 48 children with very high risk ALL, who received an allogeneic transplant from a haploidentical donor at our center. The presence or absence of the 15 KIR genes was determined by real-time PCR as described by Alves et al. (Tissue Antigens 2008) and donors were assigned the A/A or B/x haplotype. Genotypes for the centromeric (Cen) or telomeric (Tel) parts of the KIR locus were assigned according to the absence or presence of one or more B haplotype defining KIR genes.

Most of the patients (n=28) were conditioned with a non-TBI-based reduced intensity regimen, whereas 20 patients received a TBI-based preparative regimen. At time of transplant, 38 patients were in remission and 10 children had refractory leukaemia. The graft comprised mobilised peripheral stem cells depleted ex-vivo from T-cells by either CD34+ positive selection or CD3/19 negative depletion. Of the 48 donors, 12 had KIR haplotype A and 36 KIR haplotype B. Patients grafted with a KIR haplotype A donor were more likely to relapse (odds ratio 4.90; p=0.02), whereas patients who received a graft from a KIR haplotype B donor had a five-fold lower risk of relapse.

We also analysed the risk of relapse according to the haplotype B-content-score system (from 0 to 4) described by Cooley et.al (Blood 2010). Briefly, the B-content score is calculated by adding the number of CenB and/or TelB motifs in each genotype.We found a reduced risk of relapse with an increasing B-content score, with 58% of relapse for a score of 0 (n=12), 31% for a score of 1 (n=16), 20% for a score of 2 (n=15) and 0% for a score of 3 (n=5). In addition, the risk of relapse was much lower in patients grafted with a donor TelB haplotype (odds ratio 7.3, p<0.01).

We conclude from our results that a donor KIR B haplotype with a high B-content score should be favored as haploidentical donor and that KIR genotyping should be implemented in the donor selection algorithm in haploidentical transplantation for children with ALL.