KIR B haplotype donors confer a reduced risk for relapse after haploidentical transplantation in children with ALL

Natural killer (NK) cell function is tuned by an array of receptors transducing either inhibitory or activating signals.¹  Among these receptors influencing NK cell function, the killer-cell immunoglobulin-like receptors (KIRs) are of particular importance. The KIR gene family consists of 17 genes (15 genes and 2 pseudogenes), each encoding for a different receptor.²  Inhibitory KIRs recognize human leukocyte antigen (HLA) A, B, and C alleles as their ligands,¹  and in humans, they show a polymorphism similar to that of the HLA system, although their genes segregate independent of HLA genes. As a consequence, every person expresses an individual KIR pattern.³  The expression of inhibitory KIRs on NK cells led to the discovery of NK cell alloreactivity in allogeneic hematopoietic stem cell transplantation (HSCT); indeed, donor NK cells can attack patient hematopoietic cells when lacking the ligand for the corresponding inhibitory KIR. A strong graft-versus-leukemia effect mediated by alloreactive NK cells, resulting in reduced risk for relapse, was documented in adult patients with acute myeloid leukemia,⁴  as well as in children with acute lymphoblastic leukemia (ALL),^5,6  undergoing T-cell-depleted HLA-haploidentical HSCT.

Activating forms of KIRs have been also identified and cloned,^1,3  but only for KIR2DS1 and KIR2DS4 has the specificity for HLA class 1 molecules been unequivocally documented. KIR genes, located on chromosome 19, are inherited as haplotypes. Two basic KIR haplotypes can be found in humans: the group A haplotype, which has a fixed number of genes encoding inhibitory receptors (with the exception of the activating receptor KIR2DS4), and the group B haplotypes, which have variable gene content and 1 or more of the B-specific genes: KIR2DS1, KIR2DS2, KIR2DS3, KIR2DS5, KIR2DL2, and KIR2DL5.^1,7,8 

Among haplotype B individuals, a KIR B-content score can further be established⁸  on the basis of the number of centromeric and telomeric KIR B haplotype motifs. Cooley et al recently reported a reduced relapse risk in adults with acute myeloid leukemia, but not in those with ALL, given allogeneic HSCT from unrelated KIR haplotype B donors.⁸  In that study, a reduced risk for relapse was also found for patients whose donor had an increased number of centromeric and telomeric KIR haplotype B motifs.

The influence of KIR haplotypes on the outcome of children with ALL given haploidentical HSCT is still unknown. Therefore, we employed quantitative real-time polymerase chain reaction (PCR) to determine the KIR profiles of donors for 85 children with high-risk or chemotherapy-refractory ALL given T-cell-depleted haploidentical HSCT, and analyzed their effect on outcome.

All patients were younger than 18 years and had received T-cell-depleted haploidentical HSCT, as previously described,^9-11  between 1996 and 2013. Patient characteristics are detailed in Table 1. DNA samples that had been previously collected and stored were used for analysis. Written informed consent was obtained from patients’ parents/legal guardians. The study was approved by the Institutional Review Board of Tübingen University and was performed in accordance with the Declaration of Helsinki.

The presence of the 17 KIR genes was investigated according to a previously published method.^12,13  Briefly, the PCR protocol consisted of a first denaturation cycle of 20 seconds at 95°C, followed by 32 alternating cycles at 95°C for 3 seconds and 64°C for 20 seconds each. H₂O was used as negative control, and Nectin and GAL-C were used as internal controls. Quantitative real-time PCR was performed using the Bio-Rad CFX96TM Real-Time System PCR Cycler.

The risk for relapse was expressed as cumulative incidence with competing risk to adjust the analysis for the competing event (namely, nonrelapse mortality), whereas the probability of event-free survival (EFS) was calculated according to the Kaplan-Meier method. The influence of patient characteristics (sex, age, ALL immune phenotype, and state of remission at transplantation) and transplantation-related factors (type of donor, conditioning regimen, type of graft manipulation, donor NK alloreactivity, KIR haplotype, and B-content score) on relapse risk and EFS was tested in both univariate and multivariate analysis. Univariate prognostic analyses used the Log-rank test (Mantel-Cox) for EFS and Gray’s test for relapse risk. Variables having a P value ≤ 0.05 in univariate analyses were included in multivariate analyses, which were performed using the proportional subdistribution hazard regression model for relapse and the Cox proportional regression model for EFS. Acute graft-versus-host disease was analyzed as a time-dependent covariate in the univariate model for relapse. Statistical analysis was performed using Graph Pad Prism 5 and the statistical software R.^14,15 

Of the 85 patients studied, 69 (81%) were in first, second, or more advanced complete remission, and 16 (19%) were not in remission at the time of transplantation. Sixty-three (74%) donors had a KIR haplotype B and 22 (26%) a KIR haplotype A. This distribution is in line with what has already been reported in Caucasian populations (www.allelefrequencies.net/).

Twenty patients (24%) developed acute graft-versus-host disease, and 12 (14%) developed chronic graft-versus-host disease. Thirty-five patients (41%) relapsed after transplantation, and 11 (13%) experienced nonrelapse mortality. Fourteen of the 22 patients (64%) transplanted from a KIR haplotype A donor relapsed compared with only 21 (33%) of the 63 patients transplanted from a KIR haplotype B donor. The cumulative incidence with competing risk analysis demonstrated a significantly reduced incidence of relapse for patients transplanted from KIR haplotype B donors (Gray's test for relapse, P = .010) (Figure 1A).

In children transplanted in complete remission, 53% (8 of 15) of the patients transplanted from a KIR haplotype A donor relapsed compared with only 28% (15 out of 54) of those transplanted from a KIR haplotype B donor (P = .062). Neither nonrelapse mortality nor relapse incidence were influenced by donor KIR haplotype in patients transplanted with active disease (data not shown).

The 5-year EFS of patients transplanted from a KIR haplotype B or KIR haplotype A donor were 50.6% and 29.5%, respectively (P = .033; Figure 1B).

We also determined KIR B-content scores for all donors according to the system proposed by Cooley et al.⁸  These results were condensed according to the Donor KIR B-content group calculator (neutral = score of 0, better = score of 1 or 2, best = score of 3 or 4) (www.ebi.ac.uk/ipd/kir/donor_b_content.html). Twenty-two (26%) patients had a neutral score, 54 (64%) had a score of 1/2, and 9 (11%) had a score of 3/4. The cumulative incidence with competing risk of relapse showed a reduced risk for recurrence with a KIR B-content score greater than 0 (Gray’s test, P = .038; Figure 1C).

The 5-year-probability of EFS for patients transplanted from a donor with a KIR B-content score of 3/4, 1/2, and 0 was 65%, 48%, and 29%, respectively (Log-rank test for trend, P = .026; Figure 1D).

In multivariate analysis, KIR haplotype B donors, being in complete remission at time of transplantation and occurrence of acute graft-versus-host disease were associated with a lower relapse risk (Table 2). These latter 2 variables were associated with better EFS in univariate analysis (data not shown).

Our results show that using a KIR haplotype B donor is associated with lower relapse risk and better EFS after T-cell-depleted haploidentical HSCT in childhood ALL. In addition, we demonstrate a correlation of donor high KIR B-content score with better probability of EFS and lower relapse risk. The influence of donor KIR genotype has been largely investigated in adults; in particular, although acute myeloid leukemia patients given T-replete HSCT from an unrelated volunteer showed a significantly improved outcome when KIR haplotype B donors were employed; this protective effect was not found in ALL.⁸  These data are in contrast with our results in children with ALL, where we could demonstrate that KIR haplotype B donors confer a greater protection from relapse than KIR haplotype A donors. This discrepancy could be explained by the different type of graft infused (replete vs depleted) and by the observation that pediatric ALL blasts have higher expression of activating KIR ligands, rendering them more susceptible to lysis by KIR haplotype B donor NK cells than adult ALL blasts.¹⁶  It is noteworthy that in a recent study by Michaelis et al, 57 adults with hematological malignancies given T-cell-depleted haploidentical HSCT were found to have a reduced risk for relapse when transplanted from a KIR haplotype B donor.¹⁷ 

In conclusion, although the number of patients analyzed is limited, this study shows that in the setting of haploidentical HSCT, a significant advantage exists for children with ALL transplanted from a KIR haplotype B donor. Therefore, KIR haplotype B donors with high B-content score should be preferentially chosen for these patients, and KIR genotyping should be included in the selection algorithm. The advantage of using a KIR haplotype B donor, especially with a high KIR B-content score, in children with ALL given unmanipulated grafts remains to be proved.

This work was supported by grants from the Stiftung für krebskranke Kinder Tübingen e.V, the Jose-Carreras Leukaemia Foundation, the Stefan-Morsch-Stiftung, and the Deutsche Forschungsgemeinschaft (SFB 685) to P.L. and to R.H. M.M. was supported by a grant from the Fortuene programme Tübingen (2021-0-0). The project was supported by research grants awarded by the Associazione Italiana per la Ricerca sul Cancro (5 x 1000 Special Grant 9962 to L.M. and F.L.), by Progetti di Rilevante Interesse Nazionale 2010 (to F.L.), by Ospedale Bambino Gesù, Roma (Progetto di Ricerca Corrente 2012-2013; to A.B. and F.L.), and by Ministero della Salute, Progetto di Ricerca Finalizzata 2010 (to F.L.).

Contribution: L.O., S.U.M., M.M., F.L., and R.H. contributed equally to this work. S.U.M. performed KIR genotyping. L.O., S.U.M., and M.M. evaluated clinical data, prepared the manuscript, and performed statistical analysis. S.U.M. and M.M. introduced the KIR genotyping assay. P.L., J.T. A.B., M.Z., and F.L. provided donor samples and clinical data. L.M. contributed to study design. F.L. and R.H. designed the study and wrote the paper.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Rupert Handgretinger, Children’s Hospital, University of Tübingen, Hoppe-Seyler-Strasse 1, 72076 Tübingen, Germany; e-mail:rupert.handgretinger@med.uni-tuebingen.de.