KIR B or not to be?...that is the question for ALL

In this issue of Blood, Oevermann et al report on a cohort of 85 children (median age, 10 years) undergoing T-cell–depleted haploidentical hematopoietic cell transplantation (HCT) for acute lymphoblastic leukemia (ALL), where variations in the killer-cell immunoglobulin-like receptor (KIR) gene “content” of the donor were associated with significantly less relapse and improved disease-free survival.¹ 

Although antihost alloreactivity varies among HCT donors, differences are most pronounced in the human leukocyte antigen (HLA) mismatched setting, where mismatch between donor KIR and recipient KIR ligands (ie, HLA) are more frequent. It was initially assumed that natural killer (NK) cells kill targets without restriction by the HLA complex. This paradigm changed in 1990, when Kärre and Ljunggren demonstrated that NK cell killing is directed to transformed targets “missing self” HLA.²  The loss of HLA results in a lack of inhibitory signaling and an increase in cytotoxicity. Conversely, the ligation of self HLA to inhibitory KIR prevents NK cell activation (including lysis and cytokine production) and promotes NK cell tolerance.³  These mechanisms are mediated by activating and inhibitory KIR encoded by 15 separate polymorphic genes.⁴  Individuals differ in the number of genes contained within each KIR haplotype, where KIR A haplotypes have a fixed gene content and KIR B haplotypes have variable content including 1 or more of 7 KIR B-specific genes: KIR2DS2, KIR2DL2, KIR2DS1, KIR3DS1, KIR2DS3, KIR2DS35, and KIR2DL5⁵  (www.ebi.ac.uk/ipd/kir/). Inhibitory KIR2DL2, KIR2DL3, and KIR3DL1 recognize shared motifs in HLA C1, HLA C2, and HLA Bw4. Although KIR2DS1 can recognize HLA C2 and KIR2DS2 has been reported to recognize HLA A11, the ligands for most activating KIR remain unknown. Inhibitory KIR contributes to the acquisition of NK cell function via a process called licensing, or education, in which NK cells expressing KIR that recognize self HLA acquire maximum functional capacity.⁶  Signaling through activating KIR may stimulate NK cells directly or indirectly by modification of NK cell education.⁷ 

In a large cohort of recipients of T-cell–replete unrelated adult donor HCT, Cooley et al demonstrated that donor, but not recipient, KIR B/x genotype was associated with protection from acute myeloid leukemia (AML) relapse.⁸  This protection was not observed in ALL, raising the possibility that myeloid leukemia is more susceptible to NK killing. This finding is consistent with the differential effect of KIR–ligand mismatch on outcome after haploidentical HCT for AML vs ALL, reported by the Perugia group,⁹  and may be explained by the higher density of HLA (inhibitory KIR ligand) expressed on ALL blasts and by the lower density of adhesion receptors and other NK cell-activating ligands on ALL compared with AML targets.

In the cohort described by Oevermann, the KIR B/x genotype was present in 74% of the donors and was associated with 51% disease-free survival compared with 30% in the transplants from KIR A/A donors. Moreover, donors with a higher number of KIR B genes (ie, KIR B content score⁸ ) conferred the best outcomes. The authors recommend that related donors with KIR B/x genotypes be used for transplantation for pediatric patients with ALL. These findings must be reconciled with previous reports suggesting adult ALL is less susceptible to NK cell-mediated killing. There are 3 possible explanations. First, adult patients are more likely to have high-risk features at diagnosis, defined by high-risk cytogenic changes (Philadelphia chromosome, mixed-lineage leukemia gene rearrangements, and so on). These differential genetic events could affect the sensitivity of ALL to NK cell killing, as has been recently demonstrated for AML.¹⁰  Second, the conclusions from Oevermann et al are limited to the specific setting of haploidentical HCT. Because of variations in the degree of T-cell depletion and the use of total-body irradiation, it is difficult to dissect out components of the transplant platform that may also influence outcome. Last, it is possible that the results are a result of the relatively small sample size, and therefore, these data should compel other single-center or registry studies to validate these findings.

In summary, Oevermann et al present important and provocative data demonstrating a benefit in relapse protection and survival for pediatric patients with ALL transplanted with haploidentical donors with KIR B/x genotypes. Donor selection based on KIR B content is feasible because of the high frequency of KIR B genes in this population and the lack of data suggesting any down side to avoiding KIR A/A donors. Studies in different cohorts using other transplant platforms and donor cell sources are warranted to determine the extent of the benefit associated with KIR B/x donors. These findings should motivate mechanistic analyses to determine why the response of pediatric ALL to KIR B/x donors is similar to that of adult AML, and not adult ALL.