Outcomes of Pediatric Patients with Acute Leukemia Following Adult Unrelated Donor Transplant: The Impact of Donor KIR Gene Content and KIR Ligand Matching

In patients with acute leukemia, relapse remains one of the most frequent causes of treatment failure after allogeneic transplantation (allo-HCT). Early after transplant, natural killer (NK) cells predominate and thus, they have been implicated in graft vs. leukemia (GVL) reactions after allo-HCT. Killer Ig Receptors (KIR) are a family of 15 distinct genes that both positively and negatively regulate NK cell function. Individuals vary in the number of genes that they possess within their genome and thus, may vary in their GVL capacity. Accordingly, numerous studies have investigated the association of donor KIR gene content (B/x vs. A/A), KIR B gene score (neutral, better, best), KIR composite score (taking into account CenB/x and TelA/A gene status) and KIR ligand mismatching in the setting of allo-HCT. These studies have largely focused on adult transplant recipients, yielding variable outcomes. However, pediatric patients may differ from adult transplant recipients in a variety of ways including differential sensitivity of leukemic blasts to NK cytotoxicity, distinct cytogenetic changes, the higher propensity to receive myeloablative conditioning and an earlier adoption of minimal residual disease-driven timing of transplantation. The purpose of this analysis was to investigate the impact of KIR gene content and KIR ligand mismatching on outcomes of children transplanted with acute lymphoblastic and acute myeloid leukemia (ALL and AML, respectively), reported to CIBMTR. The study cohort included pediatric patients (0-<19 years, n=716), who received a first HLA 7-8/8 matched, unmanipulated allo-HCT from an adult donor for either B precursor ALL (n=372) or AML (n=344), where donor KIR typing was available. The time period studied was from 1995-2016. For ALL and AML patients, most (95% and 86%) received myeloablative conditioning, without in vivo T cell depletion (79% and 70%). GVHD prophylaxis was with calcineurin inhibitor + MTX+/- serotherapy (79% and 76%) or calcineurin inhibitor + MMF +/- serotherapy (13% and 17%). Median follow up was 59 and 54 months for ALL and AML, respectively. The cumulative incidence of relapse for ALL and AML at 3 years was 22% (18-26%) and 33% (28-39%) and the disease free survival 58% (53-63%) and 48% (43-54%), respectively. The primary endpoints of the analysis were the impact of KIR gene content on relapse and disease free survival (DFS). Other secondary endpoints included acute GVHD, treatment related mortality and overall survival. Considering the various KIR models (Bx vs AA, KIR B gene content, and the KIR composite score (i.e., Tel A/A, Cen B) and KIR ligand matching, there was no significant association with the primary (relapse or DFS) or secondary endpoints in multivariate analysis when taking into account relevant covariates. Likewise, similar results were obtained when considering AML and ALL separately. Lastly, a subgroup analysis considering just patients who received in vivo T cell depletion also showed a lack of correlation between KIR genes and transplant outcomes. Collectively, these studies do not support the selection of an unmanipulated adult URD based on KIR gene content in pediatric AML or ALL transplant recipients.