Impact of KIR/HLA Incompatibilities after Posttransplant Cyclophosphamide Based T Cell-Replete Haploidentical Hematopoietic Stem Cell Transplantation

Introduction: Posttransplantation cyclophosphamide (PTCy) based T cell-replete haploidentical (haplo) hematopoietic stem cell transplantation (HSCT) is a valid option for patients with indication for allogeneic HSCT without a human leucocyte antigen (HLA) matched donor. However, selection criteria to determine the optimal among several available haplo donors are still a matter of debate. Especially, the impact of killer cell immunoglobulin-like receptor (KIR)/human leukocyte antigen (HLA) incompatibilities (inc) in the setting of PTCy T cell-replete haplo HSCT is unclear. PTCy has been reported to eliminate most mature donor NK cells infused with the graft, including single KIR⁺ NK cells, thereby blunting NK cell alloreactivity in this setting (Russo et al., Blood 2018). Willem et al. (J Immunol 2019) reported (i) a significant loss of KIR2DL2/3+ NK cells at day +30 in patients with inhibitory KIR/HLA incompatibility (inc.) suggesting that PTCy might target responsive KIR NK cells and (ii) a correlation of genetic KIR2DL/HLA inc. with less relapse, but more graft-versus-host-disease (GvHD). Similarly, NK alloreactivity defined as KIR receptor-ligand mismatch or group B KIR haplotype with the presence of KIR2DS2 has been correlated with improved survival (Salomon et al., BBMT 2018).

Aims of our study were to evaluate the impact of (i) HLA/KIR inc, (ii) donor KIR genotype and (iii) HLA-DP mismatch status on survival and incidence of relapse, acute and chronic GvHD in our homogeneously treated, independent patient cohort.

Patients and methods: We retrospectively analyzed the outcome of 51 consecutively transplanted patients (AML/MDS (n=28/5), ALL (n=9), HD (n=2), NHL (n=5), CML (n=1), PMF (n=1)) receiving a PTCy based T cell-replete haplo HSCT between 01/2011-12/2018. All patients received a myeloablative conditioning regimen (fludarabine/total body irradiation (FTBI) or thiotepa/busulfan/fludarabine (TBF)) with unmanipulated bone marrow (98%) as the preferred graft (median CD34+ cells: 3.02 x 10⁶/kg (range, 1.50-6.90) and median CD3+ T cells: 3.54 x 10⁷/kg (range 1.52-43.74)). GvHD prophylaxis with ciclosporin A started on day 0, mycophenolate-mofetil on day +1, PTCy was applied on day +3 and +5.

Results: Patient, donor and transplant characteristics as detailed in table 1 were well balanced between the inh. KIR/HLA inc. group (n=29) vs. no inh. KIR/HLA inc. group (n=22) with the exception of the median donor age (41.7 (range, 23.4-73.7) vs. 33.6 years (range, 19.0-56.2), resp. All patients engrafted. At day +28 (range, 20-29; n=26) CD3+ cells were 88.5/nL (range, 3-665), CD3+CD4+ cells 22.5/nL (range, 0-277.0), CD3+CD8+ cells 117.0/nL (range, 7-478), CD19+ cells 1.0/nL (range, 0-12), CD56^bright cells 74.4/nL (range11.1-93.4), CD56^dim cells 25.5/nL (range, 6.4-88.9) measured by flow cytometry and without differences between the inh. KIR/HLA inc. group vs. no inh. KIR/HLA inc. group. Cytomegalovirus (CMV) reactivation occurred in 73.3% of patients at risk and median time of occurrence was 32 days (range, 12-97) without difference between groups. Median follow-up for surviving patients was 26.1 months (range, 2.8-92.8) and we found no significant differences in 2-year overall survival (OS; 65.3±10.3 vs. 89.6±7.0, p=0.311), 2-year relapse-free survival (RFS; 66.0±9.4 vs 77.8±10.2, p=0.235), GvHD- and relapse-free survival (GRFS; 48.4±9.8 vs 60.5±12.0, p=0.182) as well as cumulative incidence (CI) of relapse (23.3% vs 16.2%, p= 0.283), acute GvHD grade 2-4 (27.6% vs 31.8, p=0.563), moderate-severe chronic GvHD (22.2% vs. 9.9%, p=0.227) and NRM (16.3% vs 5.3%, p=0.283) between the inh. KIR/HLA inc. group vs. no inh. KIR/HLA inc. group. This was also the case for donor KIR genotype AA vs AB (n=46; 2-y OS: 74.9±13.0% vs. 73.0±9.9%, p=0.844; 2-y RFS: 60.0±14.8% vs 74.5±8.4%, p=0.645) and HLA-DP-identical/permissive mismatch (MM) vs non permissive MM (n=45; 2-y OS: 70.7±10.0% vs 72.7±13.4%, p=0.945; 2-y RFS: 73.2±8.2% vs 63.6.0±14.5%, p=0.798)

Conclusion: Our outcome data support the hypothesis of PTCy eliminating mature donor NK cells infused with the graft and thereby reducing the impact of alloreactivity in this setting. However, our patient number is quite small and the findings need to be validated in larger cohorts and preferably prospective studies.

View largeDownload slide

View largeDownload slide

Close modal