Platelet Engraftment (PE) in AML Patients Receiving HLA Matched Sibling Donor Allogeneic Bone Marrow Transplant (ABMT) Correlates with Inhibitory HLA Ligand Groups for Killer Immunoglobulin-Like Receptors (KIRs).

Abstract

The interaction of KIRs with target cell HLA class I molecules regulates the activity of NK cells and some T cell populations. KIR interactions are reported to influence allogeneic hematopoietic stem cell transplant outcomes, particularly for AML. We have previously reported that AML pts homozygous for C1 or C2 have superior survival and lower relapse rates than do pts heterozygous for C1/C2. However, the influence of KIR matching on PE has not been well described. We therefore analyzed the KIR ligand phenotypes of 60 AML pts who received HLA identical sibling donor myeloablative ABMT from 4/9/97 – 11/5/03. The median age was 45 years (range, 8–62 yrs). At transplant a minority (40%) were in CR. All pts received a busulfan/cyclophosphamide based preparative regimen and T-cell replete bone marrow as their stem cell source. Patient HLA KIR ligands were categorized as: 1) HLA-Cw group C1- or C2 - homozygous vs C1/C2 heterozygous; and 2) HLA-Bw4 (positive or negative) (reviewed in Farag et al Blood 100:1035, 2002). Kaplan Meier estimates of median time to PE >20 K/μL and >50 K/μL were 23 and 30 days, respectively. PE was next assessed in relation to the inhibitory HLA KIR ligand group expressed. PE >20 K/μL was superior for those C1 or C2 homozygous (n=26) compared to C1/C2 heterozygotes (n=34) (median 21 vs 26 days, p=0.049) and 31 Bw4 negative pts had superior PE compared to 29 Bw4+ pts (median 21 vs 30 days, p=0.012); these findings remained significant in multivariable analysis. A similar analysis performed for PE >50 K/μL found that Bw4 negative pts had superior PE compared to Bw4 + pts (median 26 vs 38 days, p=0.015); this remained significant in multivariable analysis. 57 cases had KIR genotyping performed for those KIRs with established HLA ligands and there were no cases in which the donor did not have at least one inhibitory KIR gene specific for expressed HLA ligands. Age at transplant, number of prior chemotherapy regimens, disease status at transplant and CMV status were not predictive of PE. Since both C1/C2 heterozygosity and Bw4+ status correlated with reduced early PE, the possible interaction of these two variables was next investigated. The analysis for PE >20 K/μL and >50 K/μL suggested an additive effect: pts lacking expression of both these variables had the most rapid PE, while those who expressed both variables had the slowest PE. Those who expressed only 1 of the variables had an intermediate time to PE as shown below:HLA KIR Ligand StatusNMedian CD34+ cell dose (x10^6/kg)Median Total Nucleated Cell Dose (x10^8/kg)Median time to PE>20K/μLMedian time to PE >50K/μLC1/C2 and Bw4 negative142.022.6419 days25 daysC1/C2 positive and Bw4 negative OR C1/C2 negative and Bw4 positive291.922.6223 days29 daysC1/C2 and Bw4 positive171.872.6031 days41 daysp= 0.54p=0.82p=0.003p=0.003These results may suggest that donor NK cells control host effector cells that delay PE. When minimal opportunity for inhibitory KIR engagement exists (C1/C2 negative, Bw4 negative) maximal NK cell control is expected and rapid PE ensues. When maximal opportunity for inhibitory KIR engagement exists (C1/C2 positive, Bw4 positive) donor NK cell controlling activity would be reduced, leading to delayed PE.