Donor natural killer cell allorecognition of missing self in haploidentical hematopoietic transplantation for acute myeloid leukemia: challenging its predictive value

Abstract

We analyzed 112 patients with high-risk acute myeloid leukemia (61 in complete remission [CR]; 51 in relapse), who received human leukocyte-antigen (HLA)-haploidentical transplants from natural killer (NK) alloreactive (n = 51) or non-NK alloreactive donors (n = 61). NK alloreactive donors possessed HLA class I, killer-cell immunoglobulin-like receptor (KIR) ligand(s) which were missing in the recipients, KIR gene(s) for missing self recognition on recipient targets, and alloreactive NK clones against recipient targets. Transplantation from NK-alloreactive donors was associated with a significantly lower relapse rate in patients transplanted in CR (3% versus 47%) (P > .003), better event-free survival in patients transplanted in relapse (34% versus 6%, P = .04) and in remission (67% versus 18%, P = .02), and reduced risk of relapse or death (relative risk versus non-NK-alloreactive donor, 0.48; 95% CI, 0.29-0.78; P > .001). In all patients we tested the "missing ligand" model which pools KIR ligand mismatched transplants and KIR ligand-matched transplants from donors possessing KIR(s) for which neither donor nor recipient have HLA ligand(s). Only transplantation from NK-alloreactive donors is associated with a survival advantage.

Figure 1

Reconstitution of potentially alloreactive KIR⁺/NKG2A⁻ NK cells of donor origin after haploidentical transplantation. Posttransplantation kinetics of KIR⁺/NKG2A⁻ (CD56⁺/CD3⁻) NK cell subsets versus KIR⁻/NKG2A⁺ (CD56⁺/CD3⁻) NK cell subsets of donor origin in a sample population of 10 recipients during the first 1 to 7 months after transplantation (mean ± SD) compared with values in their donors (left). KIR⁻/NKG2A⁺ cells (open triangles), KIR2DL2/3/S2⁺/NKG2A⁻ cells (open circles), KIR2DL1/S1⁺/NKG2A⁻ cells (open squares), KIR3DL1/S1⁺/NKG2A⁻ cells (closed circles).

Figure 2

Pretransplantation and posttransplantation immunofluorescence analyses in 2 representative transplant pairs of NK cells of donor origin expressing the KIR for which there is no class I ligand in the recipient as their only inhibitory receptor for self. Panels A (donor) and B (recipient 1–3 months after transplantation) illustrate a transplant pair in whom the donor KIR ligand HLA-C 1 was missing in the recipient. KIR genotyping showed the donor possessed only group A haplotype genes, specifically KIR2DL1, KIR2DL3, and KIR3DL1 inhibitory KIR genes. Consequently the potentially alloreactive population is represented by cells expressing only KIR2DL3 (upper left quadrants in panels A and B where the percentage is indicated). Cells expressing or coexpressing all other KIRs and/or NKG2A are shown in the right quadrants. Panels C (donor) and D (recipient 4-7 months after transplantation) illustrate a transplant pair in whom the donor KIR ligand HLA-C 2 was missing in the recipient. KIR genotyping showed the donor possessed group B haplotype genes, specifically KIR2DL1, KIR2DL2, KIR2DL3, and KIR3DL1 inhibitory genes and KIR2DS1, KIR2DS2, KIR2DS3, and KIR3DS1 activating genes. Consequently, the potentially alloreactive population is represented by cells expressing only the KIR2DL1 inhibitory receptor (lower right quadrants in panels C and D where the percentage of cells expressing KIR2DL1 with or without KIR2DS1 is indicated). Cells expressing or coexpressing all other KIRs or NKG2A or both are shown in the upper quadrants.

Figure 3

Transplantation from haploidentical NK alloreactive donors controls AML relapse in patients transplanted in any remission. (A) Relapse in patients transplanted in chemoresistant relapse from NK alloreactive versus non-NK alloreactive donors. (B) Relapse in patients transplanted in any remission from NK alloreactive versus non-NK alloreactive donors.

Figure 4

Transplantation from haploidentical NK alloreactive donors improves EFS. (A) EFS in patients transplanted in relapse from NK-alloreactive versus non-NK alloreactive donors. (B) EFS in patients transplanted in CR from NK alloreactive versus non-NK alloreactive donors.

Figure 5

EFS according to NK cell alloreactivity and to specific KIR ligand mismatches. (A) EFS after transplantation from NK-alloreactive donors versus non-NK alloreactive donors in the entire series of 112 haploidentical transplant recipients. (B) EFS after transplantation from NK-alloreactive donors according to the specific donor KIR ligand that was missing in the recipient. Solid line indicates HLA-C 1; dotted line, HLA-C 2; broken line, HLA-Bw4.

Figure 6

EFS is better predicted by NK alloreactivity than by the missing ligand model. (A) The missing ligand model breaks the non-NK alloreactive EFS curve, shown in Figure 5A, into 2 curves according to whether donors possessed an extra KIR(s) for which neither donor nor recipient had HLA ligand(s), ie, missing ligand, versus no missing ligand. The 2 curves do not differ significantly. (B) In the missing ligand model our NK alloreactive transplants (KIR ligand mismatched) are combined with the missing ligand transplants and compared with the no missing ligand transplants. No significant difference emerged (P = .28). EFS in the missing ligand cohort was worse than after transplantation from NK alloreactive donors (B versus Figure 4).