Activated NK cells have a potential therapeutic role in sustaining donor engraftment following paediatric haematopoietic stem cell transplantation for non-malignant disease.

Haematopoietic stem cell transplantation (HSCT) remains the most successful treatment for conditions that originate from abnormal proliferation or function of haematopoietic stem cells. However, following donor engraftment, many patients experience a gradual loss of full donor chimaerism (FDC), with returning host haemopoiesis leading to disease relapse. Donor natural killer (NK) cells can prevent disease relapse in malignant disease (Ruggeri et al, 2002) by eradicating malignant haematological cells. Their activation was initially described in the ‘missing self’ hypothesis (Ljunggren & Karre, 1990), which stated that inhibitory receptors [killer immunoglobulin-like receptors (KIRs)] prevent NK cell activation following ligation with specific human leucocyte antigen (HLA) molecules (Ohlen et al, 1989). When such molecules are ‘missing’ no inhibitory signals occur, resulting in NK cell activation and cytotoxicity. However, the identification of NK cell activating receptors that recognize specific human antigens (HLA class I molecules) indicates that NK cells perform direct surveillance and eradication by more sophisticated means (Chewning et al, 2007). These activating KIRs recognize specific HLA molecules and, following ligation, induce direct cytotoxic responses, initiating target cell death. The latter phenomenon suggests NK cells have the capacity to promote engraftment following HSCT for non-malignant disease. We hypothesized activated donor NK cells may promote donor engraftment by the clearance of host cells and improve rates of FDC and disease eradication in non-malignant disease. We analysed differences in peripheral blood NK cell numbers between patients with FDC and mixed donor chimaerism (MDC) and determined the effect of NK cell activation capacity on engraftment and the number of NK cells. Peripheral blood was collected from 40 paediatric patients, who had achieved initial FDC after HSCT for non-malignant disease. Donor and recipient HLA were known at class I (HLA-A, -B, -C) and class II (HLA-DRB1 and -DQB1) to four digit resolution. No marrow/peripheral blood stem cell donor was <9/10 and no cord cell unit was <5/6 HLA matched. Chimaerism analysis was performed on bone marrow aspirates or peripheral blood using short tandem repeat markers (Pindolia et al, 1999). Polymerase chain reaction (PCR) products were separated by capillary electrophoresis and analysed using the 3130 Genetic Analyser (Applied Biosystems, Warrington, UK). Patients with 95–100% chimaerism were classed as having FDC, and <95% donor chimaerism was classified as MDC (Pindolia et al, 1999). HLA and KIR genotyping was performed using LABType® PCR-RSSO (One Lambda Inc, Canoga Park, CA, USA.) KIR ligand groups were assigned according to recipient HLA type (HLA-A3/A11, HLA-Bw4 and HLA-C1/2), and donor KIR genotype (including inhibitory KIR2DL1, KIR2DL2, KIR3DL1, KIR3DL2 and activating receptors KIR2DS1, KIR2DS2, KIR 3DS1). Donor inhibitory and activating KIR and corresponding recipient HLA genotype were used to generate a previously verified inhibitory and activating KIR score, based upon the potential number of donor inhibitory/activating KIRs that could be engaged with corresponding recipient HLA ligands(Sobecks et al, 2008). One hundred microlitres of whole blood was conjugated with fluorochrome bound antibodies (CD3-fluorescein isothiocyanate, CD16-phycoerythrin (PE) and CD56-PE-Cyanin 5 (all Becton Dickinson, Oxford, UK), and flow cytometric analysis was conducted using previously published methods (Baumgarth & Roederer, 2000). EXPO32 ADC Analysis software (Beckman Coulter, High Wycombe, UK) was used to analyse data. A multivariate analysis was conducted to include the covariates – donor cell source, type of conditioning, time since transplant, type of donor and activating or inhibitory scores and their influence on engraftment and NK cell numbers was analysed. P-values of ≤0·05 were considered as statistically significant. Patient characteristics are shown in Table I. The donor cell source was bone marrow in 25 patients, peripheral blood in 5 and umbilical cord blood in 10. Type of donor included matched unrelated donors (n = 20), mismatched unrelated donors (n = 8), matched sibling donors (n = 11) and one mismatched related donor. Chimaerism levels ranged from 10% to 100% donor chimaerism where 23 patients were full donor chimaeras and 17 were mixed. Elevated activating scores were associated with increased engraftment (P = 0·018) and NK cell numbers (P = 0·014). Furthermore, NK cell numbers were an important determinant of engraftment (P = 0·020; Fig. 1). Inhibitory scores did not influence engraftment (P = 0·835) or NK cell numbers (P = 0·504). Furthermore, in our cohort, time since transplant did not influence NK cell numbers (P = 0·458) or chimaerism (0·749). Differences in peripheral blood NK cell numbers between patients with full donor chimaerism and mixed donor chimaerism. Flow cytometric analysis was conducted for the total number of CD3− CD16+ CD56+ NK cells and this demonstrated a significant elevation of these cells in the peripheral blood of patients with full donor chimaerism (P < 0·001). We provide novel data indicating that FDC may be induced via selecting donor activating KIR and corresponding recipient HLA genotypes. We theorize this may lead to donor-derived NK cell activation, resulting in improved HSCT engraftment. Evidence for the role of activated NK cells in driving engraftment is apparent in malignant disease (Hercend et al, 1986; Ruggeri et al, 2002). Our findings suggest NK cell alloreactivity may also augment engraftment and moderate disease recurrence, without causing severe graft-versus-host disease in non-malignant disease. This phenomenon may be dependent on activating mechanisms, as donor NK cell inhibition did not correlate with donor chimaerism or NK cell numbers. We also found that donor chimaerism and the number of circulating NK cells were both elevated in patients with increased NK cell activating scores. Therefore it is likely that the NK cells are of donor origin, as their elevated levels are found in FDC and fall in MDC. On these grounds we propose that donor NK cells which express activating KIRs become activated via HLA ligation on host pathological cells, which may result in NK cell cytotoxicity and host cell elimination. This may be a mechanism by which donor NK cells promote engraftment and prevent disease relapse in non-malignant disease. There are limitations to our study, which include a heterogeneous cohort and the relatively small sample size. Furthermore, our analysis remains cross-sectional and given that NK cell numbers and chimaerism can change over time, a longitudinal analysis would be interesting. The time since transplant did not influence NK cell numbers in our cohort and this is most probably due to the cross-sectional nature of the study. Having said this, our observational results indicate a potentially novel role for NK cells in HSCT for non-malignant disease and give an indication for further research in a larger cohort, minimizing heterogeneity. The authors have no conflicts of interest to disclose.