Acute myeloid leukaemia (AML) accounts for 20% of childhood leukaemias.1 Approximately 60 cases are diagnosed each year in Spain. Current therapeutic approaches are based on the administration of polychemotherapy, combining high-dose cytarabine with anthracyclines and topoisomerase inhibitors.2 Furthermore, patients that respond poorly to induction therapy, patients considered to be high-risk from the outset (with secondary AML or unfavourable cytogenetic characteristics), or intermediate-risk patients that have a HLA match related donor are candidates for allogeneic haematopoietic stem cell transplantation (HSCT) for AML in first complete remission.3,4

The survival of patients with AML has improved considerably in the past 40 years, mostly due to advances in supportive care. However, survival has plateaued at 60% in recent decades. Our knowledge of the genetic heterogeneity of AML and its importance in prognosis is increasing by the day.5,6 There are cytogenetic changes that carry a good prognosis, such as t(8;21)(q22;q22), inv(16)(p13.1q22) or t(16;16)(p13.1;q22), t(15;17)(q22;q12), the presence of which corresponds to an 80% survival with chemotherapy alone, while unfavourable changes such as 5q, t(6;9)(p23;q34), monosomy 7 and complex karyotypes are refractory to chemotherapy and have survival rates of no more than 40% even with HSCT.6–8

Thus far, the potential of these genetic changes as therapeutic targets has not had a significant impact on survival.9,10 Recently, and complementing advances on genetic and epigenetic phenomena, an immunobiological approach to AML is being developed.11,12 Such an approach is supported by different observations: (a) the crucial role of HSCT3,13; (b) the success of cellular adoptive immunotherapy with post-transplant donor lymphocyte infusions against minimal residual disease and/or mixed chimerism14,15; and (c) the favourable clinical and preclinical experience with antibodies such as gemtuzumab ozogamicin (anti-CD33) and bi-specific T-cell engagers (CD33/CD3, AMG 330).16,17

In this regard, the Perugia group led by Dr Velardi demonstrated the antileukaemic effect of haploidentical HSCT in adult patients with AML through the donor-derived natural killer (NK) cells.18,19 Later on, the St Jude group confirmed these findings in childhood AML and pioneered an NK cell infusion regimen for consolidation therapy in patients with low-to-intermediate risk AML.20 We know that the antileukaemic activity of NK cells is regulated by the recognition by inhibitory and activating receptors of their ligands in blast cells.21 These ligands correspond to human leucocyte antigen class I (HLA-I) for inhibitory receptors, and major histocompatibility complex (MHC) class I-related chain A/B (MICA/B) and UL16-binding proteins 1-4 (ULBP 1-4) for activating receptors.22

This study offers a retrospective review of our experience in the management of childhood AML and describes the expression of NK cell receptor ligands in 10 cases of acute leukaemia (5 myeloid and 5 lymphoid), comparing this expression in both types of acute leukaemia with the purpose of contributing additional data in support of incorporating NK cell adoptive immunotherapy to conventional chemotherapy protocols.

The results of our retrospective series are similar to those reported in studies of paediatric patients with AML both in Spain and abroad. The latest report of the Spanish Registry of Childhood Tumours (RETI) for the 1980–2013 period described an overall 5-year survival of 54% for years 1995–1999, 59% for 2000–2004 and 62% for 2005–2007. The European group NOPHO recently reported an EFS of 55% in their NOPHO-AML2004 protocol25; and the group at the St Jude's Hospital in Memphis, United States, a 5-year EFS of 63% in 230 patients during the 2000–2008 period in their AML-02 protocol.26 Based on our data, the event-free survival in childhood AML has not changed substantially with the past three versions of the SHOP protocol and plateaued more than two decades ago, remaining in suboptimal values. Furthermore, consistent with the current knowledge, the survival for secondary ALM (treatment-related or arising from a myelodysplastic syndrome) is much lower than survival for de novo ALM.27 On the other hand, the heterogeneity of AML is demonstrated by the EFS of promyelocytic AML (FAB M3), which far exceeds that of the rest of AMLs (80% vs 57%, P<.05), underscoring the positive impact of all-trans retinoic acid on survival and the need to find molecular targets in other types of AML.28

In our series, cytogenetic changes also had an important impact on prognosis. Thus, the presence in a third of our patients of changes considered to be favourable, such as t(8;21), inv(16) or (15;17), was associated with a higher survival that was probably due to a lower probability of relapse, although this result was not statistically significant due to the small sample size. The use of allogeneic HSCT as a treatment strategy has been increasing in recent protocols, along with a decrease in autologous HSCT. Advances in cytogenetic testing methods that allow the identification of high-risk changes may have played a role in this development. We found that allogeneic HSCT was associated with an improving trend in EFS, consistent with the findings of other case series.3,29 A greater sample size may have allowed us to find significant differences in our patients.

The causes of death in the patients in our series were relapse and treatment-related causes (infection and haemorrhage). The rate of toxic death has tended to decrease with successive protocols, while relapse has continued to be an important event with no change in frequency.30 The small sample size of our study did not allow us to identify prognostic factors associated with the probability of relapse. We believe that a 38% relapse rate is still too high. The probability of relapse is nearly 30% in the subset of patients with favourable cytogenetics. Thus, we believe that all patients could benefit from new therapeutic approaches.

In addition to this retrospective study, this article presents the preliminary results on the expression of ligands of inhibitory and activating receptors by malignant myeloid and lymphoid blasts, which we analysed with the aim of explaining the increased susceptibility of AML to NK cell cytotoxicity described by research groups in other countries.31 Consistent with the existing evidence, the mean fluorescence intensities of HLA-I, ligands of the KIR inhibitory receptors, were significantly lower in myeloid than in lymphoid blasts, suggesting a greater sensitivity to NK cell cytotoxicity mediated by their inhibitory receptors. When it came to the expression of MICA/B and ULBP ligands, we only found significant differences in the expression of ULBP-4, which suggests that activating pathways play a lesser role than inhibitory pathways, consistent with the findings of other authors.32 Along the same lines, it has been described that decreased expression of ligands of activating receptors in patients with AML constitutes an immune escape mechanism.33 Recent research has explored haploidentical NK cell transplantation for consolidation treatment of paediatric patients with low-to-intermediate-risk AML in whom HSCT is not indicated.20,34 The results have been quite promising, with a 0% toxicity and a 100% relapse-free survival after a median followup of more than 3 years.34 Biotechnology currently allows us to isolate NK cells from healthy donors by aphaeresis. Furthermore, two new cell products are currently being studied in Spain: NK cells stimulated with IL-15 obtained from haploidentical donors by aphaeresis (NK-IL-15), and activated and expanded NK cells (AENKs) from the peripheral blood of patients (autologous AENKs) or from haploidentical donors (haploidentical AENKs). To date, there is no evidence of significant toxicity in patients with advanced disease (>2 reports, based on the NCI-CTCAE v 3.0),35 so this cell-based therapeutic approach could be incorporated in a national prospective multicentric study, especially in patients in remission that do not meet the criteria for allogeneic HSCT.

In conclusion, the rate of HSCT in AML is still not optimal, and relapse constitutes a significant problem. Research needs to be conducted on diagnostic, therapeutic and monitoring procedures with the goal of improving on current survival. In this regard, NK cell transplantation could be an approach worth developing, considering the importance of immunological factors in AML.

The authors have no conflict of interests to declare