A Direct Comparison, and Prioritisation, of the Immunotherapeutic Targets Expressed by Adult and Paediatric Acute Myeloid Leukaemia Cells: A Systematic Review and Meta-Analysis

Abstract

Acute myeloid leukaemia (AML) is characterized by impaired myeloid differentiation resulting in an accumulation of immature blasts in the bone marrow and peripheral blood. Although AML can occur at any age, the incidence peaks at age 65. The pathobiology of AML also varies with age with associated differences in incidence, as well as the frequency of cytogenetic change and somatic mutations. In addition, 5-year survival rates in paediatrics are 60-75% but fall to 5-15% in older AML patients. This systematic review aimed to determine whether the altered genes in AML affect the same molecular pathways, indifferent of patient age, and, therefore, whether patients could benefit from the repurposing drugs or the use of the same immunotherapeutic strategies across age boundaries to prevent relapse. Using a PICO framework and PRISMA-P checklist, relevant publications were identified using five literature databases and assessed against an inclusion criteria, leaving 36 articles, and 71 targets for therapy, for further analysis. QUADAS-2 was used to determine the risk of bias and perform a quality control step. We then priority-ranked the list of cancer antigens based on predefined and pre-weighted objective criteria as part of an analytical hierarchy process used for dealing with complex decisions. This organized the antigens according to their potential to act as targets for the immunotherapy of AML, a treatment that offers an opportunity to remove residual leukaemia cells at first remission and improve survival rates. It was found that 80% of the top 20 antigens identified in paediatric AML were also within the 20 highest scoring immunotherapy targets in adult AML. To analyse the relationships between the targets and their link to different molecular pathways, PANTHER and STRING analyses were performed on the 20 highest scoring immunotherapy targets for both adult and paediatric AML. There were many similarities in the PANTHER and STRING results, including the most prominent pathways being angiogenesis and inflammation mediated by chemokine and cytokine signalling pathways. The coincidence of targets suggests that the repurposing of immunotherapy drugs across age boundaries could benefit AML patients, especially when used in combination with conventional therapies. However, due to cost implications, we would recommend that efforts are focused on ways to target the highest scoring antigens, such as WT1, NRAS, IDH1 and TP53, although in the future other candidates may prove successful.

Keywords: acute myeloid leukaemia; adult; immunotherapy; paediatric.

Figure 1

Prisma-P flow diagram to show the screening process that determined whether articles met the inclusion criteria at each stage; if not, they were removed. In total, 36 articles, describing 71 targets, were identified for the quality assurance step.

Figure 2

Venn diagram to illustrate which of the affected genes within adult and paediatric AML overlap with one another. Those that lie within the overlapping intersection should be assessed as potential targets for unified immunotherapy treatment approaches.

Figure 3

Prioritisation of the 63 potential treatment targets identified in AML by systematic review and found in articles that remained after quality control. Prioritisation was performed using the National Cancer Institute Pilot Project as a model [10], based on nine predefined, pre-weighted criteria. The antigens were scored to determine their suitability for further investigation as therapeutically effective cancer vaccines.

Figure 4

PANTHER was used to analyse the pathways that are linked to the 20 highest scoring targets after prioritization in both adult and paediatric AML. Pathways with the highest number of genes linked to them should be investigated further as key pathways for AML development. From those tested, only one pathway differs between paediatric and adult AML, which suggests strong similarities between the two and increases the possibility of an overlapping treatment.

Figure 5

Protein–protein interaction analysis was performed via STRING with confidence levels set to 0.7 and the species set to Homo sapiens. In both figures, the top five targets (red nodes) are shown after prioritization for adult and paediatric AML. (A) STRING network of the 20 highest scoring potential therapeutic targets for adult AML identified after prioritization. TP53 had the most connections for the adult network. (B) STRING network of the 20 highest scoring potential therapeutic targets for paediatric AML identified after prioritization. PTPN11 had the most connections for the paediatric network.