BACKGROUND

Acute myeloid leukemia (AML) is a fatal blood cancer characterized by clonal expansion of myeloid progenitor cells through stepwise acquisition of mutations that lead to aberrant self-renewal, proliferation and differentiation.¹ The disease is genetically and clinically heterogeneous affecting all age groups. While pediatric AML is uncommon representing about 20% of pediatric leukemias, it remains the leading cause of childhood leukemic mortality. The unfavorable outcomes can be attributed to insufficient prognostic classification and limited therapeutic options, highlighting the need for a better strategy to identify high-risk patients for precision treatment and the development of novel therapies to tackle the disease.

Thanks to the recent advancements in sequencing technology, many studies have elucidated the genomic landscape of AML, leading to the identification of a considerable number of genetic lesions with important clinical implications in AML.² However, as most of these observations were obtained from adult AML studies, their relevance in pediatric AML remains uncertain, given the fact that some of the most common genetic/cytogenetic alterations in adult patients are substantially less prevalent in pediatric cases.^3-5 Accordingly, we observed that the recent 2022 European LeukemiaNet (ELN) risk classification, which was established for adult AML patients,² was suboptimal in stratifying pediatric AML patients into prognostically distinct subgroups, indicating that a separate risk stratification system is required for pediatric AML.

Compared to adult AML, detailed investigations of the mutational spectrum in pediatric AML patients have been less common. In this study, we performed deep genomic characterization of 147 pediatric patients with newly diagnosed AML to decipher the biological and clinical implications of various molecular aberrations in this deadly disorder. As the previous TARGET-AML project has revealed that pediatric AML genomes are featured by structural changes including gene fusions and focal deletions,⁶ we concurrently profiled the fusion gene landscape together with mutations and copy number changes of 141 myeloid-related genes using various molecular techniques to dissect the complex patterns and relationships among structural and non-structural genomic changes in pediatric AML (Fig. 1). This study was a joint effort with clinical and scientific colleagues from the Department of Paediatrics, The Chinese University of Hong Kong. As pediatric AML is a relatively rare disease (~10 new cases diagnosed in Hong Kong per year), we have been collecting bone marrow and blood samples from pediatric patients for over 20 years to achieve the current cohort size for completion of this study. We hope that the genomic data obtained could enable us to develop new and effective approaches for risk classification and guiding management of pediatric AML patients to improve clinical outcomes.

KEY FINDINGS

To begin with our prognostic investigations, we performed univariate analysis for individual clinicopathological and molecular variables (total 28 variables analysed) to see which ones would affect the survivals of our patients. Expectedly, several known variables such as initial treatment response and adverse cytogenomic risk emerged as significant factors. Strikingly, we found that TP53 alterations (mutations and deletions), which were identified in 6 cases, were strongly prognostic in pediatric AML. None of the TP53-altered patients could survive more than 3 years, with 5 of them (83%) died in 12 months. Importantly, TP53 alterations were found to be associated with the worst overall survival (OS) among other factors in pediatric AML (Fig. 2a). Next, based on the results from multivariate analysis, we devised a superior three-category risk scoring model incorporating initial treatment response, cytogenomic risk and TP53 gene status to predict OS in pediatric AML. Compared with the low-risk group (score=0), the hazard ratio for death was 3.53 (95% confidence interval (CI)=1.65-7.56) for the intermediate-risk (score=1) and 9.79 (95% CI=4.17-22.98) for the high-risk (score ≥2) groups. The 5-year OS rate was 81% in low-risk, 52% in intermediate-risk, and 15% in high-risk patients (Fig. 2b).

Having established altered TP53 as an adverse prognostic marker in pediatric AML, it is desirable to identify potential targets in this aggressive subgroup for therapeutic interventions. To achieve this, we performed whole transcriptome analysis to compare gene expression profiles in TP53-altered and TP53-wild-type pediatric AML patients. Such analysis showed that altered TP53 was associated with enhanced cell division control in pediatric AML. By interrogating the DepMap database with confirmatory experimental studies in AML cell lines, we further revealed that BUB1B, a core spindle checkpoint kinase governing proper chromosome separation during cell division, is a potential vulnerability in TP53-altered AML (Fig. 3). How to interpret these findings? It is known that TP53 alterations are associated with increased chromosomal instability (CIN), which promotes tumor aggressiveness. However, excessive levels of CIN can reduce fitness leading to cell death. The spindle assembly checkpoint is a major fail-safe mechanism to prevent chromosome mis-segregation during cell division. It is anticipated that TP53-altered leukemic cells are more dependent on the checkpoint system to avoid exacerbation of CIN that will ultimately compromise survival. Targeting BUB1B may serve as a CIN-inducing therapy to generate catastrophic levels of CIN to kill the leukemic cells. 

TAKE HOME MESSAGE

Unlike in adult patients in which TP53 mutation screening is required to establish disease diagnosis and risk classification,² TP53 testing has long been neglected in the diagnostic investigations of pediatric AML patients. Our findings that TP53 alterations bear significant prognostic information urge the importance of concurrent TP53 mutation and deletion analysis in the workup of pediatric AML patients. Our work also illustrated the feasibility of a single assay, i.e. targeted sequencing, to perform parallel mutational and copy number analysis to streamline TP53 gene examination. Biologically, our findings suggest BUB1B as a selective vulnerability in TP53-altered AML. Notably, the kinase has also been demonstrated to be essential in aneuploid cell lines^7,8 and more highly expressed in TP53-mutated than TP53-wild-type tumors across various cancer types,⁹ further implicating the requirement of BUB1B under TP53 impairment. Identification of selective BUB1B inhibitors may thus open new therapeutic avenues for targeting TP53-altered AML.

References

1. Grove, C. S. & Vassiliou, G. S. Acute myeloid leukaemia: a paradigm for the clonal evolution of cancer? Model. Mech. 7, 941-951 (2014).
2. Döhner, H., et al. Diagnosis and management of AML in adults: 2022 ELN recommendations from an international expert panel. Blood 140, 1345-1377 (2022).
3. Harrison, C. J., et al. Cytogenetics of childhood acute myeloid leukemia: United Kingdom medical research council treatment trials AML 10 and 12. Clin. Oncol. 28, 2674-2681 (2010).
4. Ho, P. A., et al. Leukemic mutations in the methylation-associated genes DNMT3A and IDH2 are rare events in pediatric AML: a report from the children’s oncology group. Blood Cancer 57, 204-209 (2011).
5. Rau, R. & Brown, P. Nucleophosmin (NPM1) mutations in adult and childhood acute myeloid leukaemia: towards definition of a new leukaemia entity. Oncol. 27, 171-181 (2009).
6. Bolouri, H., et al. The molecular landscape of pediatric acute myeloid leukemia reveals recurrent structural alterations and age-specific mutational interactions. Med. 24, 103-112 (2018).
7. Cohen-Sharir, Y., et al. Aneuploidy renders cancer cells vulnerable to mitotic checkpoint inhibition. Nature 590, 486-491 (2021).
8. Quinton, R. J., et al. Whole-genome doubling confers unique genetic vulnerabilities on tumour cells. Nature 590, 492-497 (2021).
9. Wang, X. & Sun, Q. TP53 mutations, expression and interaction networks in human cancers. Oncotarget 8, 624-643 (2017).