Aggressive Aneuploid Acute Myeloid Leukemia Is Dependent on Alterations of P53, Gain of APC and PLK1 and Loss of RAD50

Chromosome number alterations, aneuploidy, is a hallmark of cancer. It occurs in about 15% of acute myeloid leukemia (AML) cases, is generally preserved throughout disease progression (Bochtler et al. Leukemia 2015) and correlates with adverse prognosis (Breems et al. JCO 2008, Papaemmanuil et al. NEJM 2016). This evidence highlights the need of understanding the molecular mechanisms that promote and sustain aneuploidy in AML, in order to define novel potential therapeutic targets.

In the NGS-PTL project we profiled the genomic landscape of 536 hematological samples by whole exome sequencing (WES, Illumina). Among them, we analyzed 88 and 68 samples from aneuploid (A-) FLT3-wildtype AML (isolated trisomy and monosomy, complex and monosomal karyotype) and euploid (E-) AML (normal and complex karyotype, <3 abnormalities), respectively (100 bp, paired-end). Variants were called by MuTect and Varscan 2.0. WES output was integrated with genotype data (CytoScan HD Array, Affymetrix and Nexus Copy Number analysis) and gene expression profiling (HTA 2.0 and TAC 3.0, Affymetrix).

A-AML showed an increased genomic instability, as confirmed by a higher mutation load compared with E-AML (median number of variants: 22 (range: 2-95) and 11 (range: 3-45), respectively, p<.001), which was associated with increased patients' age (median age of 62 for A-AML and 55 for E-AML, p<.05). The increased age and mutation load correlated with a mutational signature with prominence of C>A substitutions, compared with the C>T transition-related signature, which is prevalent in AML. A-AML was associated with mutations and/or heterozygous deletion of TP53 (p<.001), which co-occurred with copy number loss of both the tumor suppressor APC and the DNA repair gene RAD50 in 93% of cases (p<.001). Moreover, A-AML was enriched for a gene expression signature of p53-deficiency, independently of TP53 structural defects (p<.05, GSEA). Mutations and deregulated expression of genes involved in cell cycle contributed to the A-AML phenotype, with 68% A-AML vs. 32% E-AML patients (p<.01) carrying at least one genomic lesion affecting the process. The alterations targeted the following pathways: DNA repair (i.e. reduced RAD50 expression, p<.001), cell cycle checkpoints (i.e. mutated CHK2), regulation of PLK1 activity at G2/M transition (i.e. mutation and 2-fold upregulation of PLK1, p<.01), mitotic metaphase and anaphase (i.e. increased CDC20 level, p<.001) and separation of sister chromatids (i.e. mutated BUB1B, ESPL1, CENPO). Of note, a 3-gene signature composed of PLK1, CDC20 and RAD50, was able to discriminate 73% of patients between the A- and E-AML cohorts. This signature was confirmed at protein level. In parallel, E-AML showed a preferential dependency on epigenetic mechanisms, with recurrent genomic lesions of ASXL1/2, BCOR/L1, EZH2 and MLL, enrichment of FLT3 alterations and mutations activating RAS signal transduction (p<.05). Of note, a HOX-related signature characterized by overexpression of the HOX family members HOXA7, HOXB3 and MEIS1 identified E-AML.

We show here for the first time the molecular mechanisms promoting and maintaining aneuploidy in AML. Our results indicate that p53 deficiency, either caused by somatic mutations, copy number loss, impaired DNA damage response and enhanced PLK1 signaling synergize with APCgain, RAD50 structural or functional loss and forced progression through mitosis, to override cell cycle and mitotic checkpoints and allow the formation of daughter cells with an aberrant chromosome number. These mechanisms cooperate with recurrent mutations of genes involved in protein ubiquitination and proteasome-mediated protein catabolic process in A-AML, indicative of the attempt of aneuploid cells to override the proteotoxic stress due to the unbalanced protein load generated by the aneuploid condition. This evidence provides the rationale for exploiting proteasome inhibition (Velcade), p53 reactivation (MDM2/4 inhibitor) and targeting of the cell cycle (CHK1/2 inhibitor) downstream to p53 (WEE1 inhibitor) as strategies for novel combination therapies against aggressive aneuploid AML, which are under clinical investigation in our Institution and may serve as a model for aneuploid tumors.

GS and AP: equal contribution.

Supported by: FP7 NGS-PTL project, ELN, AIL, AIRC, PRIN, progetto Regione-Università 2010-12 (L. Bolondi).