Aneuploid acute myeloid leukemia exhibits a signature of genomic alterations in the cell cycle and protein degradation machinery

Abstract

Background: Aneuploidy occurs in more than 20% of acute myeloid leukemia (AML) cases and correlates with an adverse prognosis.

Methods: To understand the molecular bases of aneuploid acute myeloid leukemia (A-AML), this study examined the genomic profile in 42 A-AML cases and 35 euploid acute myeloid leukemia (E-AML) cases.

Results: A-AML was characterized by increased genomic complexity based on exonic variants (an average of 26 somatic mutations per sample vs 15 for E-AML). The integration of exome, copy number, and gene expression data revealed alterations in genes involved in DNA repair (eg, SLX4IP, RINT1, HINT1, and ATR) and the cell cycle (eg, MCM2, MCM4, MCM5, MCM7, MCM8, MCM10, UBE2C, USP37, CK2, CK3, CK4, BUB1B, NUSAP1, and E2F) in A-AML, which was associated with a 3-gene signature defined by PLK1 and CDC20 upregulation and RAD50 downregulation and with structural or functional silencing of the p53 transcriptional program. Moreover, A-AML was enriched for alterations in the protein ubiquitination and degradation pathway (eg, increased levels of UHRF1 and UBE2C and decreased UBA3 expression), response to reactive oxygen species, energy metabolism, and biosynthetic processes, which may help in facing the unbalanced protein load. E-AML was associated with BCOR/BCORL1 mutations and HOX gene overexpression.

Conclusions: These findings indicate that aneuploidy-related and leukemia-specific alterations cooperate to tolerate an abnormal chromosome number in AML, and they point to the mitotic and protein degradation machineries as potential therapeutic targets.

Keywords: acute myeloid leukemia; aneuploidy; cell cycle; genomics; mutation; ubiquitination; whole exome sequencing.

Figure 1

Genomic lesions in A‐AML and E‐AML. (A) Number and type of nonsilent somatic mutations detected by whole exome sequencing. (B) Frequency of A‐AML and E‐AML cases classified according to the number of mutations. (C) Pattern of genomic lesions in A‐AML and E‐AML. Rows denote genes or group of genes (other). Columns represent (from left to right) functional categories (distinguished by colors), mutated genes/groups of genes/other genomic alterations, and single patients. A‐AML indicates aneuploid acute myeloid leukemia; CN, copy number; E‐AML, euploid acute myeloid leukemia; indel, insertion/deletion; LOH, loss of heterozygosity; TF, transcription factor.

Figure 2

Spectrum of somatic mutation categories distinguishing A‐AML and E‐AML. (A) Frequency of cases carrying mutations according to functional categories. Statistical significance was determined with the Fisher exact test (*P < .05; **P < .01). (B) Distribution of mutations targeting cell cycle–related genes. Each row denotes 1 gene; columns represent (from left to right) cell cycle phases, mutated genes, and single patients. (C) Frequency of mutations according to cell cycle phases. A‐AML indicates aneuploid acute myeloid leukemia; AML, acute myeloid leukemia; CN, copy number; E‐AML, euploid acute myeloid leukemia; LOH, loss of heterozygosity.

Figure 3

Mutational signatures in A‐AML and E‐AML. (A) Mutational signatures according to the 96‐substitution classification. Mutation types are reported on the horizontal axes with different colors; the percentage of each specific mutation type is represented by vertical axes. (B) Contributions of the identified signatures to the mutational processes. A‐AML indicates aneuploid acute myeloid leukemia; E‐AML, euploid acute myeloid leukemia; S1, Signature #1; S2, Signature #2.

Figure 4

Frequency and co‐occurrence of CNAs in leukemia‐related genes in (A) euploid acute myeloid leukemia and (B) aneuploid acute myeloid leukemia. The Circos plots depict copy number gains/duplications (in red) and loss/deletions (in green) in acute myeloid leukemia–related genes associated with the aneuploid phenotype. The bar plots represent the percentages of patients with copy number events in each gene (0%‐100% scale). Links connect copy number alterations co‐occurring in the same patient; the color intensity reflects the absolute frequency of patients harboring that co‐occurrence (range, 1‐17). Mutually exclusive alterations may exist in areas that are not connected.

Figure 5

Gene expression profile analysis of A‐AML and E‐AML. (A) Gene expression differences in leukemia‐related and cell cycle– and DNA repair–related genes between A‐AML (n = 22) and E‐AML (n = 27). Data are standardized through a z‐score transform; color changes within a row indicate expression levels relative to the mean and rescaled on the transcript standard deviation. (B) Biological processes significantly enriched among differentially expressed genes in A‐AML versus E‐AML (P < .05). (C) Percentage of Ki‐67⁺ cells on bone marrow blasts of patients with A‐AML and E‐AML according to flow cytometry analysis. Statistical significance was determined with the Student t test (*P < .05). (D,E) Expression of UHRF1, UBA3, UBE2C, RAD50, PLK1, and CDC20 proteins. (D) Western blot of representative cases. (E) Densitometry after normalization for the mean value across E‐AML. Statistical significance was determined with the Student t test (*P < .05; **P < .01). (F) Signature of p53‐downregulation in A‐AML identified by gene set enrichment analysis. A and A‐AML indicate aneuploid acute myeloid leukemia; E and E‐AML, euploid acute myeloid leukemia; ES, enrichment score; NES, normalized enrichment score.

Figure 6

Mechanisms that potentially induce and support aneuploidy in AML: a model incorporating the genomic and transcriptomic results. AML indicates acute myeloid leukemia; SAC, spindle assembly checkpoint.