CBF-AML is defined by the presence of either t(8;21)(q22;q22)/RUNX1-RUNX1T1 or inv(16)(p13.1q22)/t(16;16)(p13.1;q22)/CBFB-MYH11. The resulting fusion genes require a "second hit" to initiate leukemogenesis. Although compared with other AML subtypes patients with CBF-AML have a relatively favorable prognosis, still almost 40% of CBF-AML patients experience relapses of their disease.

Mutation assessment of 84 leukemia- and/or cancer-associated genes using a customized targeted next-generation sequencing (NGS) approach was done on samples from 177 adults with CBF-AML [t(8;21), n=68; inv(16)/t(16;16), n=109]. We identified frequent mutations in the CCND1 and CCND2 genes as novel molecular alterations in AML with t(8;21). They were detected in 10 (15%) patients with t(8;21), making CCND1/2 mutations the third most frequently detected mutations in t(8;21) AML patients. In contrast, we found a single CCND2 mutation in only 1 (0.9%) patient with inv(16), and CCND1/2 mutations in only 11 of 1,426 non-CBF-AML patients (0.77%). Testing for CCND1 and CCND2 mutations using Sanger sequencing in additional 25 CBF-AML patients with t(8;21), for whom NGS analysis was not possible, identified 3 CCND1/2 mutated patients, thereby confirming the CCND1/2 mutation frequency in t(8;21) AML (15% in the NGS set, and 12% in Sanger sequencing set). All CCND2 mutations cluster around the highly conserved amino acid residue threonine 280 (Thr280), and encode a degradation-resistant CCND2 protein. Curiously, de novo germline CCND2 mutations affecting the identical amino acid residues Thr280 or Pro281 have been recently found to cause megalencephaly-polymicrogyria-polydactyly-hydrocephalus syndrome (MPPH) type 3, an extremely rare neurodegenerative disorder (Mirzaa et al., Nat Genet. 2014;46:510-5).

The variant allele fractions of the detected CCND1/2 mutations were variable, suggesting that these mutations can represent both early and later mutational events. Interestingly, three of the 11 CCND1/2 mutated patients had two different CCND1/2 mutations, suggesting a dependence of their leukemia on this mutational event.

The evaluation of mutations in other genes found in CBF-AML patients with CCND1/2 mutations did not reveal any statistically significant co-occurring or mutually exclusive mutations. The most frequent mutation co-occurring with CCND1/2 was ASXL2, detected in 4 of the 10 (40%) cases with t(8;21). Patients with CCND1/2 mutations presented with lower percentages of blood (P=0.02) and bone marrow (P=0.05) blasts than patients without these mutations, but did not differ significantly with regard to other pretreatment characteristics. No differences in the achievement of complete remission, disease-free or overall survival were observed between CCND1/2 mutated and CCND1/2 wild-type patients. Mechanistically, we show that Thr280Ala mutated CCND2 significantly increased cell growth (as assessed by TiterGlo assays and cell counts) in all experimental set-ups. Western blotting revealed an increase of retinoblastoma gene phosphorylation, indicating increased downstream signaling. Moreover, both propidium iodide mediated cell cycle analysis and immunofluorescence staining for phospho-histone H3 indicated an increased mitotic rate and a decreased G1 fraction.

In summary, recurrent mutations in CCND1/2 are consistent with the currently known mutational landscape of t(8;21) AML, which is dominated by receptor tyrosine kinase, RAS, and chromatin remodeling gene mutations─all of which can result in targeting of CCND1/2. While CCND1 and CCND2 play a crucial role in the regulation of hematopoietic differentiation and cell proliferation, and their differential activation has been implicated in leukemogenesis, frequent mutations in these genes have not previously been reported in AML. The fact that no outcome difference was seen when comparing CCND1/2-mutated versus wild-type patients may be due to the relatively small sample size, or may possibly reflect the common use of cell cycle affecting high-dose cytarabine-based therapy in CBF-AML patients. Thus, the identification of CCND1/2 mutations as frequent mutational events in t(8;21) AML may provide further justification for cell cycle-directed therapy in this disease.


Stone:Merck: Consultancy; Karyopharm: Consultancy; Jansen: Consultancy; Pfizer: Consultancy; Juno Therapeutics: Consultancy; ONO: Consultancy; Roche: Consultancy; Seattle Genetics: Consultancy; Sunesis Pharmaceuticals: Consultancy; Xenetic Biosciences: Consultancy; Celator: Consultancy; Agios: Consultancy; Amgen: Consultancy; Novartis: Consultancy; Celgene: Consultancy, Membership on an entity's Board of Directors or advisory committees; Abbvie: Consultancy, Membership on an entity's Board of Directors or advisory committees.

Author notes


Asterisk with author names denotes non-ASH members.