Genetic heterogeneity of cytogenetically normal AML with mutations of CEBPA

Abstract

Biallelic mutations of the CCAAT/enhancer binding protein α (CEBPA) gene define a distinct genetic entity of acute myeloid leukemia (AML) with favorable prognosis. The presence of GATA2 and CSF3R mutations that are specifically associated with this subgroup but not mutated in all samples suggests a genetic heterogeneity of biCEBPA-mutated AML. We characterized the mutational landscape of CEBPA-mutated cytogenetically normal AML by targeted amplicon resequencing. We analyzed 48 biallelically mutated CEBPA (biCEBPA), 32 monoallelically mutated CEBPA (moCEBPA), and 287 wild-type CEBPA (wtCEBPA) patient samples from German AML Cooperative Group studies or registry. Targeted sequencing of 42 genes revealed that moCEBPA patients had significantly more additional mutations and additional mutated genes than biCEBPA patients. Within the group of biCEBPA patients, we identified 2 genetic subgroups defined by the presence or absence of mutations in chromatin/DNA modifiers (C), cohesin complex (C), and splicing (S) genes: biCEBPA CCSpos (25/48 [52%]) and biCEBPA CCSneg (23/48 [48%]). Equivalent subgroups were identified in 51 biCEBPA patients from the Cancer Genome Project. Patients in the biCEBPA CCSpos group were significantly older and had poorer overall survival and lower complete remission rates following intensive chemotherapy regimens compared with patients in the biCEBPA CCSneg group. Patients with available remission samples from the biCEBPA CCSpos group cleared the biCEBPA mutations, but most had persisting CCS mutations in complete remission, suggesting the presence of a preleukemic clone. In conclusion, CCS mutations define a distinct biological subgroup of biCEBPA AML that might refine prognostic classification of AML. This trial was registered at www.clinicaltrials.gov as #NCT00266136 and NCT01382147.

Conflict of interest statement

Conflict-of-interest disclosure: The authors declare no competing financial interests.

© 2018 by The American Society of Hematology.

Figures

Graphical abstract

Figure 1.

Mutation spectrum of moCEBPA, biCEBPA, and wtCEBPA. Evaluation of the mutation spectrum of moCEBPA (n = 32) and biCEBPA (n = 48) patients in comparison with wtCEBPA samples (n = 287). Eight of 20 genes with a mutation frequency of ≥5% were significantly associated with ≥1 groups.

Figure 2.

Frequency of genetic alterations organized by categories of related genes and genetic groups. The heatmap includes all genes that were mutated in either the moCEBPA (red) or biCEBPA (dark blue) subgroup. biCEBPA patient samples were further separated in 2 biological groups: CCSneg (blue) and CCSpos (green). Patients with a signal ratio of FLT3-ITD ≥0.5 are marked with an asterisk.

Figure 3.

Age and CR rate of biCEBPACCSpos, biCEBPACCSneg, and moCEBPA patients. (A) Age of biCEBPACCSneg (median = 48), biCEBPACCSpos (median = 66), and moCEBPA (median = 62) patients. Scatter dot plot, median with interquartile range. (B) The CR rate was significantly higher in biCEBPACCSneg patients (91% [20/22]) than biCEBPACCSpos (61% [14/23]; P = .035) and moCEBPA patients (P = .027).

Figure 4.

Survival data depending on biogroup and TET2 status. (A) OS of biCEBPACCSneg, biCEBPACCSpos, and moCEBPA patients. (B) OS of biCEBPA patients depending on TET2 mutational status.

Figure 5.

Mutational status of biCEBPACCSposand biCEBPACCSnegremission samples (age is given in brackets). (A) Four out of 5 biCEBPACCSpos patients had persisting clones in CR, indicating a clonal hematopoiesis. (B) None of the 5 biCEBPACCSneg samples had a detectable mutation load at the time of remission. Dx, diagnosis.