[Detection of heterogeneity and evolution of subclones in t(8;21) AML by QM-FISH].

OBJECTIVE: To explore the heterogeneous subclones in acute myeloid leukemia (AML) with t(8;21) by quantitative multicolor- fluorescence in situ hybridization (QM-FISH), and to figure out whether there is putative ancestral relationship among different subclones. METHODS: Bacterial artificial chromosomes (BAC) clones that contain the targeted genes including AML1, ETO, WT1, p27 and c-kit were searched in the data base UCSC Genome Bioinformatics. Multicolor FISH probes were prepared by linking fluorescein labeled dUTP or dCTP to targeted genes by nick translation. Bone marrow mononuclear cells from t (8;21) AML patients are dropped on to the wet surface of glass slides after hypotonic treatment and fixation. After hybridization, the fluorescence signals were captured by Zeiss fluorescence microscope. The copy number of AML1, ETO, WT1, p27, c- kit and the AML1-ETO fusion gene in AML1-ETO positive cells was counted. The cells with same signals were defined as a subclone. Various subclones were recorded and their proportions were calculated, and their evolutionary relationship was deduced. The subclones in matched primary and relapsed samples were compared, the evolution of dominant clones were figured out and the genomic abnormality that is associated with relapse and drug resistance were speculated. RESULTS: In this study, 36 primary AML with t(8;21) cases and 1 relapsed case paired with the primary case were detected. In these 36 primary cases, 4 cases (11.1%) acquired additional AML1-ETO fusion signal, 3(8.3%) had additional AML1 signal, 4(11.1%) had additional ETO signal, 20(55.6%) had additional WT1 signal, 15(41.7%) had additional p27 signal and 14(38.9%) had additional c-kit signal. In addition, 10(27.8%) displayed AML1 signal deletion, and such an aberration represents statistic significance in male patients. It seems that male patients usually accompany AML1 signal deletion. Of 36 cases, 28(77.8 %) harbored at least 2 subclones (ranged from 2 to 10). According to the genetic signature of subclones, we can assemble a putative ancestral tree, and the genetic architecture is linear or branching. In particular, the clonal architecture of the relapsed sample exhibited significant clonal evolution compared to its paired sample at diagnosis, including proportion changes in dominant clone, subclone disappearance and appearance of new dominant clones. CONCLUSION: Genomic abnormality is very diverse in t(8;21) AML. Subclones have linear or complex branching evolutionary histories, and clonal architecture is dynamic.