The fusion gene Cbfb-MYH11 blocks myeloid differentiation and predisposes mice to acute myelomonocytic leukaemia

The heterodimeric CBF transcription factor genes (CBFA2 (also known as AML1) and CBFB) are the most common translocation targets in human acute myeloid leukaemia (AML), accounting for 30% of AML cases1. Approximately one-half are attributable to a chromosome 16 inversion, inv(16)(p13; q22), found consistently and recurrently in AML subtype M4Eo. This inversion disrupts CBFB and MYH11, a gene encoding the smooth muscle myosin heavy chain, creating the fusion gene CBFB-MYH11 (ref. 2). We previously generated a Cbfb-MYH11 allele in mouse embryonic stem (ES) cells by the knock-in approach3. F1 heterozygousembryos (Cbfb+/Cbfb-MYH11) failed to generate definitive haematopoiesis and died at approximately embryonic day 12.5 (ref. 3). This phenotype is similar to those found in Cbfb–/– and Cbfa2–/– embryos4,5,6, suggesting that Cbfb-MYH11 abrogates CBF function dominantly. To study the role of Cbfb-MYH11 during adult haematopoiesis and its contribution to leukaemogenesis, we analysed chimaeras generated with Cbfb+/Cbfb-MYH11 ES cells.

Bone marrow cells from two Cbfb+/Cbfb-MYH11 chimaeras were sorted based on cell-surface markers7 and then analysed by PCR. We detected Cbfb+/Cbfb-MYH11 cells in the c-kit+/TER119/Lin (Fig. 1a, lane 1) and TER119+ (Fig. 1a, lanes 2 and 3) fractions, but not in the Lin+ (–TER119) fraction (Fig. 1a, lane 4). Consistent with these findings, DNA-PCR (ref. 3), globin polymorphism8 analysis and FACScan revealed the presence of Cbfb+/Cbfb-MYH11 ES cell-derived erythrocytes, but not leukocytes, in the circulating blood (Fig. 1b,c, and data not shown). These results indicate that haematopoietic stem cells containing the knocked-in Cbfb-MYH11 gene are present in the chimaera's bone marrow, which have a selective defect in myeloid and lymphoid differentiation.

Figure 1: Cbfb-MYH11 blocks myeloid and lymphoid differentiation.

a,PCR analysis of Cbfb+/Cbfb-MYH11 ES cell contribution to sorted bone marrow progenitor populations. DNA from c-kit+/TER119/Lin (Lin, CD4/CD8/B220/Mac-1/Gr-1) (lane 1), c-kit+/TER119+/Lin (lane 2), c-kit/TER119+/Lin (lane 3) and Lin+ (lane 4) cells was analysed to detect either the Cbfb-MYH11 allele using neo specific primers (top), or Trp53 as an internal DNA control3 (bottom). b, Globin gel electrophoresis. The contribution from blastocyst-derived cells is indicated by the C57BL/6-specific 's' globin polymorphism. Contribution from ES cells (Cbfb+/Cbfb-MYH11 cells) is indicated by the 129/Sv-specific 'd' polymorphism8. H, K, M, S, T and V, Cbfb+/Cbfb-MYH11 chimaeras. c, Representative FACS analysis of ES cell contribution (Ly9.1 marker is specific for the 129/Sv lineage) to peripheral blood cells in chimaeras generated with either the untargeted ES cells (TC1) or Cbfb+/Cbfb-MYH11 ES cells (KI).

Cbfb+/Cbfb-MYH11 chimaeras did not develop any tumours in their first year of life. It is possible that Cbfb-MYH11 contributes to leukaemic transformation but additional genetic events are required. To test this hypothesis, we injected 4–16-week-old Cbfb+/Cbfb-MYH11 chimaeras with a single sublethal dose of N-ethyl-N-nitrosourea (ENU, 100 mg/kg), a potent DNA alkylating mutagen9. We found that 84% (21/25) of the treated Cbfb+/Cbfb-MYH11 chimaeras developed leukaemia 2–6 months after treatment, whereas none of the ENU-treated TC1 (the wild-type ES cell line) chimaeras developed leukaemia (Fig. 2a).

Figure 2: Development of acute myelomonocytic leukaemia in ENU-treated Cbfb+/Cbfb-MYH11 chimaeras.

a, ENU treatment-induced leukaemia. The solid line represents the survival curve of ENU-treated Cbfb+/Cbfb-MYH11 chimaeras (n=25). The dashed line represents the survival curve of ENU-treated Cbfb+/+ TC1 control chimaeras (n=10). b, Detection of Cbfb+/Cbfb-MYH11 cells in the peripheral blood of an ENU-treated chimaera by PCR. Samples were taken before ENU injection (lanes 1, 2), 1 week (lanes 3, 4), 3 weeks (lanes 5, 6), 6 weeks (lanes 7, 8), 8 weeks (lanes 9, 10), 10 weeks (lanes 11, 12) and 17 weeks (lanes 13, 14) after ENU injection. Odd numbered lanes show PCR using neo-specific primers, and even numbered lanes show PCR using Trp53-specific primers (used as internal control3). Expected amplicon sizes are shown on the side. c, Peripheral blood cells from an ENU-treated Cbfb+/Cbfb-MYH11 chimaera with full-blown leukaemia (chimaera) and a normal healthy 129/Sv×C57BL/6 mouse (WT) were analysed by FACS for the indicated cell surface markers. d, Spleen infiltration. Splenic architecture was partially to extensively effaced by the neoplastic cell population. Splenic lymphoid follicles were still identifiable in most mice (*), as well as foci of erythroid haematopoiesis (arrow). e, Bone marrow infiltration. The bone marrow was extensively effaced, although residual cell populations of normal megakaryocytes (arrows), erythroid and myeloid precursors (arrowheads) were present, with normal maturation evident. f, Meningeal infiltration. Infiltrations of leukaemic cells are evident in the meningeal space (arrow). g, Morphology of leukaemic cells in the peripheral blood. Poorly differentiated and partially differentiated leukaemic cells are indicated by arrows and arrowheads, respectively. h, High-magnification view of one 'hybrid' cell with monocytic nuclear feature and both basophilic and eosinophilic granules. i, Transmission electron microscopic analysis of a leukaemic cell showing the presence of cytoplasmic granules (arrowheads) and multivesicular bodies (arrow), structures frequently found in myelomonocytic cells.

Between two and five months after treatment, Cbfb+/Cbfb-MYH11 cells appeared in the peripheral blood (Fig. 2b). Soon after that, the chimaeras became anaemic (haematocrit, 16.3±5.1%) and leukocytotic (white cell count, 88.0±66.3×103/μl). By differential counting, more than 90% of the circulating white cells were leukaemic blasts. A predominant population of these cells in most cases was c-kit+, Mac-1, Gr-1, B220 and CD3 (Fig. 2c), suggesting the involvement of stem cells not committed to a particular lineage. A population of c-kit+/TER119+ cells was also detected in many cases (Fig. 2c). Its significance is not clear, because the leukaemic blasts in these animals do not resemble erythroblasts.

The clinical course of the disease was acute, because the interval between the first abnormal white cell count and death was usually less than four weeks. On necropsy examination, a consistent feature of the diseased chimaeras was the severely enlarged spleen (5–30-fold larger than normal by weight). Thymus and lymph nodes were generally not involved. Tumour cell infiltrates were present consistently in the spleen (Fig. 2d), bone marrow (Fig. 2e) and liver. We also found infiltrations frequently in lung, kidney and brain, in the form of focal meningeal infiltrates (Fig. 2f). AML M4Eo cells infiltrate the leptomeninges frequently in human patients10.

Microscopically, most leukaemic cells are poorly differentiated (Fig. 2g), whereas partial myelomonocytic differentiation is noted in a small percentage of cells in most mice (Fig. 2g,i). There were occasional cells of 'hybrid' nature, with nuclear characteristics of monocytes and both basophilic and eosinophilic granules in the cytoplasm (Fig. 2h). Such cells are characteristic for the M4Eo subtype of human AML (ref. 11). Overall, the morphologic findings support the diagnosis of acute myelomonocytic leukaemia in these mice.

Isogenic (C57BL/6×129/Sv-F1) recipients intravenously injected with bone marrow cells from the leukaemic chimaeras developed the same type of acute leukaemia as that of the donor 4–12 weeks after transplantation, with or without prior irradiation. By RT-PCR analysis, the leukaemia cells in these recipients expressed the Cbfb-MYH11 fusion gene, as did the original chimaeras (data not shown), suggesting they are derived from Cbfb+/Cbfb-MYH11 ES cells.

No chromosomal rearrangements were detected in leukaemic cells from four affected chimaeras by spectral karyotyping (SKY; ref. 12) and G-banding (data not shown). Therefore, the steps necessary for leukaemia formation did not include gross chromosomal changes or genomic instability. These findings are similar to the pattern of genetic changes in the human leukaemia, that is, inv(16) is frequently the only chromosomal change observed.

The data reported here indicate that Cbfb-MYH11 blocks myeloid differentiation and predisposes mice to leukaemia. The tumours in the Cbfb+/Cbfb-MYH11 chimaeras are almost exclusively acute myelomonocytic leukaemia, even though ENU mutagenizes cells in many tissues and Cbfb expresses broadly13, suggesting a strong disease specificity for the Cbfb-MYH11 oncogene. Furthermore, we showed that alteration of other critical genes is necessary to trigger leukaemogenesis. Because Cbfb-MYH11 seems to contribute to leukaemogenesis by blocking differentiation, the likely role of such cooperating genes is to control cell proliferation or apoptosis.

No mouse leukaemia model expressing CBFA2-CBFA2T1 (commonly known as AML1-ETO), the fusion gene generated by t(8;21) in AML subtype M2 (ref. 14), is available, despite intensive efforts by several groups using both conventional transgenic and Cbfa2-CBFA2T1 knock-in approaches (D. Zhang and J. Downing, pers. comm.). Because Cbfa2-CBFA2T1 also suppresses CBF function dominantly15, it is possible that Cbfa2-CBFA2T1, and potentially homozygous deletions of Cbfb and Cbfa2, contribute to leukaemogenesis but require additional 'hits', as does the Cbfb-MYH11 gene. These possibilities can now be tested in similar mouse chimaera models.


  1. 1

    Look, A.T. Science 278, 1059–1064 (1997).

    CAS  Article  Google Scholar 

  2. 2

    Liu, P. et al. Science 261, 1041–1044 (1993).

    CAS  Article  Google Scholar 

  3. 3

    Castilla, L.H. et al. Cell 87, 687–696 (1996).

    CAS  Article  Google Scholar 

  4. 4

    Wang, Q. et al. Cell 87, 697–708 (1996).

    CAS  Article  Google Scholar 

  5. 5

    Wang, Q. et al. Proc. Natl Acad. Sci. USA 93, 3444–3449 (1996).

    CAS  Article  Google Scholar 

  6. 6

    Okuda, T., van Deursen, J., Hiebert, S.W., Grosveld, G. & Downing, J.R. Cell 84, 321–330 (1996).

    CAS  Article  Google Scholar 

  7. 7

    Orlic, D., Fischer, R., Nishikawa, S., Nienhuis, A.W. & Bodine, D.M. Blood 82, 762–770 (1993).

    CAS  PubMed  PubMed Central  Google Scholar 

  8. 8

    Lyon, M.F., Barker, J.E. & Popp, R.A. J. Hered. 79, 93–95 (1988).

    CAS  Article  Google Scholar 

  9. 9

    Russell, W.L. et al. Proc. Natl Acad. Sci. USA 76, 5818–5819 (1979).

    CAS  Article  Google Scholar 

  10. 10

    Holmes, R. et al. Blood 65, 1071–1078 (1985).

    CAS  PubMed  Google Scholar 

  11. 11

    Bitter, M.A. et al. Am. J. Clin. Pathol. 81, 733–741 (1984).

    CAS  Article  Google Scholar 

  12. 12

    Liyanage, M. et al. Nature Genet. 14, 312–315 (1996).

    CAS  Article  Google Scholar 

  13. 13

    Wang, S. et al. Mol. Cell. Biol. 13, 3324–3339 (1993).

    CAS  Article  Google Scholar 

  14. 14

    Nucifora, G. & Rowley, J.D. Blood 86, 1–14 (1995).

    CAS  PubMed  Google Scholar 

  15. 15

    Yergeau, D.A. et al. Nature Genet. 15, 303–306 (1997).

    CAS  Article  Google Scholar 

Download references


We thank A.R. Moser for advice on the ENU mutagenesis; F. Collins and L. Brody for critical reading of the manuscript; S. Hoogstraten-Miller, A. Cheng and T. Hernendez for technical assistance; and T. Blake, K. Henning and S. Lyons for helpful discussions.

Author information



Corresponding author

Correspondence to P. Paul Liu.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Castilla, L., Garrett, L., Adya, N. et al. The fusion gene Cbfb-MYH11 blocks myeloid differentiation and predisposes mice to acute myelomonocytic leukaemia. Nat Genet 23, 144–146 (1999).

Download citation

Further reading


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing