Enforced expression of the homeobox gene Mixl1 impairs hematopoietic differentiation and results in acute myeloid leukemia

Abstract

Mixl1, the sole murine homologue of the Xenopus Mix/Bix family of homeobox transcription factors, is essential for the patterning of axial mesendodermal structures during early embryogenesis. Gene targeting and overexpression studies have implicated Mixl1 as a regulator of hematopoiesis arising in differentiating embryonic stem cells. To assess the role of Mixl1 in the regulation of adult hematopoiesis, we overexpressed Mixl1 in murine bone marrow using a retroviral transduction/transplantation model. Enforced expression of Mixl1 profoundly perturbed hematopoietic lineage commitment and differentiation, giving rise to abnormal myeloid progenitors and impairing erythroid and lymphoid differentiation. Moreover, all mice reconstituted with Mixl1-transduced bone marrow developed fatal, transplantable acute myeloid leukemia with a mean latency period of 200 days. These observations establish a link between enforced Mixl1 expression and leukemogenesis in the mouse.

Genes that regulate blood formation in the embryo can also play key roles in the regulation of adult blood stem cells and the specification of hematopoietic cell lineages and are frequently implicated in the genesis of hematopoietic malignancies such as leukemia and lymphoma. One such family of genes is the homeobox gene family of DNA-binding proteins. Class I homeobox (Hox) genes consist of 13 paralogous groups organized into four clusters that direct anteroposterior patterning during embryogenesis. Class II divergent homeobox genes are dispersed throughout the genome, generally have restricted patterns of gene expression, and function to direct embryonic development, organogenesis, or cellular differentiation. Gene knockout studies have shown that members of both classes of homeobox proteins play important roles in hematopoiesis, influencing hematopoietic stem cell renewal, lineage commitment, and differentiation. Enforced overexpression studies in mice, together with gene expression profiling of human leukemias, have implicated homeobox proteins in leukemic transformation (1).

Mixl1, a class II homeobox gene of the Paired-like subclass, is the sole murine orthologue of the Xenopus Mix/Bix gene family (2, 3). Xenopus and zebrafish Mix/Bix genes are induced by TGF-β family ligands in the gastrulation-stage embryo and have been implicated in mesendodermal development (4–13). Mix.1, the founding member of the Xenopus Mix/Bix gene family, was first identified as an immediate-early response gene to activin stimulation and was rediscovered during a screen for factors that ventralized the Xenopus embryo (4, 14). Overexpression of Mix.1 altered cell fate in the developing Xenopus embryo, resulting in increased blood formation and suggesting that it may be a regulator of hematopoiesis (14).

During murine embryogenesis, Mixl1 expression is detectable in mesodermal progenitors that give rise to the first blood cells and in fetal liver colony-forming cells (CFCs) and can be induced by TGF-β family ligands (3, 15–17). Mixl1-null embryos die at embryonic day 8 with complex defects in axial mesendodermal structures, precluding detailed analysis of the role of Mixl1 in hematopoiesis (18). The in vitro differentiation of murine embryonic stem cells as embryoid bodies recapitulates many aspects of hematopoietic development (19). Early in embryoid body differentiation, Mixl1 is expressed in the Brachyury⁺/Flk⁺ cell population that gives rise to hemangioblasts. Moreover, hemangioblasts are enriched in the Mixl1-expressing population, and Mixl1-null embryonic stem cells display a defect in differentiation into hematopoietic cells characterized by reduced and delayed Flk1 expression and decreased formation of CFCs (15, 16). Recently, the precocious induction of Mixl1 in differentiating embryoid bodies was found to increase the numbers of mesodermal, hemangioblast, and hematopoietic progenitors (20).

Human MIXL was shown to activate Xenopus α-T4 globin in Xenopus animal cap assays when basic fibroblast growth factor was used to induce mesoderm, indicating that MIXL1 could drive mesoderm to a hematopoietic cell fate (21). MIXL1 expression has been reported in hematopoietic tissues, and MIXL1 protein was detected in leukemic cell lines and biopsy samples from patients with high-grade lymphoma (21, ¶).

To determine the consequences of enforced Mixl1 expression on adult hematopoiesis, lethally irradiated mice were reconstituted with bone marrow cells transduced with a Mixl1-expressing retrovirus. We found that enforced Mixl1 expression perturbed hematopoietic commitment and differentiation, resulting in an accumulation of cells of the myeloid lineage and blocking the production of erythroid and lymphoid cells. Moreover, after a latency period, the mice developed acute myeloid leukemia (AML). The results demonstrate that enforced Mixl1 expression dysregulates normal hematopoiesis and implicate Mixl1 as a putative oncogene.

Results

Mixl1 Overexpression Alters Hematopoietic Differentiation, Expanding Primitive Cells and Myeloid Progenitors and Impairing Erythropoiesis and Lymphopoiesis.

Mixl1 cDNA was cloned into a murine stem cell virus (MSCV)-based retrovirus carrying an enhanced green fluorescent protein (GFP) reporter (22) (Fig. 1A and B). To assess the effect of Mixl1 overexpression on hematopoiesis in vivo, 5-fluorouracil (5-FU)-treated bone marrow was infected with MSCV (control) or Mixl1-expressing retrovirus and transplanted into lethally irradiated syngeneic hosts. Primary recipients were chimeras, because they received a mixture of transduced and nontransduced cells. The percentage of peripheral blood leukocytes expressing retrovirus was determined by flow cytometry at 8, 12, and 16 weeks after reconstitution. In mice receiving control-transfected bone marrow, the percentage of GFP-positive cells remained constant, whereas in mice receiving Mixl1-transduced bone marrow the percentage of GFP-positive, Mixl1-expressing cells increased with time (Fig. 1C).

Fig. 1.

Overexpression of Mixl1 in murine bone marrow. (A) MSCV–IRES–GFP control vector. (B) MSCV–Mixl1–IRES–GFP (Mixl1) vector. AD, activation domain; FLAG, N-terminal FLAG epitope; HD, homeodomain; LTR, long terminal repeat; IRES, internal ribosomal reentry site. (C) Lethally irradiated mice were reconstituted with 5-fluorouracil-treated bone marrow transfected with control or Mixl1 retrovirus. The percentage of peripheral blood leukocytes expressing GFP in individual recipient mice was monitored by flow cytometry. In contrast to control mice, in which the percentage of retrovirally transduced cells remained constant, the percentage of Mixl1-transduced cells increased over time. (Mixl1 n = 12, MSCV n = 3.)

Healthy recipient mice between 14 and 20 weeks of age were killed for analysis. GFP-positive cells accounted for 25–80% of bone marrow cells in both the experimental (Mixl1) and control (vector only) groups of animals. Hematocrits and total white blood cell counts were similar in both groups, but platelet counts were reduced in mice receiving Mixl1-transduced bone marrow (control, 824 ± 256, n = 7; Mixl1, 421 ± 272, n = 16; P < 0.05). Gross and histological examinations were normal, with the exception of the spleens of mice transplanted with Mixl1-transduced bone marrow, which were increased in size and contained elevated numbers of granulocytic cells in the red pulp. Peripheral blood, bone marrow, and spleen from recipient mice were analyzed by flow cytometry using a panel of lineage-specific fluorescent antibodies and GFP as a marker for transduction. In control mice, the lineage compositions of the GFP-positive and -negative cellular fractions were similar. In contrast, although the GFP-negative fractions from peripheral blood, bone marrow, and spleen of Mixl1-transduced mice contained granulocytic, erythroid, and B and T lymphoid cells, the GFP-positive fractions exhibited a marked expansion of Mac-1/⁺Gr-1⁻ cells and decreased numbers of erythroid and lymphoid cells. Twenty to 50% of Mixl1-expressing bone marrow and spleen cells expressed c-kit (Fig. 2). In keeping with the immunophenotypic profile, differential counts of cytospins of GFP-positive cells from mice receiving Mixl1-transduced bone marrow showed an increased proportion of blast cells and immature myeloid cells.

Fig. 2.

Enforced Mixl1 expression in bone marrow skews hematopoietic differentiation. (A) Mice reconstituted with control or Mixl1-transfected bone marrow were killed 14–20 weeks after reconstitution, and peripheral blood, bone marrow, and spleen were analyzed by flow cytometry, gating on GFP-positive and -negative cells. GFP-gated FACS contour plots are shown. The percentage of Mac-1⁺/Gr-1⁻ and c-kit⁺ cells was elevated in the Mixl1-expressing (GFP-positive) fraction. The results are representative of 15 pairs of mice. (B) Percent of control (n = 3) or Mixl1-transduced (n = 12) GFP-positive bone marrow or spleen cells expressing B cell (B220⁺), erythroid (Ter119⁺), myeloid (Mac-1⁺/Gr-1⁺), and granulocytic (Mac-1⁺/Gr-1⁻) T cell (CD4⁺ or CD8⁺) markers.

To examine the effect of enforced Mixl1 expression on defined bone marrow progenitors, we isolated lineage-negative (lin⁻) bone marrow cells of Mixl1-transduced or control animals and used immunophenotypic criteria to determine the percentage of early myeloid progenitors in the GFP-positive and -negative fractions (23, 24). In Mixl1-expressing bone marrow, the percentages of hematopoietic stem cell-enriched (lin⁻/Sca-1⁺/c-kit⁺) and myeloid progenitor-enriched (lin⁻/IL-7R-α⁻/Sca-1⁻/c-kit⁺) cells were increased (Fig. 3A). In control-transduced bone marrow the three typical groups of myeloid progenitors: common myeloid progenitors (CMPs), lin⁻/IL-7R-α⁻/Sca-1⁻/c-kit⁺/CD34⁺/FcR-γ^int; granulocyte/macrophage progenitors (GMPs), lin⁻/IL-7R-α⁻/Sca-1⁻/c-kit⁺/CD34⁺ /FcR-γ^hi; and megakaryocyte erythroid progenitors (MEPs), lin⁻/IL-7R-α−/Sca-1⁻/c-kit⁺/CD34⁻/FcR-γ^low could be distinguished. In contrast, the Mixl1-transduced lin⁻/Sca-1⁻/c-kit⁺ population was composed of atypical myeloid progenitor populations with an increased percentage of cells with the phenotypic features of GMPs, reduced CMPs, and an abnormal MEP-like population (Fig. 3B).

Fig. 3.

Increased numbers of abnormal early myeloid progenitors in Mixl1-transduced bone marrow. (A) Percent of lin⁻/Sca-1⁺/c-kit⁺ and myeloid progenitor-enriched (lin⁻/Sca-1⁻/c-kit⁺) GFP-positive fractions of Mixl1- (n = 7) and control-transduced (n = 7) bone marrow. ∗, P < 0.01; ∗∗, P < 0.001. (B) Early myeloid progenitor populations in GFP-positive fractions of bone marrow from mice reconstituted with control (MSCV) or Mixl1-transduced bone marrow. In control transduced bone marrow, typical common myeloid progenitor (CMP) (lin⁻/IL-7R-α⁻/Sca-1⁻/c-kit⁺/CD34⁺/FcR-γ^low), GMP (lin⁻/IL-7R-α⁻/Sca-1⁻/c-kit⁺/CD34⁺/FcR-γ^hi), and megakaryocyte erythroid progenitor (MEP) (lin⁻/IL-7R-α⁻/Sca-1⁻/c-kit⁺/CD34⁻/FcR-γ^low) populations were present. Mixl1-transduced bone marrow contained an increased percentage of GMPs. Results are representative of three experiments.

CFCs in the GFP-positive and -negative fractions of bone marrow from mice reconstituted with control or Mixl1-transduced bone marrow were assayed by culture in semisolid agar with a range of cytokines. Whereas GFP-positive and -negative fractions from bone marrow transduced with the control virus contained similar numbers and types of CFCs, blast and granulocytic colonies were increased up to 10-fold in the GFP-positive fraction from Mixl1-transduced bone marrow stimulated by granulocyte/macrophage colony-stimulating factor (GM-CSF), IL-3, or stem cell factor plus IL-3 and erythropoietin. Strikingly, when stimulated by GM-CSF, Mixl1-transduced cells formed blast colonies (not normally observed with GM-CSF stimulation) and increased numbers of granulocytic colonies. The granulocytic colonies were of abnormal size and shape, but none exhibited autonomous proliferation. Numbers of granulocyte/macrophage, macrophage, and megakaryocytic colonies tended to be reduced when compared with the GFP-negative fraction (Table 1). The frequency of the colony-forming unit-spleen (CFU-S), an early pluripotent progenitor, in GFP-positive and -negative fractions was assayed. There were no significant differences in spleen weights or day 12 CFU-S numbers between mice receiving GFP-positive and -negative fractions from control or Mixl1-transduced bone marrow (data not shown). Taken together, these results indicate that Mixl1 overexpression in bone marrow stem cells results in intrinsic abnormalities of hematopoietic differentiation. Hematopoietic differentiation is skewed with an expansion of abnormal myeloid progenitor cells and reduced numbers of erythroid and lymphoid cells.

Table 1.

Colony formation by Mixl1-transduced and control bone marrow cells

Bone marrow fraction	Stimulus	No. of colonies
		B	G	GM	M	EO	MEG
Mixl1-transduced GFP-positive (n = 10)	GM-CSF	3 ± 4	13 ± 9	1 ± 2	2 ± 2	0 ± 0
	G-CSF		2 ± 1	0 ± 0	0 ± 0
	M-CSF		0 ± 0	0 ± 0	0.8 ± 1.6
	IL-3	5 ± 4	7 ± 6	1 ± 1	3 ± 3	0 ± 0	0.1 ± 0.3
	SCF + IL-3 + Epo	12 ± 7	11 ± 8	1 ± 1	7 ± 8	0.2 ± 0.4	0.4 ± 0.7
	Saline		0 ± 0	0 ± 0	0 ± 0
Mixl1-transduced GFP-negative (n = 10)	GM-CSF	0.1 ± 0.3	2 ± 1	0.8 ± 1	4 ± 4	0 ± 0
	G-CSF		0.6 ± 0.5	0 ± 0	0 ± 0
	M-CSF		0 ± 0	0.3 ± 0.7	6 ± 5
	IL-3	0.5 ± 0.5	1 ± 1	0.2 ± 0.4	2 ± 1	0 ± 0	0.5 ± 0.7
	SCF + IL-3 + Epo	0.6 ± 0.5	2 ± 1	0.6 ± 0.7	2 ± 1	0 ± 0	0.9 ± 0.5
	Saline		0 ± 0	0 ± 0	0 ± 0
MSCV-transduced GFP-positive (n = 4)	GM-CSF	1 ± 2	5 ± 4	0.8 ± 1.0	4 ± 3	0.3 ± 0.5
	G-CSF		1 ± 0	0 ± 0	0 ± 0
	M-CSF		0 ± 0	0 ± 0	9 ± 6
	IL-3	1 ± 1	3 ± 1	0.5 ± 1.0	2.5 ± 1.7	0 ± 0	0.3 ± 0.5
	SCF + IL-3 + Epo	1.3 ± 0.6	3.7 ± 1.0	0.8 ± 1.0	2.3 ± 1.5	0.3 ± 0.5	1.0 ± 1.4
	Saline		0 ± 0	0 ± 0	0 ± 0
MSCV-transduced GFP-negative (n = 4)	GM-CSF	0 ± 0	3 ± 1	0.8 ± 0.5	3 ± 2	0 ± 0
	G-CSF		1 ± 1	0 ± 0	0 ± 0
	M-CSF		0 ± 0	0 ± 0	6 ± 2
	IL-3	0.8 ± 0.5	2.5 ± 1.3	0.8 ± 0.6	2.5 ± 1.3	0 ± 0	0.5 ± 0.6
	SCF + IL-3 + Epo	1.5 ± 1.0	3.3 ± 2.1	1.0 ± 0.8	2.5 ± 1.3	0.3 ± 0.5	3.0 ± 1.6
	Saline		0 ± 0	0 ± 0	0 ± 0

Lethally irradiated mice were reconstituted with bone marrow transduced with MSCV–Mixl1–GFP or control (MSCV) retrovirus. The mice were sacrificed 14–20 weeks later, and bone marrow was sorted into GFP-positive and -negative fractions. Cells (n = 2,000) were plated in semisolid agar cultures with cytokines as listed, and colonies were enumerated after 7 days. Colonies were fixed, stained, and examined microscopically to identify them as blast (B), granulocyte (G), granulocyte macrophage (GM), macrophage (M), eosinophil (EO), or megakaryocyte (MEG). Numbers in bold indicate P < 0.001. GM-CSF, granulocyte/macrophage colony-stimulating factor; G-CSF, granulocyte colony-stimulating factor; M-CSF, macrophage colony-stimulating factor; SCF, stem cell factor; Epo, erythropoietin. Numbers in bold: P < 0.001. Blank spaces indicate assays in which colonies are not stimulated by the cytokine added.

Enforced Mixl1 Expression Results in AML.

Cohorts of mice receiving control or Mixl1-expressing bone marrow were monitored for the development of disease. In contrast with control mice, which remained well, all recipients of Mixl1-transduced bone marrow developed a fatal AML with a median onset of 200 days (range of 95–325 days) (Fig. 4). Disease onset varied with the percentage of GFP-positive cells in the peripheral blood at 12 weeks. Full blood examination revealed anemia, thrombocytopenia, and circulating blast cells. The mice exhibited marked splenomegaly and variable hepatomegaly (Table 2). Histological analysis revealed massive invasion of the bone marrow, spleen, and liver by leukemic blast cells and a variable component of more differentiated granulocytic cells. Other organs were infiltrated less frequently (Fig. 5). Flow cytometric analysis showed that the majority of GFP-positive cells from leukemic bone marrow or spleen were negative for lymphoid, erythroid, and megakaryocytic antigens with a variable proportion of cells expressing c-kit (range of 10–33%) or Mac-1 (range of 5–30%), indicating myeloid differentiation (Fig. 6). Differential counts indicated that the majority of GFP-positive cells were blast cells with some immature myeloid cells. These features fit the diagnostic criteria for AML with maturation (25). Clonal cultures of GFP-positive cells from bone marrow or spleen of leukemic mice gave rise mainly to abnormal, small colonies composed of loosely dispersed cells with a blast-like morphology. The leukemic CFCs did not exhibit an extended capacity for self-generation and most were factor-dependent, although some factor-independent CFCs arose in unstimulated cultures. Cultures of GFP-negative cells from leukemic mice exhibited normal colony formation (Table 3). In all cases tested (n = 12), the leukemia could be transplanted to nonirradiated syngeneic mice. Recipient mice receiving s.c. or i.v. cells developed leukemia with a mean latency of 50 days. The cells of the transplanted leukemias were GFP-positive, and their immunophenotype was similar to that of the primary leukemias except that fewer c-kit⁺ or Mac-1⁺ cells were present. Histological examination of organs from mice receiving transplants of leukemic cells showed pathology similar to that observed in the primary mice. Southern analysis demonstrated that the leukemias were clonal or oligoclonal (data not shown). This, together with the long latency period before development of disease, indicated that the initial genetic change engineered in the hematopoietic stem cells was insufficient to induce disease.

Fig. 4.

Enforced expression of Mixl1 results in AML. Kaplan–Meier plot showing the survival of cohorts of mice reconstituted with bone marrow transduced with control (n = 14) or Mixl1 (n = 16) retrovirus.

Table 2.

Phenotype of mice developing leukemia after reconstitution with Mixl1-transduced bone marrow

Mouse ID	Weeks to leukemia	% GFP in bone marrow	% GFP in spleen	Hb, gm/liter	WBCs, gm/ml	Plaletes, 106 per ml	Spleen, mg	Transplantable*
302	36	84	94	2.0	6.3	11	305	Yes
1204	27	73	83	3.4	23.4	16	650	Yes
1201	29	59	56	8.4	10.6	36	1200	Yes
104	42	69	79	3.5	5.6	20	684	Yes
801	33	73	79	8.7	31.2	21	811	Yes
409	43	74	44	10.2	9.0	16	437	Yes
1303	30	89	49	10.8	5.4	31	531	Yes
1005	41	Necrotic	81	7.7	20.0	13	750	Yes

*Bone marrow or spleen cells from primary leukemic mice were injected into nonirradiated syngeneic mice and observed for the development of leukemia.

Fig. 5.

Histology of leukemia arising in mice receiving Mixl1-transduced bone marrow. (A) Peripheral blood film from leukemic mouse showing leukocytosis and increased immature blast cells. (B) Bone marrow section showing a hypercellular appearance due to the replacement of normal marrow cells with leukemic blasts. (C) Liver showing widespread perivascular infiltration (arrows) by leukemic cells. (D) Extensive infiltration of leukemic cells (arrow) in pancreas. is, Islet. (E) Effacement of normal spleen architecture by leukemic cells. Only small clusters of nucleated erythroid cells remain. (F) High-power image of spleen in E showing extensive infiltration by leukemic cells. Occasional macrophages are seen, but few lymphoid cells remain.

Fig. 6.

Leukemia arising in mice receiving Mixl1-transduced bone marrow. Representative immunophenotype of Mixl1-induced leukemia. Leukemic cell (GFP-positive) population was negative for erythroid (Ter119) and lymphoid (B220, CD19, CD4, and CD8) markers and contained c-kit⁺ and Mac-1⁺ subpopulations.

Table 3.

Colony formation by GFP-positive and -negative cells from leukemic Mixl1-transduced mice

Leukemic marrow cells (n = 4)	Stimulus	No. of colonies formed
		B	G	GM	M	EO	MEG	Comments
GFP-positive	GM-CSF	10 ± 10	3 ± 3	0.5 ± 1.0	0 ± 0	0 ± 0		Atypical
	G-CSF	4 ± 3	2 ± 2	0 ± 0	0 ± 0
	IL-3	17 ± 6	4 ± 4	0.5 ± 0.6	0 ± 0	0 ± 0	0 ± 0	Atypical
	M-CSF	2 ± 2	0 ± 0	0 ± 0	0 ± 0
	SCF	10 ± 13	0 ± 0	0 ± 0	0 ± 0
	SCF + IL-3 + Epo	51 ± 33	5 ± 5	4 ± 7	0 ± 0	0 ± 0	0 ± 0	Atypical
	Saline	4 ± 3	0 ± 0	0 ± 0	0 ± 0
	GFP-negative	GM-CSF	0 ± 0	2 ± 2	0.5 ± 0.6	3 ± 1	0.3 ± 0.5		Normal
	G-CSF		2 ± 1	0 ± 0	0 ± 0
	IL-3	0.5 ± 1.0	3 ± 3	1 ± 2	1 ± 1	0.3 ± 0.5	0 ± 0	Normal
	M-CSF		0 ± 0	0 ± 0	9 ± 11			Normal
	SCF	0.3 ± 0.5	1 ± 2	0 ± 0	0 ± 0			Normal
	SCF + IL-3 + Epo	1 ± 2	3 ± 3	2 ± 2	1 ± 1	0 ± 0	1 ± 1	Normal
	Saline		0 ± 0	0 ± 0	0 ± 0

Lethally irradiated mice were reconstituted with bone marrow transduced with MSCV–Mixl1–GFP retrovirus. Leukemic mice were killed, and bone marrow was sorted into GFP-positive and -negative fractions. Cells (n = 2,000) were plated in semisolid agar cultures with cytokines as indicated. B, blast; G, granulocyte; GM, granulocyte macrophage; M, macrophage; EO, eosinophil; MEG, megakaryocyte; GM-CSF, granulocyte/macrophage colony-stimulating factor; G-CSF, granulocyte colony-stimulating factor; M-CSF, macrophage colony-stimulating factor; SCF, stem cell factor; Epo, erythropoietin. Blank spaces indicate assays in which colonies are not stimulated by the cytokine added.

Discussion

Enforced expression of Mixl1 in hematopoietic stem cells markedly perturbed hematopoiesis, expanding immature and mature myeloid cells and inhibiting erythroid, megakaryocytic, and lymphoid differentiation. In lineage-depleted Mixl1-transduced bone marrow, Sca-1⁺/c-kit⁺ cells and Sca-1⁻/c-kit⁺ cells were increased, and the normal sequence of myeloid maturation, as represented by differences in cell surface markers that correlate with cell fate commitment, was perturbed. The effect of Mixl1 expression on hematopoiesis and the development of leukemia prevented meaningful analysis of competitive repopulation assays. However, colony-forming unit-spleen (CFU-S) numbers in Mixl1-transduced bone marrow were not altered, suggesting that the increase in lin⁻/Sca-1⁺/c-kit⁺ and lin⁻/Sca-1⁻/c-kit⁺ cells most likely reflects a difference in the numbers of multipotent progenitors, rather than repopulating cells. The findings contrast with the results of experiments in which precocious Mixl1 expression in embryoid bodies was shown to accelerate and expand hematopoiesis without bias or skewing of hematopoietic colony types as judged by marker gene expression, flow cytometry, and CFC assays (20). This indicates that the transcriptional effects of Mixl1 are strongly modified by their cellular context.

After a latency period of several months, mice reconstituted with Mixl1-transduced bone marrow cells developed a transplantable AML characterized by anemia, thrombocytopenia, splenomegaly, and widespread infiltration of organs by blast cells. The time of disease onset depended on the size of the progenitor population at risk (percent of retrovirally transduced cells), and the delayed onset of leukemia indicated that additional genetic events are required for malignant transformation. Aberrant expression of HOX genes is associated with leukemic transformation in both mice and humans. Retroviral transduction/transplantation studies have demonstrated roles for Hoxa9, HOXA10, HOXB3, HOXB6, and HoxB8 genes in leukemic transformation in mice (26–30), and gene expression profiling of acute leukemias has linked aberrant expression of HOX genes, in particular HOXA9, to myeloid leukemogenesis (31, 32). More recently, HOX genes have also been implicated in lymphoid malignancy (33). In contrast, nonclustered homeobox genes have been associated only rarely with leukemogenesis (1). Interestingly, ectopic expression of the protooncogene Cdx2, a nonclustered homeobox gene, was recently shown to cause myeloid leukemia in mice (34). This may be due to the role of Cdx2 as an upstream regulator of Hox genes (35). A critical question asks what distinct Mixl1 functions direct the predisposition to leukemia development. The in vivo hematopoietic effects induced by Mixl1 are reminiscent of the effects of the retroviral overexpression of HOX genes, as well as leukemia fusion genes such as NUP98-HOXD13, and it is possible that enforced expression of Mixl1 may effect an altered balance of expression and function of Hox genes.

MIXL1 protein has been detected in a number of human leukemic cell lines and is reported to be present in high-grade lymphoma (21, ¶). The absence of lymphoma phenotype in this work is probably a result of the enforcement of myeloid commitment in early stem cells. To date, MIXL1 expression has not been linked to human leukemia. The findings reported here make it important to ascertain whether transcriptional deregulation of MIXL1 is involved in the pathogenesis of human hematopoietic malignancy.

Materials and Methods

Generation of Retrovirus and Southern Analysis.

An MSCV-based retroviral construct containing the internal ribosome reentry site from the encephalomyocarditis virus and GFP cDNA (22) was modified by the insertion of sequence encoding a FLAG epitope and an Mlu1 restriction site into the multiple cloning site. Mixl1 cDNA was cloned into the Mlu1 site. Retroviral supernatants were prepared from Phoenix E cells as previously described (36). Retroviral supernatants were collected 48 and 72 h after transfection, filtered, and stored at −80°C until use. DNA extracted from the spleen of repopulated animals was digested with EcoRI and analyzed by Southern blot hybridization using a GFP probe to estimate the number of integrated provirus.

Bone Marrow Transduction and Transplantation.

Transduction of bone marrow cells with retroviral supernatant and transfer of these cells into lethally irradiated recipients were performed as described (36). In brief, bone marrow cells were harvested from C57BL/6.SJL [Ptprc^aPep3^b (Ly5.1)] donor mice 4 days after i.v. injection of 150 mg/kg 5-FU and cultured for 24 h in Dulbecco's modified Eagle's medium supplemented with 10% FCS, murine stem cell factor (100 ng/ml), IL-3 (6 ng/ml), IL-6 (10 ng/ml), IL-1-β (4 ng/ml), and IFN-γ (1 ng/ml). The following day, 1–2 × 10⁶ cells per milliliter in the same media were transferred to 12-well plates coated with Retronectin (Takara Bio, Tokyo, Japan), retroviral supernatant (50% vol/vol) and polybrene (6 μg/ml) were added, and the cells were transduced by spinfection (37). After a further round of spinfection, 2.5 × 10⁵ to 1.0 × 10⁶ cells were transplanted via the tail vein into lethally irradiated (2 × 550 Gy) C57BL/6 [Ptprc^bPep3^a (Ly5.2)] mice. To measure reconstitution and viral expression, peripheral blood was collected by retroorbital puncture 12 weeks after transplantation, and the percentage of GFP-positive leukocytes was determined. For colony-forming unit-spleen (CFU-S) assays, 7.5 × 10⁴ purified bone marrow cells were transplanted into lethally irradiated mice. After 12 days the mice were killed, and spleens were removed, weighed, fixed in Carnoy's solution, and examined for macroscopic colonies. All experiments were approved by the Melbourne Health Animal Ethics Committee.

Flow Cytometry.

Cells from bone marrow, spleen, thymus, lymph node, or peripheral blood were prepared as described (38) and analyzed by flow cytometry with monoclonal antibodies against the following antigens used as conjugates to allophycocyanin (APC), phycoerythrin (PE), Alexa Fluor 594, or biotin followed by PE–avidin or peridinin–chlorophyll–protein complex (PerCp)–Cy5.5–avidin: Ly5.2 (AL1-4A2), Ly 5.1 (A20-1.1), Mac-1 (M1/70), Gr-1 (RA6-8C5), FcR-γII/III (2.4G2), B2220 (RA3-6B2), CD19 (ID3), Ter119, CD4 (GK1.5), CD8 (YTS 169.4), c-kit (2B8), Sca-1 (E13-161-7), IL-7R-α (A7R34), CD34 (RAM34), and CD71 (C2). The cells expressing retrovirus were detected by GFP fluorescence. All antibodies were purchased from BD Biosciences (Franklin Lakes, NJ). Lin⁻/Sca-1/c-kit cells were sorted based on the Lin⁻/Sca-1^hi/c-kit^hi immunophenotype. Common myeloid progenitor (CMP), GMP, and megakaryocyte erythroid progenitor (MEP) populations from bone marrow were analyzed by immunomagnetic bead depletion of lin⁺ cells followed by staining with c-kit–APC, Sca-1–Alexa Fluor 594 FcR-γII/III–PE and CD34–biotin followed by PE–Cy7–avidin as described (24, 25). Sorting and analysis was performed on a DiVa high-speed sorter or an LSR flow cytometer (BD Biosciences).

Hematological Analysis.

Analysis of orbital plexus blood was performed by using an ADVIA 120 blood analyzer equipped with a mouse analysis software module (Bayer, Tarrytown, NY). Differential counts were performed manually on May–Grünwald–Giemsa-stained blood smears and cytocentrifuge preparations of sorted populations from femoral bone marrow and spleen. Clonal cultures of hematopoietic cells were performed as described (38). To assess the transplantability of leukemias, 1 × 10⁶ bone marrow or spleen cells from leukemic mice were injected i.v. or s.c. into nonirradiated C57BL/6 mice, and the mice were monitored for the onset of disease.

Abbreviations

CFC

colony-forming cell

AML

acute myeloid leukemia

MSCV

murine stem cell virus

GMP

granulocyte/macrophage progenitor.