Suppressor screen in Mpl^-/- mice: c-Myb mutation causes supraphysiological production of platelets in the absence of thrombopoietin signaling

Abstract

Genetic screens in lower organisms, particularly those that identify modifiers of preexisting genetic defects, have been used successfully to order components of complex signaling pathways. To date, similar suppressor screens have not been used in vertebrates. To define the molecular pathways regulating platelet production, we have executed a large-scale modifier screen with genetically thrombocytopenic Mpl^-/- mice by using N-ethyl-N-nitrosourea mutagenesis. Here we show that mutations in the c-Myb gene cause a myeloproliferative syndrome and supraphysiological expansion of megakaryocyte and platelet production in the absence of thrombopoietin signaling. This screen demonstrates the utility of large-scale N-ethyl-N-nitrosourea mutagenesis suppressor screens in mice for the simultaneous discovery and in vivo validation of targets for therapeutic discovery in diseases for which mouse models are available.

Mice and humans are physiologically similar, are afflicted by many of the same diseases, and show sufficient conservation at the genomic level that mutations in humans can often recapitulate disease when introduced into the germline of mice. With the discovery that N-ethyl-N-nitrosourea (ENU) is a potent germline mutagen in mice, genetic screens have now become feasible in a mammalian model of direct relevance to human health and disease (1). Some ENU mutagenesis screens in mice have focused on particular traits such as circadian rhythm or on particular regions of the genome, whereas others have screened for anomalies in a wide range of tissues and organs (2–6). Unlike screens in lower organisms (7–9), to date all of these studies have begun with wild-type mice and have isolated mutants with abnormal traits. The successful application of modifier screens in lower organisms, with prominent examples including the dissection of sevenless-dependent eye development in Drosophila melanogaster (10) and vulval development in Caenorhabditis elegans (11), led us to explore the use of suppressor screens in mice.

In this study, we describe a large-scale suppressor screen, the aim of which was to identify mutations capable of ameliorating thrombocytopenia, a lack of blood platelets. Blood platelets are shed by megakaryocytes into the circulation, where they are required for blood clotting and hemostasis. Thrombopoietin (TPO), acting through its specific cell surface receptor c-Mpl, is considered the principal cytokine controlling megakaryocyte and platelet numbers (12, 13). Mice and humans lacking functional Tpo or Mpl genes are profoundly thrombocytopenic and have a corresponding reduction in the numbers of megakaryocytes, megakaryocyte progenitor cells, and stem cells (14–21). We screened 1,575 Mpl^-/- mice harboring random ENU-induced mutations for amelioration of thrombocytopenia, resulting in isolation of two independent partial loss-of-function alleles of c-Myb. When homozygous, these mutations produced a supraphysiological expansion of megakaryocyte and platelet production in the absence of signaling by the major regulator of thrombopoiesis, TPO.

Materials and Methods

Generation and Screening of Mutant Mice. Male Mpl^-/- C57BL/6 mice (14) were treated with a total dose of 200–400 mg/kg ENU divided into one, two, or three weekly injections, as described (22). Four weeks after the final injection, ENU-treated mice were mated with one or two isogenic female mice to yield first-generation (G₁) progeny. At 7 weeks of age, blood from G₁ mice was collected from the retroorbital plexus into tubes containing potassium EDTA (Sarstedt), and the number of platelets in the peripheral blood was determined by using an Advia 120 automated hematological analyzer (Bayer, Tarrytown, NY).

Test for Linkage Between Plt3 and Plt4. To determine whether Plt3 and Plt4 were genetically linked, heterozygous Plt3/+ Mpl^-/- and Plt4/+ Mpl^-/- mice on a C57BL/6 background were crossed, and the progeny were bled at 7 weeks of age to determine their platelet count. These mice were then bred with +/+ Mpl^-/- mice, and the numbers of platelets in these progeny were determined at 7 weeks of age.

Analyses of Epistasis. Plt4/Plt4 Mpl^-/- mice were mated to +/+ Mpl^+/+ mice to produce Plt4/+ Mpl^+/- animals, which were then intercrossed. The Mpl genotype of the progeny was determined by Southern blot (14), and the genotype of the Plt4 locus was inferred from one or two generation progeny testing. The phenotypes of Plt4/Plt4 Mpl^-/-, Plt4/Plt4 Mpl^+/+, and +/+ Mpl^-/- mice were then compared to assess the epistatic relationship of the genes.

Genetic Mapping. Affected heterozygous Plt4/+ Mpl^-/- mice on a C57BL/6 background were crossed to Mpl^+/- mice on a 129Sv background. Plt4/+ Mpl^-/- F₁ animals were then identified at 7 weeks of age because of their elevated platelet counts and intercrossed to produce 65 mice in the F₂ generation. At 7 weeks of age, F₂ mice were bled and their peripheral blood platelet numbers were determined. Mice were killed, and DNA was prepared from a piece of liver according to described methods (23). One hundred forty-eight simple sequence length polymorphisms (SSLPs) spaced evenly throughout the genome were amplified and analyzed, essentially as described (24). Plt4 was found to reside on chromosome 10, and its location was refined through analysis of additional markers in the region. Tight genetic linkage of Plt3 to Plt4 was confirmed by crossing Plt3/+ Mpl^-/- mice on a C57BL/6 background to +/+ Mpl^+/- mice on a 129Sv background. Plt3/+ Mpl^-/- F₁ animals were then identified at 7 weeks of age because of their elevated platelet counts and were backcrossed to +/+ Mpl ^-/- mice on a 129Sv background to generate N₂ mice. The platelet counts of 68 Mpl^-/- N₂ mice were determined at 7 weeks of age, and their genotype at SSLP markers closely linked to Plt4 was assessed.

DNA Sequencing. Mice generated by intercrossing either Plt3/+ Mpl^-/- or Plt4/+ Mpl^-/- mice were typed as Plt3/Plt3, Plt3/+, Plt4/Plt4, Plt4/+, or +/+ at 7 weeks of age by determining their peripheral blood platelet number. Using a tail biopsy taken at ≈3 weeks of age, DNA was prepared, and each of the exons of the c-Myb gene (25) was amplified by PCR and sequenced on an Applied Biosystems automatic sequencer according to the manufacturer's instructions.

Transactivation Assays. To compare the activity of c-Myb^Plt3 and c-Myb^Plt4 with that of wild type c-Myb and a constitutively active truncation mutant (CT3) of c-Myb (26), the respective cDNAs were cloned into the pEF-BOS vector (27). The c-Myb expression constructs were then cotransfected into 293 cells with a reporter construct containing five consecutive high-affinity c-Myb-binding sites from the chicken Mim-1 promoter upstream of the CAT gene (a gift of Joe Lipsick, Stanford University, Stanford, CA), and transactivation activity was measured as described (28). The expression of the various c-Myb alleles was measured by Western blotting by using the Mab1.1 and Mab 5.1 antibodies as described (29).

Hematological Analysis. The hematocrit, platelet and white cell count, and differential were determined by using either manual or automated (Advia 120, Bayer) counting techniques. Clonal cultures of hemopoietic cells were performed as described (23). Cultures of 2.5 × 10⁴ adult bone marrow cells or 5 × 10⁴ spleen cells in 1 ml of 0.3% agar in DMEM supplemented with newborn calf serum (20%) were stimulated with a mixture of 100 ng/ml murine stem cell factor, 10 ng/ml murine IL-3, and 4 units/ml human EPO and incubated for 7 days at 37°C in a fully humidified atmosphere of 10% CO₂ in air. Agar cultures were fixed; sequentially stained for acetylcholinesterase, Luxol fast blue, and hematoxylin; and the cellular composition of each colony determined at ×100–400. Megakaryocyte counts were performed by manual counting from sections of sternum and spleen after staining with hematoxylin/eosin. A minimum of 10 high-power fields (×200) were scored. Colony-forming units (spleen) were enumerated by i.v. injection of bone marrow cells into recipient mice that had been irradiated with 11 Gy of γ-irradiation in two equal doses given 3 hours apart from a ¹³⁷Cs source (Atomic Energy, Ottawa). Transplanted mice were maintained on oral antibiotic (1.1 g/liter neomycin sulfate; Sigma). Spleens were removed after 12 days, fixed in Carnoy's solution (60% ethanol/30% chloroform/10% acetic acid), and the numbers of macroscopic colonies were counted.

Flow Cytometry. Single-cell suspensions of spleen and bone marrow cells were depleted of erythrocytes by lysis with 156 mM ammonium chloride (pH 7.3). Cells were stained with a saturating concentration of IgM-FITC and B220–phycoerythrin (PE) or Ter119-PE and CD71-FITC (30) (BD Pharmingen), and analyses were performed on an LSR flow cytometer (BD Biosciences, San Diego). Dead cells were excluded based on propidium iodide staining.

Results

A Large-Scale Random Mutagenesis Screen to Isolate Suppressors of Thrombocytopenia. The aim of this genetic screen was to identify mutations capable of ameliorating thrombocytopenia. The mice we used had markedly reduced platelet numbers (117 ± 52 × 10⁶/ml, n = 783 vs. 1,168 ± 230 × 10⁶ platelets per ml, n = 100 for isogenic C57BL/6 controls, Fig 1 a and b), caused by a loss-of-function deletion in the c-Mpl gene, which encodes the cell surface receptor for the platelet-regulating cytokine, TPO (12).

Fig. 1.

Identification of ENU-mutant mice with ameliorated thrombocytopenia. Peripheral blood platelet counts at 7 weeks of age for 100 Mpl^+/+mice (a), 783 Mpl^-/- mice (b), 1,575 G₁ Mpl^-/- mice harboring random ENU-induced mutations (c), the Plt3 G₁ mouse (d), the G₂ progeny derived from mating the G₁ Plt3 mouse to Mpl^-/- mice (e), the progeny derived from crossing heterozygous Plt3/+ Mpl^-/- mice (f), the G₁ Plt4 mouse (g), the G₂ progeny derived from mating the G₁ Plt4 mouse to Mpl^-/- mice (h), the progeny derived from intercrossing heterozygous Plt4/+ Mpl^-/- mice (i), the progeny derived from mating homozygous Plt4/Plt4 Mpl^-/- mice to Mpl^-/- mice (j), and the progeny derived from intercrossing homozygous Plt4/Plt4 Mpl^-/- mice (k).

Three hundred Mpl^-/- C57BL/6 mice were injected with 200–400 mg/kg ENU in one, two, or three weekly doses. Four weeks after injection, males were mated with isogenic Mpl^-/- C57BL/6 female mice to produce first-generation (G₁) offspring. At 7 weeks of age, we analyzed the peripheral blood of 1,575 G₁ Mpl^-/- mice and found five mice with >300 × 10⁶ platelet/ml, a count >3 SD higher than the mean observed in Mpl^-/- C57BL/6 mice (Fig. 1 c, d, and g).

To determine whether suppression of thrombocytopenia was heritable, the five G₁ animals were crossed to untreated Mpl^-/- mice, and the platelet numbers in their progeny were measured. The progeny of two of the five mice (named Plt3 and Plt4) contained individuals with the suppressed phenotype, and this trait showed simple Mendelian inheritance consistent with a dominant phenotype (Fig. 1 e and h). In one case, the G₁ mouse died before yielding progeny, whereas in the remaining two cases, suppression was not heritable (data not shown). Because each of the affected G₁ mice was derived from a different ENU-treated Mpl^-/- mouse, this suggested that independent dominant suppressors of thrombocytopenia had been isolated.

Plt3 and Plt4 Cause Semidominant Phenotypes with Homozygous Mice Exhibiting Supraphysiological Increases in Platelet Numbers. Although heterozygous Plt3 and Plt4 mutations led to amelioration of thrombocytopenia, platelet counts did not reach wild-type levels. We wished to determine whether platelet counts were further elevated in homozygous mice. Offspring from intercrossing Plt4/+ Mpl^-/- mice displayed either the low platelet counts typical of Mpl^-/- mice, mild amelioration of thrombocytopenia, or supraphysiological platelet levels (Fig. 1i). These phenotypes were shown to correspond to +/+ Mpl^-/-, Plt4/+ Mpl^-/-, and Plt4/Plt4 Mpl^-/- genotypes by breeding experiments. Notably, when Plt4/Plt4 Mpl^-/- mice were crossed to +/+ Mpl^-/- mice, all of the progeny had mild suppression of thrombocytopenia characteristic of Plt4/+ Mpl^-/- mice; whereas if Plt4/Plt4 Mpl^-/- mice were intercrossed, then all of the pups exhibited supraphysiological platelet numbers characteristic of their parents (Fig. 1 j and k). Although the proportions of these three phenotypes were consistent with full penetrance of the phenotype and viability of the homozygotes for Plt4, a lower than expected frequency of homozygotes was observed for Plt3 (Fig. 1f), and this was shown to be due to embryonic or neonatal death of a proportion of the homozygous mice rather than a lack of penetrance of the phenotype (data not shown).

Plt3 and Plt4 Are Independent Hypomorphic Alleles of c-Myb. To determine whether Plt3 and Plt4 were genetically linked, Plt3-Plt4 compound heterozygotes were produced and mated to Mpl^-/- mice. All progeny of the Plt3-Plt4 compound heterozygotes exhibited mild amelioration of thrombocytopenia, typical of Plt3/+ Mpl^-/- and Plt4/+ Mpl^-/- mice, demonstrating that the Plt3 and Plt4 mutations are tightly linked and could be alleles of the same gene (data not shown).

To obtain a chromosomal localization for the closely linked Plt3 and Plt4 mutations, Plt4/+ Mpl^-/- mice (C57BL/6 background) were mated with +/+ Mpl^+/- mice on a 129Sv background and then Plt4/+ Mpl^-/- (C57BL/6 × 129Sv)F₁ progeny were intercrossed to produce 65 F₂ generation mice. Platelet counts in these F₂ mice varied from the very low levels expected of +/+ Mpl^-/- mice, intermediate levels characteristic of Plt4/+ Mpl^-/- mice, and exceptionally high levels characteristic of Plt4/Plt4 Mpl^-/- mice. Using a set of 148 simple sequence length polymorphisms, the only region of the genome in which linkage was observed was at the centromeric end of chromosome 10 between markers D10Mit213 and D10Mit214 (Fig. 2a). Analysis of 68 similarly generated backcross progeny demonstrated that this interval was also closely linked to the Plt3 mutation (Fig. 2a).

Fig. 2.

Plt3 and Plt4 are tightly linked on chromosome 10 and are alleles of c-Myb. (a) To map Plt4, a (C57BL/6 × 129Sv)F₂ generation of 65 mice was produced, bled at 7 weeks of age, and categorized as having low platelets (<150 × 10⁶/ml) characteristic of +/+ Mpl^-/- mice, moderate numbers of platelets (150 × 10⁶ to 2,000 × 10⁶/ml) characteristic of Plt4/+ Mpl^-/- mice, or extremely high platelets (>2,000 × 10⁶/ml) characteristic of Plt4/Plt4 Mpl^-/- mice. Animals were then genotyped, and markers found to be homozygous 129/Sv are shown in white, markers that were heterozygous are shown in gray, and markers homozygous C57BL/6 are shown in black. The number of animals with each haplotype is shown below. The Plt4 mutation was localized to between D10Mit213 and D10Mit 214. To confirm Plt3 was located close to Plt4, we produced 68 (C57BL/6 × 129Sv)N₂ generation mice, measured platelet numbers in these animals, and genotyped them by using the markers most closely linked to Plt4. (Right) Correlation between genotype and phenotype for markers at the centromeric region of chromosome 10. (b) Sequence of PCR-amplified exons and intron boundaries of the c-Myb gene showing a single A to T transversion in Plt3 and Plt4, resulting in an Asp to Val substitution at positions 152 and 382, respectively. Representative traces are shown; a total of three Plt3/Pl3 Mpl^-/-, three Plt3/+ Mpl^-/- mice, three Plt4/Plt4 Mpl^-/- mice, three Plt4/+ Mpl^-/- mice, three +/+ Mpl^-/- mice, and three +/+ Mpl^+/+ mice were analyzed. (c) The transactivation activity of wild-type c-Myb, Plt3 and Plt4 mutant c-Myb, and a constitutively activated truncation of c-Myb (CT3; ref. 26) were compared by measuring production of chloramphenicol acetyltransferase from a c-Myb responsive promoter (28). The activity of both Plt3 and Plt4 c-Myb was significantly lower than wild type; (P = 0.017 and P = 0.0003, respectively; n = 9).

Within this chromosomal region, the c-Myb locus represented a compelling candidate for the location of the Plt3 and Plt4 mutations, because c-Myb mutations have been found to elevate platelet numbers (31, 32). The coding regions of the entire c-Myb gene from Plt3/+, Plt3/Plt3, Plt4/+, and Plt4/Plt4 Mpl^-/- mice were sequenced and compared with that from +/+ Mpl^-/- and +/+ Mpl^+/+ mice. A single A to T transversion in the c-Myb coding sequence was discovered in both mutants, resulting in substitution of the valine for an aspartic acid codon at residue 152 of the c-Myb DNA-binding domain in the Plt3 allele and at residue 384 in the leucine-zipper domain of the Plt4 allele (Fig. 2b). Northern blot analysis revealed that c-Myb^Plt3 and c-Myb^Plt4 mRNAs were expressed in hematopoietic tissues at normal levels and, on transfection of expression vectors into 293 cells, the Plt3 and Plt4 c-Myb proteins were produced at levels similar to wild-type c-Myb (data not shown). In contrast, compared to wild-type c-Myb and a constitutively active c-Myb mutant, there was a profound reduction in the activity of c-Myb^Plt4 and a more modest reduction of c-Myb^Plt3 proteins in a transactivation assay using a c-Myb-responsive reporter gene (Fig. 2c). The conclusion that partial loss-of-function of c-Myb can ameliorate thrombocytopenia was confirmed by taking advantage of a previously generated mouse (33) with deletion of one allele of the c-Myb gene (c-Myb^+/-) and showing that, similar to c-Myb^Plt3/+ Mpl^-/- and c-Myb^Plt4/+ Mpl^-/- mice, c-Myb^+/- Mpl^-/- mice had significantly higher numbers of platelets than c-Myb^+/+Mpl^-/- mice (Fig. 3a).

Fig. 3.

Mutation of c-Myb results in an elevation in progenitor cells, megakaryocytes, and platelets independent of Mpl. (a) The numbers of colony-forming units (spleen), a measure of multipotential progenitor cells, in the bone marrow (Far Left), clonogenic megakaryocyte progenitor cells (Left), megakaryocytes (Right) and platelets (Far Right)in c-Myb^+/+ (+/+), c-Myb ^Plt4/+ (4/+), c-Myb ^Plt4/Plt4 (4/4), c-Myb ^Plt3/+ (3/+), and c-Myb ^Plt3/Plt3 (3/3) mutants on a Mpl^-/- or Mpl^+/+ background are shown. Assays were performed as described in Materials and Methods with the error bars representing the SD from the mean of data from 3–7 [colony-forming units (spleen)], 2–6 (progenitor cell data), 2–7 (megakaryocyte data), and 4–50 (platelet data) mice. (b) Flow cytometric analysis of B lymphoid cells in the spleen and erythroid cells in the bone marrow of Plt4 mutant mice showing marked reductions in preB (B220⁺IgM^-) and B (B220⁺IgM⁺) lymphocytes and accumulation of more immature erythroid cells (CD71^hiTer119^med and CD71^hiTer119^hi; ref. 31) in Mpl^-/- c-Myb^Plt4/Plt4 mice. (c) Histological sections of spleens from Mpl^+/+ c-Myb^+/+, Mpl ^-/- c-Myb^Plt3/Plt3, Mpl^-/- c-Myb^Plt4/Plt4, and Mpl^+/+ c-Myb^Plt4/Plt4 mice. Note the poor development of lymphoid follicles and expanded red pulp displaying reduced cellularity and disrupted architecture.

Plt3 and Plt4 Ameliorate Thrombocytopenia by Increasing Megakaryocytopoiesis. To investigate the biological basis for the amelioration of Mpl^-/- thrombocytopenia due to loss of function of c-Myb, we analyzed megakaryocytopoiesis in Plt3 and Plt4 mutant mice. Significant increases in the numbers of megakaryocyte progenitor cells and megakaryocytes in the bone marrow and spleen of heterozygous and, to a greater extent, homozygous Plt3 and Plt4 mice were observed relative to Mpl^-/- controls (Fig. 3a), consistent with the increase in platelet numbers in these mice being the result of expanded cellular production within the megakaryocyte lineage. The numbers of progenitor cells committed to other hemopoietic lineages were not significantly altered in heterozygous Plt3 or Plt4 mutants; however, in homozygous Plt3 and Plt4 mice, the numbers of all progenitor cells were elevated compared with control Mpl^-/- mice (Table 1). Perturbations in the multipotent progenitor compartment seem likely to account for this observation, because significant increases in the numbers of spleen colony-forming cells accompanied these changes (Fig. 3a). Although the numbers of spleen colony-forming units in the bone marrow of homozygous Plt3 and Plt4 mutants were elevated, the spleen colonies derived from them were smaller than those observed in mice transplanted with heterozygous or Mpl^-/- cells.

Table 1. Hematological profile of Plt3 and Plt4 mutant mice.

	Mpl-/-					Mpl+/+
	c-Myb+/+	c-MybPlt4/Plt4	c-MybPlt4/+	c-MybPlt3/Plt3	c-MybPlt3/+	c-Myb+/+	c-MybPlt4/Plt4
Peripheral blood
Platelets (×10-6 ml)	116 ± 52	4,662 ± 851	507 ± 132	4371 ± 718	449 ± 77	1,479 ± 128	3936 ± 618
Hematocrit, %	52.7 ± 2.1	43.9 ± 2.3	50.1 ± 5.3	43.3 ± 2.5	50.0 ± 5.9	51.3 ± 1.8	45.3 ± 4.8
White cells (×10-3/ml)	6.30 ± 1.57	3.26 ± 1.09	7.26 ± 2.7	4.76 ± 1.41	6.67 ± 1.22	7.80 ± 1.94	4.73 ± 0.35
Neutrophils	0.48 ± 0.46	0.79 ± 0.35	0.89 ± 0.71	1.60 ± 0.59	0.58 ± 0.61	0.43 ± 0.19	1.53 ± 0.52
Lymphocytes	5.54 ± 1.16	2.06 ± 0.73	5.70 ± 2.56	1.94 ± 0.85	4.92 ± 2.95	6.89 ± 1.86	2.49 ± 0.21
Monocytes	0.39 ± 0.45	0.39 ± 0.20	0.38 ± 0.15	0.89 ± 0.52	0.52 ± 0.42	0.36 ± 0.29	0.60 ± 0.15
Eosinophils	0.07 ± 0.10	0.02 ± 0.02	0.11 ± 0.07	0.22 ± 0.28	0.15 ± 0.01	0.11 ± 0.08	0.12 ± 0.09
Colony-forming progenitor cells
BM (per 2.5 × 104 cells)
Blast	4 ± 3	17 ± 4	4 ± 2	21 ± 4	10 ± 3	11 ± 4	12 ± 6
Granulocyte	12 ± 4	44 ± 15	14 ± 4	11 ± 5	17 ± 6	19 ± 4	23 ± 20
Granulocyte/macrophage	9 ± 4	19 ± 9	11 ± 5	31 ± 9	11 ± 4	12 ± 2	27 ± 1
Macrophage	7 ± 3	14 ± 12	7 ± 3	17 ± 5	10 ± 5	8 ± 5	14 ± 1
Eosinophil	0.9 ± 0.9	0 ± 0	1 ± 1	0 ± 0	0.7 ± 1.2	3 ± 3	0 ± 0
Megakaryocyte	9 ± 6	113 ± 16	26 ± 6	96 ± 24	25 ± 14	21 ± 6	96 ± 25
Spleen (per 5×104 cells)
Blast	0.1 ± 0.2	5 ± 4	1 ± 0.8	3 ± 4	0.2 ± 0.3	1 ± 1	2 ± 1.4
Granulocyte	0.4 ± 0.5	12 ± 5	0 ± 0	4 ± 3	0 ± 0	0.6 ± 0.8	8 ± 1.4
Granulocyte/macrophage	0.5 ± 0.8	7 ± 5	0 ± 0	3 ± 3	0 ± 0	0.2 ± 0.4	12 ± 6
Macrophage	0.4 ± 0.8	5 ± 4	0.5 ± 0.1	4 ± 7	0.2 ± 0.3	0.2 ± 0.4	15 ± 8
Eosinophil	0.1 ± 0.4	0 ± 0	0 ± 0	0 ± 0	0 ± 0	0 ± 0	0 ± 0
Megakaryocyte	4 ± 5	100 ± 23	6 ± 5	91 ± 105	3 ± 2	7 ± 7	126 ± 24

Means ± SD of data from 15—50 (blood data) and 2—6 (progenitor cell data) mice of each genotype. BM, bone marrow.

Plt3 and Plt4 Homozygous Mice Exhibit Defects in Erythroid and B Lymphoid Lineages. With the exception of megakaryocytopoiesis, this multilineage myelodysplastic state did not result in excessive production of other types of mature blood cells in homozygous mutants (Table 1). The proportions of morphologically recognizable cells in the granulocyte series were slightly elevated in bone marrow and spleen of Plt3 and Plt4 mutants, but this was likely to be at least in part an indirect effect of reduced lymphoid cell production (see below), and monocyte numbers were relatively unchanged (data not shown). Previous studies (31, 32) have found that reduced expression of c-Myb results in defective production of cells in the lymphoid and erythroid lineages. Our analyses of the effects of reduced c-Myb function in Plt3 and Plt4 mutants supports these observations. Histological examination of the spleen revealed that abnormally small lymphoid follicles and expanded red pulp with atypical architecture accompanied the greatly expanded numbers of megakaryocytes (Fig. 3c). Significantly reduced numbers of B lymphoid cells were evident in the bone marrow and spleen (Fig. 3b and data not shown), and homozygous Plt3 and Plt4 mice were leukopenic, due largely to reduced numbers of circulating lymphocytes (Table 1). Similarly, a large increase in the numbers of colony-forming unit erythroid was observed in the spleens of homozygous mutant mice (Mpl^-/- c-Myb^Plt3/Plt3, 155 ± 143; Mpl^-/- c-Myb^Plt4/Plt4,154 ± 92; Mpl^-/- c-Myb^+/+, 3 ± 3 per 10⁵ cells), defects in erythroid maturation were evident in the bone marrow (Fig. 3b), and the animals were mildly anemic (Table 1).

c-Myb Is Epistatic to Mpl. An advantage of modifier screens is their capacity to order genes in a pathway using analyses of epistasis. Accordingly, we assessed the phenotype of homozygous Plt4 mutants on a wild-type background (c-Myb^Plt4/Plt4 Mpl^+/+) in comparison with that of wild-type (c-Myb^+/+Mpl^+/+) and c-Myb^Plt4/Plt4 Mpl^-/- mice. Remarkably, the supraphysiological production of platelets, megakaryocytes, and megakaryocyte progenitors observed in c-Myb^Plt4/Plt4 mice was independent of the Mpl genotype (Table 1 and Fig. 3).

Discussion

Previous studies have implied a complex role for c-Myb in regulation of hemopoiesis (31–33). Deletion of the c-Myb gene results in embryonic death due to failure of fetal liver hemopoiesis (33). In contrast, mice bearing a hypomorphic allele of c-Myb survive and exhibit increased megakaryocyte production but diminished production of erythroid and lymphoid lineages (32). This spectrum of phenotypes is reproduced in mice with Plt3 or Plt4 mutations consistent with partial loss of Myb function. However, the availability of hypomorphic point mutations in Myb should allow a finer dissection of the physiological roles of specific c-Myb subdomains. In this regard, a similar phenotype to that in Plt3/Plt3 and Plt4/Plt4 mice was observed when a germline mutation in the KIX domain of p300, which disrupts Myb binding, was established on a Myb^+/- background (31). These data suggest that disrupted interaction of Myb with p300 predisposes to thrombocytosis. It will therefore be of interest to determine whether the mutations that arose in Plt3 and Plt4 mice affect the interaction of Myb with p300. Moreover, the data here establish, quite unexpectedly, that reduction of c-Myb function results in a supraphysiological megakaryocytopoiesis and platelet production in the complete absence of signaling by TPO.

The question arises whether inhibition of c-Myb and TPO represents part of the same pathway or lies on independent routes leading to platelet formation. To begin to address this issue, we assessed the epistatic relationship of c-Myb and Mpl by comparing the number of platelets observed in c-Myb^Plt4/Plt4 Mpl^+/+ mice with those of c-Myb^Plt4/Plt4 Mpl^-/- and c-Myb^+/+ Mpl^+/+ mice. We observed a similar increase in platelet number to between 3,000 and 4,000 × 10⁶/ml in c-Myb^Plt4/Plt4 mice whether c-Mpl was present or absent. One interpretation of this result is that a critical step in the pathway by which TPO generates megakaryocytes and platelets is down-regulation of c-Myb expression or activity; hence when c-Myb activity is reduced by mutation, the hemopoietic system of the animal responds in a manner analogous to exposure to a high concentration of TPO. Biochemical analyses of c-Myb expression and action after TPO stimulation of hemopoietic progenitor cells and megakaryocytes will be required to test this hypothesis.

Comprehensive genome-wide and targeted mutagenesis screens using wild-type mice have been reported recently (2, 3, 5, 6). This paper describes a large-scale modifier screen in vertebrates and demonstrates that the strategies that have proven so valuable in yeast, worms, and flies (7–9) are also applicable in higher organisms. In addition to their use in dissecting complex biological processes, we propose that suppressor screens in vertebrates are of potential value for the identification of targets for drug discovery. Just as most ENU-induced mutations cause loss of function, most small-molecule therapeutics also reduce the function of proteins to which they bind. Accordingly, screens for genes that, when mutated, lead to amelioration of disease should provide genome-wide access to novel in vivo validated targets for pharmaceutical discovery in diseases with unmet clinical need for which mouse models are available.