Ott1(Rbm15) has pleiotropic roles in hematopoietic development

Abstract

OTT1(RBM15) was originally described as a 5′ translocation partner of the MAL(MKL1) gene in t(1,22)(p13;q13) infant acute mega karyocytic leukemia. OTT1 has no established physiological function, but it shares homology with the spen/Mint/SHARP family of proteins defined by three amino-terminal RNA recognition motifs and a carboxyl-terminal SPOC (Spen paralog and ortholog carboxyl-terminal) domain believed to act as a transcriptional repressor. To define the role of OTT1 in hematopoiesis and help elucidate the mechanism of t(1,22) acute megakaryocytic leukemia pathogenesis, a conditional allele of Ott1 was generated in mice. Deletion of Ott1 in adult mice caused a loss of peripheral B cells due to a block in pro/pre-B differentiation. There is myeloid and megakaryocytic expansion in spleen and bone marrow, an increase in the Lin⁻Sca-1⁺c-Kit⁺ compartment that includes hematopoietic stem cells, and a shift in progenitor fate toward granulocyte differentiation. These data show a requirement for Ott1 in B lymphopoiesis, and inhibitory roles in the myeloid, megakaryocytic, and progenitor compartments. The ability of Ott1 to affect hematopoietic cell fate and expansion in multiple lineages is a novel attribute for a spen family member and delineates Ott1 from other known effectors of hematopoietic development. It is plausible that dysregulation of Ott1-dependent hematopoietic developmental pathways, in particular those affecting the megakaryocyte lineage, may contribute to OTT1-MAL-mediated leukemogenesis.

Infant acute megakaryocytic leukemias in non-Down's syndrome children harbor a unique chromosomal translocation, t(1,22)(p13;q13) (1, 2). This translocation fuses two novel genes believed to be transcription factors, OTT1 (RBM15) and MAL (MKL1, BSAC, or MRTF-A), and results in expression of a chimeric protein (3, 4). The breakpoint occurs in the first intron of OTT1 and in either the third or fourth intron of MAL. The chimeric protein OTT1-MAL (RBM15-MKL1) contains almost the entire coding sequence of both genes. Neither gene had previously been described. Although MAL has recently been implicated in serum response, thus far no physiologic roles have been elucidated in hematopoiesis for either gene (5).

Ott1 is ubiquitously expressed and belongs to a larger gene family related to Drosophila spen (split ends) (3, 4). The spen family is defined by three amino-terminal RNA recognition motifs and a carboxyl-terminal SPOC (Spen paralog and ortholog carboxyl-terminal) transactivation/repression domain (6). spen is involved in a wide range of developmental processes including neural and eye development and was identified as a positive regulator of egfr and ras pathways (7, 8). spen modulates hox gene expression in head and thorax development and down-regulates the Notch effector suppressor of hairless (9, 10). spen is thought to regulate developmental cell fate and spatial positioning processes through these pathways (8–11).

The human and murine homologs of spen, SHARP (SMRT/HDAC1-associated repressor protein), and Mint (Msx-2-interacting nuclear target protein) are thought to provide similar functions in humans and mice, respectively (12, 13). SHARP is a component of a corepressor complex including the proteins SMRT (Silencing mediator for retinoic and thyroid hormone receptors) and NCoR (Nuclear receptor Corepressor), which are required for RBP-Jκ-mediated repression (14–16). Mint acts as a transcriptional repressor in concert with the homeodomain protein Msx-2 (13). Mint also interacts with RBP-Jκ and is a suppressor of Notch signaling through a competitive binding mechanism (17). Structural analysis of the SHARP SPOC domain suggests that the regions responsible for binding SMRT/NCoR are highly conserved between all spen family members, including OTT1 (18, 19). SHARP also associates with SRA (steroid receptor RNA coactivator) and modulates steroid response (12). SHARP, OTT1, and OTT3 associate with EB2 (Epstein–Barr virus early protein), an RNA export factor (20). Although in vitro studies of Mint demonstrate repressor activity on the Osteocalcin promoter, transcriptional up-regulation is observed when tested with a basal promoter, raising the possibility of dual repressor/transactivator function (13).

Mint deficiency in mice is lethal around embryonic day 13.5 (E13.5) (17). Defects are noted in the heart, pancreas, and skin and the size of the fetal liver, but the frequency of hematopoietic CFU from fetal liver cells is normal. When fetal liver cells are transplanted into lymphocyte-deficient recipients, there is an increase in splenic marginal zone B lymphocytes at the expense of follicular B lymphocytes that is thought to reflect loss of Notch2 signaling by Mint. Taken together, the products of the spen family have similar and overlapping functions.

We have generated a murine allele enabling conditional deletion of Ott1 in vivo to gain an understanding of the physiologic role of OTT1 in hematopoiesis and its contribution to OTT1-MAL-associated acute megakaryocytic leukemia.

Results

Generation of a Conditional Ott1 Allele in Mice.

Ott1 comprises two exons and is expressed as two major transcripts differing by a facultative intron with all but 23 bp included in the first exon. Therefore, by flanking the first exon with loxP sites, conditional deletion of Ott1 can be achieved through expression of Cre recombinase [supporting information (SI) Fig. 5A]. The Ott1 allele was targeted in ES cells (SI Fig. 5 B and C), and chimeric mice were produced capable of germ-line transmission of the conditional allele, Ott1^flox. Mice with the Ott1^flox allele were crossed with transgenic EIIa-Cre, which excises broadly in the early embryo, generating mice with complete excision of the floxed allele, Ott1^null (SI Fig. 5 D and E).

Deletion of Ott1 Is Embryonic Lethal.

Mice heterozygous for the Ott1^null allele were bred together to generate homozygous null animals (Table 1). No viable Ott1^null/null offspring were obtained. In timed matings, there was expected Mendelian frequency of viable Ott1^null/null embryos through E9.5. However, Ott1^null/null embryos after E9.5 showed degeneration and resorption, indicative of an essential embryonic requirement for Ott1 during E9.5–E10.5.

Table 1.

Genotype of embryos from Ott1^null/wt × Ott1^null/wt mice

Days after conception*	No. of litters	No. of embryos	No. of wt/wt mice (%)	No. of null/wt mice (%)	No. of null/null mice (%)
8.5	4	28	9 (32)	13 (46)	6 (22)
9.5	8	73	19 (26)	38 (52)	16 (22)
10.5	10	75	22 (29)	31 (41)	21† (28)
11.5	1	7	3 (43)	2 (28.5)	2† (28.5)
12.5	1	8	1 (12.5)	4 (50)	3† (37.5)
13.5	2	13	5 (38)	8 (62)	0 (0)
Term	10	65	18 (28)	47 (72)	0 (0)

*The day of the vaginal plug is counted as E0.5.

^†All embryos observed undergoing degeneration.

Deletion of Ott1 in Adult Mice Causes Peripheral B Cell Lymphopenia.

Ott1^flox animals were crossed with Mx1-Cre transgenic mice that enable efficient conditional induction of Cre expression with polyinosinic-polycytidylic acid (pIpC) in the hematopoietic compartment of adult mice (21). Ott1^flox/null Mx1-Cre mice were bred in which a single excision event was required to generate homozygosity of the null allele.

Six-week-old mice were treated with pIpC and monitored with weekly nonlethal retroorbital eyebleeds. Total WBC gradually decreased over a 6-week period in Ott1^flox/null Mx1-Cre mice, compared with pIpC-treated littermates with at least one wild-type Ott1 allele (Fig. 1A). No abnormalities were observed between wild-type and hemizygous animals, or with Mx1-Cre alone (data not shown). The decrease in WBC resulted from a striking decline in lymphocytes (Fig. 1A). There were no significant differences in neutrophil count, hematocrit, or platelet counts (data not shown).

Fig. 1.

Loss of peripheral B lymphocytes after deletion of Ott1 in adult mice. Ott1^null/flox Mx1-Cre mice and control littermates with a wild-type allele were induced with pIpC and bled on a weekly basis. (A) Total leukocyte count (WBC). (B) Absolute lymphocyte count. Open squares, Ott1^null/flox Mx1-Cre (n = 4); open triangles, littermates (n = 5). (B) Representative flow-cytometry plots of peripheral blood at 6 weeks after pIpC induction with percentage of cells noted in each quadrant. (C) Absolute numbers of peripheral blood subsets at 6 weeks after pIpC. Gray bars, Ott1^null/flox Mx1-Cre (n = 3); white bars, littermates (n = 5). (D) Agarose gel electrophoresis of PCR from peripheral blood DNA samples to detect Ott1 flox or null alleles. (Left) PCR from before and 1 week after pIpC induction blood DNA. (Right) PCR from weeks 4–8 after pIpC.

Flow-cytometric analysis of peripheral blood 6 weeks after pIpC induction showed a dramatic decrease in CD19⁺/B220⁺ peripheral B cells in Ott1^flox/null Mx1-Cre mice (Fig. 1 B and C). Although the absolute numbers of B220⁺ cells were decreased 20-fold, the relative percentages of IgM⁺IgD⁺ and IgM⁻IgD⁺ were preserved (Fig. 1B). CD4⁺ and CD8⁺ T cell populations were not significantly changed, nor were cells with monocytic and granulocytic markers (Fig. 1C and SI Fig. 6).

After 8 weeks, the lymphocyte counts in Ott1^flox/null Mx1-Cre mice began to rise and eventually returned to baseline levels (data not shown). Although the Ott1^flox allele was undetectable by PCR in DNA isolated from the peripheral blood 1 week after pIpC, the allele reappeared after 8 weeks (Fig. 1D). The correlation between the rise in lymphocytes and the recurrence of the Ott1^flox allele indicates a strong negative selective pressure in the B lymphocyte population with deletion of Ott1.

Ott1 Is Required for the Development of Pro/Pre-B Cells.

Flow cytometry of the bone marrow B cell compartment showed that induced Ott1^flox/null Mx1-Cre mice had equivalent numbers of B220⁺/CD43⁺ pro- to early pre-B cells (Fig. 2 A and B). Later pre-B development is characterized by loss of CD43 and acquisition of CD19 and CD25 (22). Ott1^flox/null Mx1-Cre mice show a >50% decrease in B220⁺CD43⁻, B220⁺CD19⁺, and B220⁺CD25⁺ populations (Fig. 2 A and B). Bone marrow from induced Ott1^flox/null Mx1-Cre mice showed ≈10-fold fewer pre-B cells in vitro when plated in IL-7-containing methylcellulose medium (Fig. 2C). Single-colony PCR of the few pre-B colonies that did develop from Ott1^flox/null Mx1-Cre mice yielded exclusively unexcised Ott1^flox/null genotypes, suggesting a strong selective advantage for cells with intact Ott1 (data not shown). In contrast, Ott1^flox/wt Mx1-Cre pre-B colonies showed complete excision within the pre-B compartment, with 100% conversion of Ott1^flox into Ott1^null (data not shown). Excision of all bone marrow cells was >95% as assessed by Southern blot. Splenic excision, however, was incomplete, consistent with prior Mx1-Cre studies showing reduced efficiency within the spleen (SI Fig. 7A) (21). Taken together these data indicate that Ott1 is essential for progression beyond the pro/pre-B stage.

Fig. 2.

Ott1 deletion results in a block of pre-B lymphocyte differentiation. (A) Representative flow-cytometry plots of Ott1^null/flox Mx1-Cre and littermate control bone marrow 6 weeks after pIpC. (B) Absolute numbers of bone marrow cell subsets. B220/CD19, Ott1^null/flox Mx1-Cre, n = 5; control, n = 7. B220/CD43, Ott1^null/flox Mx1-Cre, n = 7; control, n = 6. CD25/B220, n = 3. (C) Pre-B colonies from 5 × 10⁴ bone marrow cells of mice 6 weeks after pIpC plated in duplicate in IL-7-containing methylcellulose. Gray bar, Ott1^null/flox Mx1-Cre (n = 4); white bar, littermate controls (n = 9).

Deletion of Ott1 Results in Megakaryocytic Expansion.

Spleen histopathology from Ott1^flox/null Mx1-Cre mice 6 weeks after pIpC treatment showed expansion of the splenic red pulp (Fig. 3A). Of particular interest, a ≈5-fold increase in megakaryocytes was observed in the red pulp (Fig. 3C). The megakaryocytes also showed abnormal clustering (Fig. 3A Inset), a feature associated with myelodysplasia (23). Likewise, an ≈2-fold increase in megakaryocytes was observed in Ott1^flox/null Mx1-Cre bone marrow (Fig. 3 B and C).

Fig. 3.

Megakaryocytic expansion in spleen and bone marrow after Ott1 deletion. Samples from Ott1^null/flox Mx1-Cre and control littermate mice 6 weeks after pIpC. Shown are photomicrographs of splenic (A) and femoral bone marrow (B) H&E-stained sections of Ott1^null/flox Mx1-Cre and control littermate mice. (Scale bars: 100 μm.) (C) Graph of megakaryocyte density in spleen and bone marrow histology sections. Megakaryocytes were scored by morphology in 25 random hpf (×600) of splenic red pulp (Ott1^null/flox Mx1-Cre, n = 3; control, n = 5) or bone marrow (n = 4 for both). (D) Number of CFU-Meg obtained from 1 × 10⁵ bone marrow cells plated in duplicate in MegaCult-C assay (Ott1^null/flox Mx1-Cre, n = 3; control, n = 5).

Bone marrow from Ott1^flox/null Mx1-Cre mice 6 weeks after pIpC treatment was plated in a collagen-based medium with IL3, IL6, IL11, and TPO to assess numbers of CFU-megakaryocyte (CFU-Meg). Unexpectedly, equivalent numbers of CFU-Megs were found in comparison to pIpC-treated littermate controls despite increased numbers of megakaryocytes observed in vivo (Fig. 3D). These data indicate that loss of Ott1 results in megakaryocyte expansion but suggest that there may be cell-nonautonomous contributions to this phenotype in vivo.

Myeloid Expansion After Ott1 Deletion.

Myeloid expansion was observed in the bone marrow of Ott1^flox/null Mx1-Cre mice after pIpC treatment (Fig. 3B). Flow cytometry of the spleen or bone marrow demonstrated a 100% increase and a 26% increase, respectively, in Mac-1⁺/Gr-1⁺ cells, consistent with histopathology (Fig. 4 A and B and SI Fig. 7B). No tissues other than the spleen showed extramedullary hematopoiesis (data not shown). Although the hematocrit did not differ between Ott1^flox/nullMx1-Cre and control animals (data not shown), there was an increase in erythroid precursors, marked by Ter119, in spleens derived from Ott1^flox/null Mx1-Cre animals and a comparable decrease in Ter119⁺ cells in Ott1^flox/null Mx1-Cre bone marrow (Fig. 4A).

Fig. 4.

Myeloid expansion and analysis of progenitors and LSK clonogenicity after Ott1 deletion. Shown are bone marrow or spleen samples from Ott1^null/flox Mx1-Cre and control littermate mice 6 weeks after pIpC. (A) Absolute numbers of Mac-1⁺/Gr-1⁺ and Ter119⁺ cells. Gray bars, Ott1^null/flox Mx1-Cre (n = 5); white bars, controls (n = 7). (B) Myeloid colony formation from duplicate platings of 2 × 10⁴ bone marrow cells in methylcellulose culture. Gray bars, Ott1^null/flox Mx1-Cre (n = 3); white bars, control (n = 7). (C) Representative flow-cytometry plots denoting the LSK, CLP, CMP, GMP, and MEP populations. Numbers are percentages of total bone marrow cells. (D) Absolute numbers of bone marrow progenitor cells. Gray bars, Ott1^null/flox Mx1-Cre (n = 5); white bars, control (n = 4). (E) Clonogenicity of sorted LSK cells. One hundred double-sorted LSK cells were plated onto methylcellulose medium plus growth factors in triplicate. Colony type was assessed by morphology after 7–10 days in culture (n = 3).

In vitro CFU assays of bone marrow were used to determine whether an increase in precursors could explain the myeloid expansion. Bone marrow from Ott1^flox/null Mx1-Cre mice 6 weeks after treatment with pIpC or littermate controls was plated in methylcellulose to assess myeloid and erythroid CFU potential (Fig. 4B). Absolute numbers of nucleated bone marrow cells did not significantly differ between the two groups (data not shown). A 60% increase in total colonies was observed by using the Ott1^flox/nullMx1-Cre bone marrow (Fig. 4B). The difference in colony number was attributable to increases in monocytic, granulocytic, and combined granulocyte-macrophage (GM) colonies but not erythroid or granulocyte, erythroid, monocyte, and megakaryocyte colonies.

Single-colony PCR of individual myeloid colonies showed 100% conversion of Ott1^flox/null genotype colonies to a homozygous Ott1 ^null/null genotype (SI Fig. 7C). These data indicate that Ott1 is not required for terminal myeloid differentiation in adult hematopoiesis and may inhibit myeloid development and/or proliferation.

Deletion of Ott1 Results in Increased Numbers of Lin⁻IL-7R⁻Sca-1⁺c-Kit⁺ (LSK) Cells and a Proclivity for Granulocytic Maturation.

Because Ott1 deletion affected multiple hematopoietic lineages, we studied the hematopoietic stem cell (HSC) and progenitor populations, including LSK, common lymphoid progenitor (CLP), common myeloid progenitor (CMP), megakaryocyte–erythroid progenitor (MEP), and granulocyte–monocyte progenitor (GMP) (24–26). Bone marrow from Ott1^flox/wt Mx1-Cre mice 6 weeks after pIpC treatment, and corresponding pIpC-treated littermates, was lineage-depleted (Lin⁻) and labeled with anti-Sca-1, c-Kit, IL-7R, and CD34 to separate the LSK/progenitor populations (Fig. 4C) (26).

The LSK compartment, which encompasses the HSC population, showed a 48% (P < 0.01) increase in Ott1^flox/null Mx1-Cre bone marrow (Fig. 4 C and D). There were no significant changes in the CLP, CMP, or MEP populations; however, numbers of GMPs rose by 50% (P = 0.01). These correlated with the colony-forming potential of a statistically significant increase in CFU-G/M/GM colonies from whole bone marrow that derive from GMPs (Fig. 4B).

To assess clonogenicity after Ott1 deletion, LSK cells from Ott1^flox/null Mx1-Cre mouse bone marrow 6 weeks after pIpC treatment were sorted onto methylcellulose containing IL-3, IL-6, SCF, Epo, Tpo, and GM-CSF. Equivalent colony numbers were obtained between the control group and the Ott1^flox/null Mx1-Cre group (data not shown). Single-colony PCR showed >80% excision efficiency in granulocytic, monocytic, erythroid, and megakaryocytic progeny (data not shown). These data imply that Ott1 does not impede the ability of LSK cells to differentiate and is not required for myeloid differentiation.

Interestingly, there was an increase in granulocytic differentiation from Ott1^flox/null Mx1-Cre LSK cells compared with controls (25.7 ± 3.4% vs. 6.0 ± 2.0%; P < 0.01) (Fig. 4E). The increase in granulocytic maturation occurred at the expense of combined granulocyte/monocyte colonies (51.5 ± 4.1% compared with 63.8 ± 6.9%; P < 0.01). These data suggest that Ott1 inhibits granulocytic maturation at the GMP stage or earlier. Megakaryocyte hyperplasia in vivo does not appear to be due to increased numbers of MEPs or a “steal” phenomenon from the erythroid lineage, because clonogenicity of the LSK cell toward erythroid and megakaryocytic cells is identical regardless of the presence of Ott1 (Fig. 4 B and E).

Discussion

Ott1 is essential for embryonic development beyond E9.5. Loss of a single allele resulted in expected numbers of viable embryos and normal phenotype indicating no gene–dosage effect in early development. In contrast, deletion of Mint in mice results in embryonic lethality occurring later around E13.5, signifying a nonredundant developmental function between Mint and Ott1 (17). Ott3 is more closely related to Ott1 than Mint, but these data show that Ott3 does not have functional redundancy with Ott1 in early development.

We examined the role of Ott1 in hematopoiesis using a conditional allele to bypass embryonic lethality. Ott1-deleted bone marrow has normal numbers of pro-B cells, but the later pre-B population identified by CD19, CD25, and loss of CD43 significantly declines, suggestive of a requirement for survival or proliferation. The defect is cell-autonomous in that Ott1-deleted bone marrow cannot form pre-B colonies in vitro. Although there was efficient excision in this compartment, rapid reconstitution of B lymphopoiesis with cells that contained an intact floxed allele underscores a marked selective advantage for B cells with intact Ott1.

The pro- to pre-B transition is regulated by several processes including cell fate commitment and maintenance using the transcription factors E2A, EBF, Pax5, and Lef1 (22, 27); Rag function and nonhomologous end joining pathways mediating successful V(D)J recombination (28, 29); and proliferation via signaling from the pre-BCR complex (30). Interference with any of the above pathways can result in a block similar to that observed after deletion of Ott1.

Ott1 could also be an antagonist of inhibitors of B cell development, such as Id or the Notch pathway (31, 32). Transgenic expression of Id1 and Id2 early lymphoid cells results in an increase of the bone marrow B220⁺CD43⁺ population but loss of maturing B220⁺CD43⁻ cells similar to loss of Ott1 (33). Notch1 is less likely as a target of Ott1 because overexpression of Notch1 in the lymphoid compartment results in an abundance of bone marrow T cells and a complete block before an identifiable B cell precursor (34).

In contrast to B cells, myeloid cells had no selection for unexcised cells and instead expanded within the spleen and bone marrow. Furthermore, there was an increase in the number of CFU-G/M/GM, and in vitro clonogenicity of Ott1-deleted HSCs demonstrated a preference for granulocytic maturation with a reciprocal loss of monocytic differentiation. Negative regulation of Id proteins by Ott1 could explain both the increase in granulocytic cell fate and pre-B cell block upon Ott1 loss (35). C/EBPα has also been shown to play a critical role in granulocytic vs. monocytic lineage decision (36, 37). In addition, Id1 is up-regulated by C/EBPα, and deletion of C/EBPα causes an increase in LSK/HSC repopulation that could explain Ott1 effects on pre-B and LSK cells, respectively (38). There was an increase in LSK cells in Ott1-deficient animals. Sorted LSK cells were capable of forming megakaryocytic, erythroid, and myeloid colonies in the absence of Ott1; therefore, Ott1 is dispensable for terminal differentiation.

Ott1 loss thus affects the LSK compartment, granulocytic, monocytic, megakaryocytic, and B lymphoid lineages. No known single pathway explains these pleiotropic effects in adult hematopoiesis. The extensive role of Ott1 in hematopoietic development is unique among spen family members, and study of interactions involving the RNA recognition motif and SPOC domains should yield new mechanistic insights.

In regard to the OTT1-MAL fusion in acute megakaryocytic leukemia, deletion of Ott1 results in an increased number of megakaryocytes in the bone marrow and spleen, suggesting that Ott1 is a negative regulator of megakaryocyte development. One hypothetical mechanism would be that the OTT1-MAL fusion contributes to the acute megakaryocytic leukemia phenotype by impairing OTT1 function possibly through a dominant-negative effect rather than solely through dysregulation of MAL pathways. The increase in megakaryocyte proliferation was not observed in vitro, suggesting a TPO-independent or cell-nonautonomous mechanism.

In summary, OTT1 plays an important role in hematopoietic development, with pleiotropic effects in the stem cell, lymphoid, and myeloid compartments. Further analysis of the interplay between Ott1 and other alleles involved in hematopoietic development are needed to improve our understanding of lineage determination and inform our understanding of the pathogenesis of leukemia initiated by the OTT1-MAL fusion gene.

Materials and Methods

Generation of Targeted Mice and Cre Deletion.

The targeting vector (SI Fig. 5) was electroporated into R-1 ES cells and injected into C57BL/6 blastocysts to produce chimeric animals that were bred for germ-line transmission. The floxed Ott1 allele was excised in the germ line by breeding with EIIa-Cre transgenic mice (The Jackson Laboratory, Bar Harbor, ME). Mice heterozygous for the null allele were bred with Mx1-Cre transgenic mice (The Jackson Laboratory) to generate Ott1^null/wt Mx1-Cre animals that were bred with Ott1^flox/wt or Ott1^flox/flox mice to produce Ott1^flox/null Mx1-Cre animals. Conditional excision in 4- to 6-week-old mice was achieved by i.p. injections of 250 μg of pIpC (Sigma–Aldrich, St. Louis, MO) every other day for three doses. Animals were maintained in microisolator cages, and experiments were performed with the approval of the Institutional Animal Care and Use Committee.

Southern and PCR Genotyping.

Genomic DNA from tail snips or ES cells (Puregene DNA isolation kit; Gentra Systems, Minneapolis, MN) was used to confirm a homologous recombination into the 5′ and 3′ ends of the Ott1 gene by PCR and Southern hybridization, respectively (SI Fig. 5). DNA from peripheral blood was obtained by using a Blood MiniAmp Kit (Qiagen, Valencia, CA). Single-colony DNA was generated by picking colonies into Eppendorf tubes, washing with PBS, resuspending into 50 μl of dH₂O, boiling 5 min, and then using 5 μl for PCR. Floxed or endogenous alleles were detected by using primers 5′-TGCTGTGACCAAGAGAGTTTGGC-3′ and 5′-CTTCTAAGACTGAGTAGAGAATG-3′, which span the 5′ loxp site, giving 250-bp and 216-bp products, respectively. The fully excised null allele was detected with the same forward primer and the reverse primer 5′-TTGTAACAAAGACAAGAGGGGAA-3′ based 3′ to the final loxp site and amplifying a 320-bp product.

Flow-Cytometric Analysis.

Single-cell suspensions of bone marrow and spleen were prepared as described (39) and frozen in 90% FCS and 10% dimethyl sulfoxide. Thawed cells were washed and blocked with 2.4G2 hybridoma supernatant (American Type Culture Collection, Manassas, VA). Flow reagents with the exception of allophycocyanin-CD4 (Caltag, Burlingame, CA) and phycoerythrin-streptavidin (Immunotech, Marseille, France) were purchased from BD Pharmingen (San Diego, CA). Analysis was performed on a four-color FACSCalibur cytometer (Becton Dickinson, Mountain View, CA) with CELLQUEST software, and a minimum of 10,000 events per experiment were acquired.

Blood and Tissue Analysis.

Blood was obtained through nonlethal eyebleeds under anesthesia with isoflurane per institutional guidelines. Complete blood counts were obtained with a Hemavet cell counter (Drew Scientific, Oxford, CT). Bone marrow was obtained from both femurs and tibias of each mouse. Organs from killed animals were preserved in PBS with 10% neutral, buffered formalin and then processed to H&E slides.

In Vitro CFU Assays.

Megakaryocyte CFUs were assayed by using MegaCult-C (StemCell Technologies, Vancouver, BC, Canada). Whole bone marrow was plated according to the manufacturer's instructions with 10 ng/ml rmIL-3, 20 ng/ml rmIL-6, 50 ng/ml rhIL-11, and 50 ng/ml rmTPO (StemCell Technologies). After 6 days, slides were dehydrated, fixed, and stained with acetylthiocholiniodide (Sigma–Aldrich).

Myeloid CFUs were quantitated by using Methocult M3434 (StemCell Technologies) methylcellulose medium containing 3 units per ml Epo, 10 ng/ml rmIL-3, 10 ng/ml rmIL-6, and 50 ng/ml rmSCF. Duplicate plates containing 2 × 10⁴ bone marrow cells were incubated at 37°C with 5% CO₂ for 10 days. Colonies were scored by morphology and verified by Wright–Giemsa stains of single-colony cytospins. Pre-B colony assays were performed by plating bone marrow cells in Methocult M3630 (StemCell Technologies) that contains IL-7 and scoring after 6 days of incubation.

Progenitor Analysis.

Bone marrow cells were lineage-depleted and stained for progenitor analysis as described (25, 40) and then analyzed on a modified double-laser FACS (Moflo-MLS; Cytomation, Fort Collins, CO). Cell populations were identified as follows: CMP, Lin⁻Sca-1⁻c-Kit⁺CD34⁺FcγRII/III^lo; GMP, Lin⁻Sca-1⁻c-Kit⁺CD34⁺FcγRII/III^hi; MEP, Lin⁻Sca-1⁻c-Kit⁺CD34⁻FcγRII/III^lo; HSC, Lin⁻Sca-1⁺c-Kit⁺; and CLP, Lin⁻Sca-1^loc-Kit^loIL-7R⁺.

LSK clonogenicity was determined by double-sorting 100 LSK directly onto plates containing Methocult M3234 (StemCell Technologies) supplemented with 3 units per ml Epo, 10 ng/ml rmIL-3, 10 ng/ml rmIL-6, 50 ng/ml rmSCF, 10 ng/ml rhIL-11, 10 ng/ml rmGM-CSF, and 10 ng/ml rmTPO (StemCell Technologies). Triplicate plates were incubated as above for 7–10 days and scored by morphology.

Statistics.

P values were calculated by using a two-tailed Student's t test, and error bars represent one SD.

Abbreviations

CLP

common lymphoid progenitor

CMP

common myeloid progenitor

GMP

granulocyte–monocyte progenitor

HSC

hematopoietic stem cell

LSK

Lin⁻IL-7R⁻Sca-1⁺c-Kit⁺

MEP

megakaryocyte–erythroid progenitor

pIpC

polyinosinic-polycytidylic acid

En

embryonic day n.