Regulation of hematopoietic stem cells by their mature progeny

Abstract

Thrombopoietin (TPO), acting through its receptor Mpl, has two major physiological roles: ensuring production of sufficient platelets via stimulation of megakaryocyte production and maintaining hematopoietic stem cell (HSC) quiescence. Mpl also controls circulating TPO concentration via receptor-mediated internalization and degradation. Here, we demonstrate that the megakaryocytosis and increased platelet mass in mice with mutations in the Myb or p300 genes causes reduced circulating TPO concentration and TPO starvation of the stem-cell compartment, which is exacerbated because these cells additionally exhibit impaired responsiveness to TPO. HSCs from Myb^Plt4/Plt4 mice show altered expression of TPO-responsive genes and, like HSCs from Tpo and Mpl mutant mice, exhibit increased cycling and a decline in the number of HSCs with age. These studies suggest that disorders of platelet number can have profound effects on the HSC compartment via effects on the feedback regulation of circulating TPO concentration.

The control of circulating platelet number is critical to health. Deregulated platelet production accompanies some leukemias and high platelet counts characterize many myeloproliferative disorders. Conversely, low platelet count, or thrombocytopenia, can occur as the result of autoimmune disorders or hematological diseases such as idiopathic thrombocytopenia purpura and is often a side effect of cytotoxic cancer treatments, where episodes of bleeding can put patients at risk as well as interrupt and compromise therapy (1). Thrombopoietin (TPO) is the primary regulator of platelet production. Loss-of-function mutations in the genes for Tpo or its receptor, Mpl, result in a severe reduction in platelets, megakaryocytes, and megakaryocyte progenitor cells (2, 3). In mice, reduced TPO signaling also leads to increased hematopoietic stem cell (HSC) cycling, reduced HSC function in transplantation assays and, over time, a deficit in HSC numbers (4–7). Likewise, in humans, mutations in Mpl result in congenital amegakaryocytic thrombocytopenia and, eventually, aplastic anemia (8). Taken together, these observations highlight the key role for TPO signaling in control of platelet number and regulation of HSCs.

TPO is produced primarily in the liver (9) and Mpl is expressed on platelets, megakaryocytes and their progenitors, and within the HSC compartment, including on cells with long-term hematopoietic reconstituting capacity (6, 10, 11). Several lines of evidence support a model in which TPO production is constant and that the concentration of available circulating TPO is controlled by the mass of Mpl receptors available to internalize the cytokine. For example, TPO transcription remains relatively constant during periods of thrombocytopenia or thrombocytosis (12–14), and serum TPO levels correlate with platelet and megakaryocyte mass rather than absolute platelet count (15, 16). In healthy individuals, this system provides an effective feedback mechanism in which the availability of TPO reflects platelet need.

The transcription factor Myb and its coregulator p300 have diverse roles in hematopoietic regulation. Knockout mouse studies have demonstrated an absolute requirement for Myb/p300 in hematopoiesis and a series of elegant conditional mutant studies in mice have defined key roles for Myb in progression of erythropoiesis and lymphopoiesis (17–19). In contrast, Myb/p300 appears to play a repressive role in megakaryocytopoiesis. We recently identified the hypomorphic alleles Myb^Plt3, Myb^Plt4, and p300^Plt6, which each cause a marked thrombocytosis (20, 21). This phenotype has also been observed in independently derived mutant alleles of Myb and p300 (22–24) and appears to arise from excessive production of megakaryocytes and their progenitor cells from multipotent precursors (25).

In this article, we show that mice with mutations in Myb/p300 not only have a low circulating TPO level, consistent with the thrombocytosis characteristic of these mutants, but that Myb^Plt4/Plt4 stem cells also express subnormal amounts of Mpl and respond poorly to TPO. Consistent with reduced TPO signaling, Myb/p300 mutant HSCs exhibit altered expression of TPO-responsive genes, are more actively cycling than normal stem cells in vivo, and become depleted with age, phenotypes that are ameliorated by transgenic overexpression of TPO. Our data highlight a unique link between platelet number and the regulation of HSCs and suggests that unrestrained elevation in platelet production may lead to significant perturbations of the stem-cell compartment via TPO depletion.

Results

Myb/p300 Deficiency-Induced Thrombocytosis Leads to a Reduction of Serum TPO Levels.

To explore the relationship between platelet number and circulating TPO level, serum TPO concentrations and platelet counts were measured in Myb^Plt4/Plt4 mice on either wild-type or Mpl^−/− genetic backgrounds and compared with control cohorts of wild-type, Mpl^−/−, and Tpo transgenic (Tpo^Tg) mice (Fig. 1 A and B). As expected, TPO levels were high in Mpl^−/− mice and unaffected by the thrombocytosis in Mpl^−/− Myb^Plt4/Plt4 mutants. This result is because of the inability to actively clear circulating TPO in the absence of Mpl expression. In Tpo^Tg mice, which constitutively express TPO from a liver-specific transgene (SI Materials and Methods), TPO production was elevated beyond the consumption ability of even the large numbers of platelets found in these mice, resulting in a high circulating concentration. In Mpl^+/+ Myb^Plt4/Plt4 mice, TPO levels were approximately half that observed in wild-type mice, consistent with the marked thrombocytosis typical of Myb mutant mice driving increased TPO clearance. A similar result was observed for irradiated mice reconstituted with stem cells from p300^Plt6/Plt6 mice, which are also thrombocytotic (21). This finding reinforces the TPO-independent nature of the expanded thrombopoiesis in Myb and p300 mutant mice and demonstrates the capacity of excessive numbers of platelets to consume circulating TPO.

Fig. 1.

Deficiencies in the TPO/Mpl pathway in Myb^Plt4/Plt4 mice. (A) Platelet counts and (B) circulating serum TPO levels in 8- to 12-wk-old mice of each genotype; TP indicates data from mice 16 wk after bone marrow transplant with bone-marrow donors as indicated. n = 6–21 mice per genotype. (C) Mpl expression measured by quantitative real time RT-PCR in cell populations purified by flow cytometry. Data shown are expression of Mpl relative to Hprt expression, measured by the ΔΔCt method. n = 5–8 mice per genotype. (D) Thymidine incorporation in lineage-negative (Lin⁻) cells from mice of the indicated genotypes after 4 d of culture in TPO. n = 5 for Mpl^+/+, n = 2 for Mpl^−/− mice. Data are shown as means ± SDs. ***P < 0.001; **P < 0.01; *P < 0.05; and ns, not significant, for comparisons between Mpl^+/+ Myb^+/+ and Mpl^+/+ Myb^Plt4/Plt4 data.

Myb^Plt4/Plt4 Cells Respond Poorly to TPO.

In addition to controlling megakaryocyte and platelet production, TPO also has a role in HSC regulation (4–7). Because thrombocytosis in Myb^Plt4/Plt4 mice resulted in a significant reduction in circulating TPO, we examined TPO-dependent activities in Myb^Plt4/Plt4 HSCs. The lineage (Lin)⁻ Sca-1⁺ c-Kit⁺ (LSK) compartment of the bone marrow has been shown to contain all of the HSCs capable of hematopoietic reconstitution (26), but this population is heterogeneous and can be divided by cell-surface markers into subsets with differing repopulation potential. Quantitative RT-PCR analysis of phenotypically defined LSK CD34⁻ Flt3⁻ long term (LT)-HSCs, LSK CD34⁺ Flt3⁻ short term (ST)-HSCs, and LSK CD34⁺ Flt3⁺ multipotent progenitors (MPPs) (27, 28) revealed that Mpl expression was reduced in all three subsets of LSK cells in Myb^Plt4/Plt4 mice compared with that in wild-type cells (Fig. 1C).

We next compared responses to TPO of Lin⁻ cells collected from Myb^Plt4/Plt4 and wild-type mice. We found significantly less proliferation of Myb^Plt4/Plt4 Lin⁻ cells compared with their wild-type counterparts at each of two concentrations of TPO analyzed (Fig. 1D). As a control, we confirmed that both Mpl^−/− Myb^+/+ and Mpl^−/− Myb^Plt4/Plt4 Lin⁻ cells were unable to respond to TPO. This finding shows that the TPO/Mpl pathway is highly perturbed in Myb^Plt4/Plt4 mice: Myb^Plt4/Plt4 LSKs have reduced Mpl expression, are exposed to lower TPO concentrations, and respond poorly to TPO in vitro.

Composition of the LSK Populations in Myb^Plt4/Plt4 Mice.

Flow cytometric analysis revealed that Mpl^+/+ Myb^Plt4/Plt4 femurs had five times the number of LSK cells than observed in wild-type controls (Fig. 2), with no significant change in overall cellularity. The overall increase in LSK number was primarily because of increased ST-HSCs, which were 7.5 times more numerous in Mpl^+/+ Myb^Plt4/Plt4 bone marrow than Mpl^+/+ Myb^+/+ bone marrow (Fig. 2); however, there were also four times as many MPPs in Myb^Plt4/Plt4 bone marrow. No significant alteration in the number of LT-HSCs in Mpl^+/+ Myb^Plt4/Plt4 mice was observed compared with wild type, but as in Mpl^−/− mice, they constituted a reduced proportion of the LSK compartment. As expected from previous studies (5), the numbers of LSKs and LSK subsets were significantly decreased in Mpl^−/− bone marrow compared with wild type. The effect of Myb mutation on these populations was not influenced by the presence or absence of Mpl (Fig. 2).

Fig. 2.

Myb^Plt4/Plt4 mice have more prospective ST-HSCs and MPPs. (A) Representative flow cytometric analyses of viable Lin⁻ cells (Upper) and LSK cells (Lower) from bone marrow of mice of the indicated genotypes. The gates used to distinguish prospective LT-HSCs (CD34⁻ Flt3⁻ LSK), ST-HSCs (CD34⁺ Flt3⁻ LSK), and MPPs (CD34⁺ Flt3⁺ LSK) are shown. Percentages of gated cell populations within the parent populations are indicated. (B) Graphs show mean ± SD of the number per femur of total LSKs and LSK subsets. n = 5–6 mice per genotype. ***P < 0.001; ns, non significant.

Mpl^−/− and Myb/p300-Deficient LSKs Share a Common Set of Gene-Expression Changes.

To explore gene-expression changes in the stem-cell compartment, we collected LSK cells from wild-type, Mpl^−/−, Myb^Plt4/Plt4, and TPO^Tg mice, as well as mice reconstituted with p300^Plt6/Plt6 bone marrow (and as a control, from mice reconstituted with p300^+/+ bone marrow), and isolated RNA that was then hybridized to Illumina Mouse-WG microarrays. Although there were many genes that were differentially expressed when comparing LSKs from each mutant to LSKs from wild-type mice (380 for Myb^Plt4/Plt4 vs. Myb^+/+; 333 for p300^Plt6/Plt6 vs. p300^+/+; 446 for Mpl^−/− vs. Mpl^+/+), we focused on genes for which expression was consistently deregulated in all of the mutants. We found 137 probes that were significantly up- or down-regulated in each of Myb^Plt4/Plt4, p300^Plt6/Plt6, and Mpl^−/− LSKs compared with wild type (Fig. 3A and Dataset S1). These genes, the vast majority of which were down-regulated in mutant LSKs, tended to be up-regulated in cells from TPO^Tg mice. Thus, the changes in expression of this gene set in Myb^Plt4/Plt4 and p300^Plt6/Plt6 LSK cells is likely because of reduced TPO exposure. The data imply that normal expression of these genes not only requires TPO, but that they are inducible by excess TPO. We have termed this gene set the “TPO-response signature.”

Fig. 3.

Common gene-expression signature in Myb/p300 deficient and Mpl^−/− LSK cells. (A) Heat map visualization of genes that are differentially expressed in Myb^Plt4/Plt4, p300^Plt6/Plt6, and Mpl^−/− LSK cells in the same direction (either up or down) compared with wild type (P < 0.05). Genes are distributed horizontally and grouped vertically according to genotype. The scale represents log₂-fold change differences for each mutant compared with wild type. (B) Quantitative real-time RT-PCR of gene expression in MPPs (black bars), ST-HSCs (dark gray bars), and LT-HSCs (light gray bars) purified by flow cytometry from Myb^+/+ mice (^+/+) and Myb^Plt4/Plt4 mice (Plt4/Plt4). Expression is presented as relative to that in Myb^+/+ ST-HCS, set as 1. Data, expressed as mean and SD, represent one experiment with two to five independent cDNA samples. CD34 and Flt3 were used to define the populations, but not shown to be differentially expressed on the microarrays. All other genes were identified as down-regulated by microarray analysis. Genes that were unable to be detected as expressed in a particular subset are marked with #. ***P < 0.001; *P < 0.05; ns, not significant in comparison with Myb^+/+ to Myb^Plt4/Plt4. (C) Hematopoietic expression pattern of genes selected as down-regulated in Myb^Plt4/Plt4, p300^Plt6/Plt6 and Mpl^−/− LSK cells. The scale represents normalized expression across the cell types for each gene. Cell types are murine, as defined in Chambers et al. (33), and sourced from http://franklin.imgen.bcm.tmc.edu/loligag/, where SP HSC is side population LSK, N Ery, nucleated erythrocyte; Gran, granulocytes; Mono, monocytes; T CD8 A and T CD4 A, activated T cell; T CD8 N and T CD4 N, naive T cell; NK Cell, natural killer cell.

Subsets of the LSK compartment, LT-HSCs, ST-HSCs, and MPPs, are known to exhibit differential gene expression (29). Of five TPO-response signature genes tested, Col4a1, Esam1, Ryk, Socs2, and Zfp532, each showed reduction in expression when tested by RT-PCR in Myb^Plt4/Plt4 ST-HSCs and MPPs compared with wild-type cells. In LT-HSCs, two of these genes, although detected in Myb^+/+ cells, were no longer expressed in Myb^Plt4/Plt4 cells, and two genes were modestly, but not significantly, reduced in mutant cells (Fig. 3B). The majority of the genes in the TPO-response signature displayed intermediate expression in Myb^Plt4/+ LSK cells (Fig. 3A), in which the relative frequency of LT-HSCs, ST-HSCs, and MPPs is normal (Fig. 2). These data suggest that genes in the TPO-response signature were detected as differentially expressed as a result of expression changes across the mutant LSK subpopulations, rather than selective expression in a disproportionately represented cell subset. Although a number of targets of Myb have been identified, with diverse roles including differentiation, cell survival, and proliferation, none of the genes in the TPO-response signature are known to be direct targets of Myb (30–32). One member of the TPO-response signature, Cdkn1c, was reported to be up-regulated in human CD34⁺ cells following Myb expression knockdown (32), but has been independently reported to be inducible by TPO (7). The down-regulation of Cdkn1c in Myb-mutant LSK cells implies that the influence of TPO on expression may outweigh any effects of Myb activity. Other genes within the TPO-response signature have not previously been associated with TPO signaling.

Next, we investigated the expression of genes in the TPO-response signature in a publicly available collection of microarray data (33) from diverse hematopoietic cell types (Fig. 3C). This collection included the “side population” (SP) LSKs, which are highly enriched for HSCs. Our analysis revealed that the TPO-response signature is most highly expressed in the SP-LSKs and down-regulated as cells mature. Gene sets derived from gene ontology terms (34) sourced from MSigDB (35) were tested with Fisher's exact test to see if any were over represented in the down-regulated set of genes. We found that genes from the “Hematopoietic Stem Cell” [up-regulated in mouse HSCs compared with differentiated brain and bone marrow cells (36)] and “Cell Proliferation” (34) gene sets were overrepresented among the genes down-regulated in the TPO response signature (Table 1). This finding implies that the TPO-response signature genes are associated with proliferation, are enriched for those that are down-regulated as stem cells mature, and that these genes have specificity to the hematopoietic system.

Table 1.

Gene sets over-represented in the TPO-response signature

Category	Genes
Cell proliferation	Mpl, Ephb4, Fscn1, Ndn, Cdkn1c, Serpinf1, Dlg5, Ebi3, Maged1
Hematopoietic stem cell	Irf6, Traf1, Tcf15, Trib3, Tgm2, Emcn, Pkia, Efna1, Il11ra1, Galt, Tie1, Ppap2b, Mpl, Gjb3, Foxa3, Zfp30, Tead2, Cdkn1c, Phlda2, Vegfc, Col4a1, Col4a2, Cx3cl1, Ryk, Socs2, Serpinf1, Chrnb1, Aldoc, Ptrf, Ndrg2, Mapk12, Ltb, Ebi3, Ctsf, Fhl1, Bgn, Maged1

Gene sets from MSigDB (35).

Increased Cycling of Myb^Plt4/Plt4 LSK Cells.

To explore cell cycle status in Myb^Plt4/Plt4 LSKs, we used a combination of DAPI to measure cell DNA content and an antibody specific for K_i-67, a nuclear antigen associated with proliferation (Fig. 4A). We found that a decreased proportion of Mpl^−/− and Myb^Plt4/Plt4 LSKs were in the quiescent G0 phase, compared with wild-type mice and, reciprocally, that there was a modest increase in the proportion of LSKs in G0 in TPO^Tg mice (Fig. 4B). In contrast, the Lin⁻ Sca-1⁻ c-Kit⁺ fraction of the bone marrow, which only expresses low levels of Mpl (Fig. 4C), showed no difference in the number of cells that were in cycle for any of the genotypes examined (Fig. 4D).

Fig. 4.

Cycling is enhanced in Myb^Plt4/Plt4 cells. (A) Representative flow cytometric analyses of bone marrow from wild-type, Mpl^−/−, Myb^Plt4/Plt4, and TPO^Tg mice showing LSK and Lin-S-K⁺ gates (Upper) and K_i-67/DAPI staining profiles for LSK cells (Lower). (B) Mean percentages of LSK cells that are in G0, G1, or S-M phase of the cell cycle. n = 4–8 mice of each genotype. (C) Relative expression of Mpl in hematopoietic progenitor populations as determined by qRT-PCR. Expression is relative to Hprt expression as determined by the ΔΔCt method. n = 3 independent samples per genotype. (D) Mean percentages of Lin⁻ Sca-1⁻ c-Kit⁺ cells that are in G0, G1, or S-M phase. n = 6–10 mice of each genotype. Data shown are mean ± SD. ***P < 0.001; ns, not significant.

Decline in HSC Number in Myb^Plt4/Plt4 Mice with Age.

TPO regulates HSC quiescence and maintenance of HSC number throughout life (5, 7). In wild-type mice, LSK cell number did not vary greatly between 7 wk and 13 mo of age; however, in Myb^Plt4/Plt4 mice there was a marked expansion in LSK number that peaked at ≈9 mo of age and subsequently declined between 9 and 13 mo (Fig. 5A). Although the wild-type LT-HSC compartment showed a fivefold increase in numbers from 7 wk to 13 mo of age, no significant increase was observed in Myb^Plt4/Plt4 mice (Fig. 5B). Unlike wild-type bone marrow, in which progenitor cell numbers increased from 7 wk to 13 mo of age, in Myb^Plt4/Plt4 marrow, progenitor cell numbers declined modestly with age (Fig. 5C). These observations suggest that increased cycling of LT-HSCs in Myb mutant mice leads to a transitory expansion of the ST-HSC and MPP compartments at the expense of long-term maintenance of the LT-HSC compartment. Over time, this process leads to reduced ability to maintain the hematopoietic system.

Fig. 5.

Age-related deficits in Myb^Plt4/Plt4 bone marrow. Counts of (A) LSK cells per femur and (B) LT-HSCs per femur from Myb^Plt4/Plt4 and Myb^+/+ mice at 7 wk, 9 mo, and 13 mo of age. (C) Ratio of the number of bone marrow-derived colony-forming cells in Myb^Plt4/Plt4 and Myb^+/+ mice at 13 mo vs. 7 to 10 wk of age. Bone-marrow cells were cultured in SCF/IL3/EPO. Meg, megakaryocyte; G/M/Eo, Granulocyte/Macrophage/Eosinophil. Data shown are mean ± SD. n = 3–9 mice per genotype and age. ***P < 0.001; **P < 0.01; *P < 0.05; ns, not significant.

TPO Rescues Loss of Quiescence in Myb^Plt4/Plt4 Mice.

To confirm that the loss of HSC quiescence observed in Myb^Plt4/Plt4 mice was a result of TPO starvation, we introduced the TPO transgene onto the Myb^Plt4/Plt4 background, which exacerbated thrombocytosis above that observed in either the Myb^Plt4/Plt4 or the TPO^Tg mice (Fig. S1A). Despite the marked thrombocytosis in these mice, and hence increased TPO clearance, TPO production from the transgene was sufficient to ensure that the cytokine remained at supraphysiological levels in the serum (Fig. S1B). As described above, Myb^Plt4/Plt4 LSKs showed a decrease in quiescence, with fewer cells in G0 than evident in wild-type populations. TPO^Tg LSKs had a similar cell cycle profile to the wild-type mice, suggesting that additional TPO over that in wild-type mice does not influence HSC cycling. On a Myb^Plt4/Plt4 background, the additional TPO produced from the transgene resulted in partial rescue of quiescence, with more LSK cells in G0 and fewer in G1, compared with Myb^Plt4/Plt4 mice (Fig. 4B). Additionally, the number of LT-HSCs in compound Myb^Plt4/Plt4 TPO transgenic mice was increased compared with that observed in Myb^Plt4/Plt4 mice (Fig. S1C). These data support a model in which reduced circulating TPO contributes significantly to the HSC defects observed in Myb^Plt4/Plt4 mice.

Discussion

Production of platelets is regulated by a negative feedback loop in which the concentration of free TPO in the circulation is controlled by the rate of its internalization via Mpl receptors expressed on the surface of megakaryocytes and platelets. Under thrombocytopenic conditions, this process leads to an accumulation of TPO and greater production of megakaryocytes via stimulation of megakaryocyte progenitors in the bone marrow. Reciprocally, in situations of thrombocytosis there are reduced TPO levels and hence a reduction in megakaryocyte production. However, TPO not only has effects on cells committed to the megakaryocyte lineage, but is also required to maintain the appropriate balance of quiescence and activity in HSCs (5, 7). Thus, fluctuations in circulating TPO concentration are likely to effect stem cell behavior.

As expected, thrombocytosis driven by Myb or p300 mutations resulted in low circulating TPO concentrations. The likelihood that HSCs in these mice experience TPO deprivation was further supported by the observation that Mpl expression on Myb^Plt4/Plt4 LSK cells was reduced. Previous studies have noted reduced Mpl expression on platelets in mice administered with TPO, concluding that TPO has little or no effect on Mpl transcription or translation but that excess cytokine can stimulate greater receptor catabolism (37). In contrast, our data suggest that in the chronic low-TPO environment of Myb^Plt4/Plt4 mice, HSCs express reduced Mpl RNA. This finding may reflect differences in regulation of receptors in stem cells and platelets, or might reflect an intrinsic contribution of Myb deficiency to altered regulation of Mpl expression. In any event, reduced receptor expression on stem cells from Myb^Plt4/Plt4 mice appeared functionally significant because Myb^Plt4/Plt4 LSK cells proliferated poorly in response to TPO in vitro. Myb^Plt4/Plt4 stem cells also showed multiple signs of TPO deprivation in vivo. Transcriptional profiling revealed changes in gene expression in Myb^Plt4/Plt4 LSK cells that were shared with cells from Mpl^−/− mice and reciprocally regulated in mice exposed to excess TPO. These studies allowed definition of a unique TPO-response signature, a pattern of gene expression reflecting the degree of stimulation of hematopoietic stem cells by TPO. The majority of the genes within this TPO-response signature have not previously been implicated in TPO signaling; however, there was an overrepresentation of genes previously associated with HSCs, and a preponderance of proliferation-associated genes.

Studies in mice lacking TPO or Mpl have established that in the absence of TPO signaling, a greater than normal proportion of HSCs are in cycle at steady state (5, 7). Increased cycling characterizes HSCs in Myb^Plt4/Plt4 mice and was also observed in Myb^M303V/M303V mice (24). Increased HSC cycling is consistent with the fact that many of the cell proliferation-associated genes that were down-regulated in the TPO-response signature are cell cycle related; our data suggest that genes such as Ndn (38), Cdkn1c (39), and Maged1 (40) are important mediators of the balance between HSC quiescence and activity regulated by TPO. In support of this regulatory pathway, changes in HSC Cdkn1c expression have independently been associated with administration of TPO in mice (7). Consistent with TPO deprivation playing a key role in the HSC phenotype in Myb^Plt4/Plt4 mice, transgenic supplementation of TPO expression resulted in amelioration of the cell cycle deregulation characteristic of Myb^Plt4/Plt4 LSK cells. Also consistent with observations in TPO-deficient mice (41), the age-related increase in the number of phenotypically defined LT-HSCs typical of wild-type animals was absent in Myb^Plt4/Plt4 mice. The decline in the numbers of progenitor cells in aging Myb^Plt4/Plt4 mice is consistent with reduced stem cell activity, but the precise changes in functional capacity of Myb^Plt4/Plt4 stem cells with age requires further studies.

We conclude that the loss of stem-cell quiescence phenotype associated with Myb mutation arises primarily because of a feedback loop in which platelet numbers, through regulation of available TPO levels, regulates the entry of HSCs into cycle. Such a feedback mechanism may provide protection in times of hematopoietic stress associate with bleeding. Platelets will be lost directly following an episode of bleeding and indirectly as a result of concomitant vessel repair. The resulting increase in the levels of TPO will drive committed progenitor cells to produce megakaryocytes and platelets to replace those lost. Because emergency demand on the hematopoietic system runs the risk of exhausting the stem-cell compartment and compromising future hematopoietic production, the action of TPO to promote the quiescence of the most primitive stem cells may represent the mechanism by which acute demand for new blood cells can be balanced against future requirement.

It is clear from these experiments, and that of others, that deficits in the Myb/p300 complex mediate other changes to stem cell function that are TPO independent, such as the expansion of the ST-HSC population and bias in lineage commitment. The complexity of hematopoietic regulation makes it difficult to tease out which of these phenotypes are driven by direct functions of the Myb/p300 complex, and which result from feedback loops. The data presented here demonstrates that the enhanced cycling of stem cells is likely to be largely indirect and to arise from a unique feedback loop in which platelet number regulates the behavior of stem cells via control of circulating TPO concentration. To isolate the direct effects of Myb/p300 loss on the function of HSCs and progenitors from those caused by the bone-marrow environment, it may be useful to perform reciprocal transplant experiments using mice and cells deficient for Myb, Mpl, and TPO.

Materials and Methods

Mice and Hematological Analysis.

Mice carrying the Myb, p300, and Mpl alleles were derived as previously described (3, 20, 21). TPO transgenic mice were created as described in SI Materials and Methods. All animal experiments were approved by the Walter and Eliza Hall Institute of Medical Research Animal Ethics Committee.

Manual or automated blood cell counts were performed on blood collected from the retro-orbital plexus into sample tubes coated with EDTA (Sarstedt). Mice with hematopoiesis derived from engrafted p300^Plt6/Plt6 stem cells were generated as previously described (21). Clonal analysis of bone marrow cells in semisolid agar cultures were performed as previously described (42) using a combined stimulus of 100 ng/mL SCF, 10 ng/mL IL-3 (both prepared in our laboratories), and 4 IU/mL EPO (Janssen-Cilag).

For serum TPO measurement, mice were bled by cardiac puncture, the blood was allowed to clot at room temperature, and serum was separated by centrifugation. Commercial monoclonal antibody-based ELISA test kits were used to detect TPO (Quantikine M mouse TPO, MTP00; R&D Systems).

Proliferation Assays.

Twenty-thousand flow-sorted Lin⁻ cells were collected into DMEM with 10% FCS with recombinant TPO and incubated at 37 °C in 5% CO₂ in air. Cells were cultured for 4 d before they were pulsed with 1 μCi (0.037 MBq) [3H]-thymidine for 16 h, harvested onto Inotech glass filters, and counted in a TopCount NXT Microplate Scintillation Counter (Packard).

Cell Cycle Analysis.

Bone-marrow suspensions were enriched for mononuclear cells by centrifugation over a Ficoll-Paque gradient. Cells were then incubated with biotinylated CD4, CD8, B220, Gr-1 and Ter119 antibodies (Pharmingen), before staining with fluorochrome-conjugated Sca-1, c-Kit, and streptavidin (Pharmingen). Cells were made permeable by incubation with Cytofix/Cytoperm solution (BD Biosciences) on ice before washing with Perm/Wash solution (Pharmingen) and incubating with K_i-67 antibodies for 18 h at 4 °C. DAPI (4',6-diamidino-2-phenylindole, 5 mg/mL) was then added to each sample for room temperature incubation for 30 min. Cells were then resuspended and filtered before flow cytometric analysis.

Gene Expression Analysis.

RNA for microarrays was extracted from LSK cells using MicroRNeasy (Qiagen) purification columns according to the manufacturer's protocol, then labeled, amplified, and hybridized to Illumina MouseWG-6 V1.1 or V2.0 Expression BeadChips according to Illumina standard protocols at the Australian Genome Research Facility, Melbourne. Each sample was derived from bone-marrow LSK cells from at least six donor mice. Data were analyzed in R and subjected to a variance stabilizing transformation and quantile normalization (43). Linear modeling using an empirical Bayes approach was applied to the data (44). Data were corrected for multiple testing using Benjamini and Hochberg correction. Microarray data are available at Array Express (www.ebi.ac.uk/arrayexpress/) under accession number E-TABM-1050. The Myb^Plt4/Plt4 and Myb^+/+ microarrays data were those previously described (19). Methods for real-time quantitative RT-PCR were described in SI Materials and Methods.

Statistical Tests.

The two-sided Student t test was used for all statistical analyses of two means. One-way ANOVA was used when three or more means were compared. P values were adjusted for multiple testing by the Holm-Bonferroni correction, unless otherwise specified (45).