Roles for c-Myc in self-renewal of hematopoietic stem cells.

Notch and HOXB4 have been reported to expand hematopoietic stem cells (HSCs) in vitro. However, their critical effector molecules remain undetermined. We found that the expression of c-myc, cyclin D2, cyclin D3, cyclin E, and E2F1 was induced or enhanced during Notch1- or HOXB4-induced self-renewal of murine HSCs. Since c-Myc can act as a primary regulator of G(1)/S transition, we examined whether c-Myc alone can induce self-renewal of HSCs. In culture with stem cell factor, FLT3 ligand, and IL-6, a 4-hydroxytamoxifen-inducible form of c-Myc (Myc/ERT) enabled murine Lin(-)Sca-1(+) HSCs to proliferate with the surface phenotype compatible with HSCs for more than 28 days. c-Myc activated by 4-hydroxytamoxifen augmented telomerase activities and increased the number of CFU-Mix about 2-fold in colony assays. Also, in reconstitution assays, HSCs expanded by c-Myc could reconstitute hematopoiesis for more than 6 months. As for the mechanism of c-myc induction by Notch1, we found that activated forms of Notch1 (NotchIC) and its downstream effector recombination signal-binding protein-J kappa (RBP-VP16) can activate the c-myc promoter through the element between -195 bp and -161 bp by inducing the DNA-binding complex. Together, these results suggest that c-Myc can support self-renewal of HSCs as a downstream mediator of Notch and HOXB4.

Hematopoietic stem cells (HSCs) 1 are characterized by two distinct abilities: self-renewal ability and pluripotentiality. With these activities, HSCs supply all lineages of hematopoietic cells throughout their life. In a murine experimental model, only a single HSC with the CD34 low/Ϫ c-Kit ϩ Sca-1 ϩ phenotype can reconstitute whole hematopoiesis in vivo (1,2). To maintain homeostasis of hematopoiesis and protect exhaustion of HSC population, most of the HSCs are kept quiescent, and only a limited number of cells enter the cell cycle to supply mature blood cells (3). As external regulatory factors, various cytokines such as stem cell factor (SCF), Flt3 ligand (FL), thrombopoietin, IL-3, and IL-6 promote the growth of HSCs (4 -6). The combination of SCF, FL, thrombopoietin, and IL-6, in particular, was reported to efficiently induce the in vitro cell division with characteristics of self-renewal in HSCs (7). In contrast, transforming growth factor-␤ is known to inhibit the growth of HSCs (8,9). Besides these cytokines, the direct interactions with stromal cells and extracellular matrix in the bone marrow microenvironment also influence the fates of HSCs (10,11). As one mechanism responsible for this effect, Bernstein and co-workers (12) showed that the activation of Notch transmembrane receptors expressed on HSCs by their ligand (Jagged 1 or Jagged2) on stromal cells promotes the self-renewal of HSCs.
Furthermore, accumulated evidence indicated that the cell cycle state of HSCs is regulated by intrinsic transcription factors such as c-Myb and GATA-2. c-Myb promotes the growth of HSCs, and c-Myb-deficient mice die at embryonic day 15.5 due to the defect of definitive hematopoiesis (13). Similarly, GATA-2 Ϫ/Ϫ mice are embryonic lethal around embryonic day 11.5 because of the defect in the development and/or maintenance of HSCs (14), whereas functional roles of GATA-2 in the growth of HSCs are still controversial (15,16). Also, HOXB4, a member of Hox family of transcription factors, was reported to induce the marked in vitro expansion of HSCs (17,18).
During the last decade, a number of cell cycle regulatory molecules such as cyclins, cyclin-dependent kinases (CDKs), and CDK inhibitors have been identified and characterized in various types of cells. As for the roles of these molecules in primitive hematopoietic cells, p21 WAF1 and p27 Kip1 keep the quiescence of HSCs and of progenitor cells, respectively, thereby governing their pool sizes (19,20). Meanwhile, inactivation and/or deletion of p16 INK4A and p15 INK4b genes are supposed to contribute to leukemogenesis in a substantial proportion of all cases (21,22). These results imply that appropriate cell cycle control, particularly at the early stage of stem/progenitor cells, is required for maintaining normal hematopoiesis.
c-Myc belongs to a family of transcription factors containing basic, helix-loop-helix, and leucine zipper domains. The expression of c-Myc is absent in quiescent cells and is rapidly induced by growth factors in normal tissues, whereas it is continuously overexpressed in a variety of neoplasms (23)(24)(25). In addition, c-Myc is highly expressed in various tissues during embryogenesis, and c-myc Ϫ/Ϫ mice die at embryonic day 10.5, indicating that c-Myc is essential for normal embryonic development (26). c-Myc, in conjunction with its heterodimer partner Max, regulates gene transcription through the element called the E box (CACGTG) (27,28). c-Myc induces a number of target molecules involved in G 1 /S transition such as E2Fs, CDC25A, CDK2, CDK4, Cul1, Id2, and Rb (29 -35). Also, c-Myc activates cyclin E/CDK2 complexes alone or in combination with Ras-mediated signals (36 -38). With these activities, c-Myc works at several points during the cell cycle. In fact, fibroblasts isolated from c-myc-deficient rats revealed a markedly prolonged cell doubling time accompanied by the drastic reduction of CDK4/6 and CDK2 activities (32,33). Furthermore, ectopic expression of c-Myc in quiescent cells under some conditions is sufficient for inducing S phase entry (39). These lines of evidence indicates that c-Myc plays a central role in G 1 /S transition as an upstream regulator of cell cycle regulatory molecules (40).
Once HSCs are induced to enter cell cycle under certain conditions, HSCs are obliged to select either self-renewal or differentiation during cell division. In this process, Notch signals and HOXB4 promote self-renewal of HSCs rather than differentiation as described above. Therefore, it is speculated that there must be a set of genes induced by Notch signals or HOXB4 contributing to self-renewal of HSCs. However, at present, their critical target molecules remain to be determined. Moreover, it remains unknown how this process is regulated by cell cycle regulatory molecules.
To clarify the roles of cell cycle regulatory molecules in the cell division of HSCs, in this study, we investigated the changes of their expression during Notch-or HOXB4-induced self-renewal of HSCs. We found that c-Myc was transcriptionally induced by Notch and HOXB4. In addition, we found that its ectopic expression induced the growth of HSCs without disrupting their biologic properties in terms of surface phenotypes, colony-forming activities, and reconstituting activities. Together, these results suggest that c-Myc plays a major role in the self-renewal of HSCs as an effector molecule of Notch signals and HOXB4.
Flow Cytometry-The DNA content of cultured cells was examined by staining with propidium iodide and analyzed on FACSort (BD Biosciences). Cell cycle analysis was performed with the program Modfit LT2.0 (BD Biosciences). Surface phenotypes of cultured cells were examined with FACSCalibur (BD Biosciences).
Luciferase Assays-To construct reporter genes containing various fragments of murine c-myc promoter, we performed PCR and subcloned PCR products into the luciferase plasmid. Luciferase assays were performed with a Dual-Luciferase reporter system (Promega, Madison, WI) as described previously (42). In short, 293T cells (2 ϫ 10 5 cells) cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum were seeded in a 60-mm dish and transfected with the appropriate effector gene (2 g) and reporter gene (2 g) together with 5 ng of pRL-CMV-Rluc, an expression vector of renilla luciferase, by the calcium phosphate coprecipitation method. After 12 h, the cells were washed and then serum-deprived for 24 h. Then, the cells were lysed and subjected to the measurement of the firefly and renilla luciferase activities on a luminometer LB96P (Berthold Japan, Tokyo, Japan). The relative firefly luciferase activities were calculated by normalizing transfection efficiency according to the renilla luciferase activities. The experiments were performed in triplicate, and similar results were obtained from at least three independent experiments.
Electrophoretic Mobility Shift Assay (EMSA)-Seven types of doublestranded oligonucleotides were used as probes or competitors, and their sequences are shown in the legend for Fig. 6.
Assays for Telomerase Activities-Telomerase activities were measured by the stretch PCR method with a TeloChaser system (TOYOBO, Osaka, Japan) according to the manufacturer's instructions.
Retrovirus Transfection into Murine Bone Marrow Cells-At first, we prepared conditioned medium containing high titer virus particles using Plat-E cells as described previously (42). The isolated Lin Ϫ Sca-1 ϩ cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum in the presence of mIL-3 (10 ng/ml), mSCF (50 ng/ml), hIL-6 (50 ng/ml), and hFL (30 ng/ml) for 72 h. Then, the cultured cells were cultured with conditioned medium containing high titer retrovirus in the presence of polybrene (8 g/ml). After 24 h, the cells were washed and cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum containing the same cytokines and 1 mg/ml G418 for 72 h. After the selection with G418, retrovirus-infected cells were cultured under the conditions as indicated.
Statistical Analysis-Statistical analysis was performed with Student's t test.

The Effects of Notch1 and HOXB4 on the Expression of Cell Cycle Regulatory Molecules in Murine Lin
we analyzed how self-renewal is regulated by cell cycle regulatory molecules in HSCs. For this purpose, we introduced a 4-HT-inducible form of Notch1 (N/ERT, a chimeric molecule consisting of the intracellular domain of Notch1 fused to the estrogen receptor) and HOXB4 into murine Lin Ϫ Sca-1 ϩ cells since both molecules were reported to induce self-renewal in HSCs (17,43). In the absence of 4-HT, N/ERT-transduced cells could not keep proliferating for 14 days under any cytokine combinations (i.e. SCF alone, SCF ϩ FL, or SCF ϩ FL ϩ IL-6) (Fig. 1A, upper panel). In contrast, 4-HT-treated cells kept growing under the same cytokine combinations during 14 days. Similarly, HOXB4-transduced cells could proliferate for up to 14 days, whereas mock (an empty vector)-transduced cells could not (Fig. 1A, lower panel). In DNA content analysis, the 4-HT treatment increased the proportion of the cells in S-G 2 /M phase from 11.6 to 22.2% in N/ERT-transduced cells under the culture with SCF (Fig. 1B, left panel). Also, HOXB4 augmented the proportion of the cells in S-G 2 /M phase from 12.4 to 26.5% under the culture with SCF ϩ FL ϩ IL-6 ( Fig. 1B, right panel).
Next, we examined the effect of Notch1 and HOXB4 on the expression of cell cycle regulatory molecules by semiquantitative RT-PCR analysis (Fig. 2). With or without the 4-HT treatment, total cellular RNA was extracted from N/ERT-transfected cells after a 7-day culture with SCF. Also, HOXB4transfected cells were subjected to RT-PCR analysis after a 7-day culture with SCF. At first, we adjusted the amount of cDNAs according to the amounts of the ␤-actin PCR products (Fig. 2, bottom panel). Then, we evaluated the expression levels of c-myc, cyclin D1, cyclin D2, cyclin D3, cyclin E, and E2F1. After 29 -35 cycles of PCR, all of these PCR products were amplified exponentially. As shown in Fig. 2, the 4-HT treatment markedly induced or enhanced the expression of c-myc, cyclin D2, cyclin D3, cyclin E, and E2F1, whereas it could not induce the expression of cyclin D1. Similarly, we found that HOXB4 enhanced the expression of c-myc, cyclin D2, cyclin D3, cyclin E, and E2F1 but not of cyclin D1.
c-Myc Enhances and Supports the Growth of Murine Lin Ϫ Sca-1 ϩ Cells-c-Myc is a primary regulator of G 1 /S transition, and its ectopic expression can drive quiescent cells to S phase in fibroblasts (as described under the Introduction). Therefore, we examined the effect of c-Myc on the growth and biologic properties of HSCs. We retrovirally transduced a 4-HTinducible form of Myc (Myc/ERT) to murine Lin Ϫ Sca-1 ϩ cells. When treated with 4-HT, Myc/ERT-transduced cells proliferated for more than 21 days under the culture with SCF, SCF ϩ FL, or SCF ϩ FL ϩ IL-6 ( Fig. 3A). In contrast, 4-HT-untreated cells died within 21 days under the same cytokine combinations. In DNA content analysis, the 4-HT addition prominently increased the cells in S-G 2 /M phase in Myc/ERT-transduced cells after 7 days (SCF, from 11.7 to 49.3%; SCF ϩ FL, from 11.9 to 52.2%; SCF ϩ FL ϩ IL-6, from 12.2 to 53.2%) (Fig. 3B). Next, we examined the effects of c-Myc on the expression profile of cell cycle regulatory molecules. With or without the 4-HT treatment, we isolated total cellular RNA from Myc/ERT-transfected cells after a 7-day culture with SCF and performed semiquantitative RT-PCR analysis. In Myc/ERT-transduced cells, 4-HT-activated c-Myc induced or enhanced the expression of cyclin D1, cyclin D2, cyclin D3, cyclin E, and E2F1 mRNA (Fig. 3C), suggesting a possibility that c-Myc would be a primary regulator of Notch-or HOXB4-mediated cell growth among various cell cycle regulatory molecules.
Characterization of the Cells Expanded by Myc/ERT Using in Vitro Assays-The 4-HT treatment was able to keep Myc/ERTtransduced cells viable even after a 28-day culture with SCF or SCF ϩ FL ϩ IL-6 (data not shown). To investigate the biologic properties of these cells, we initially examined their surface phenotypes by fluorescence-activated cell sorter. Before the addition of 4-HT, 77.2% of the Myc/ERT-transduced cells were Sca-1 ϩ c-Kit ϩ , and 94.7% (64.9 ϩ 29.8) were Lin Ϫ (including CD34-positive and negative cells). After a 28-day culture with SCF, more than 50% of the cultured cells still revealed immature surface phenotypes: Sca-1 ϩ c-Kit ϩ cells, 56.6%; Lin Ϫ cells, 63.4% (24.5 ϩ 38.9) (Fig. 4A). Similar results were also observed after a 28-day culture with SCFϩFLϩIL-6: Sca-1 ϩ c-Kit ϩ cells, 47.7%; Lin Ϫ cells, 81.3% (32.0 ϩ 49.3). Next, we analyzed the effects of c-Myc on the colony-forming ability of HSCs. After the 3-day selection with G418, Myc/ERT-transduced Sca-1 ϩ Lin Ϫ cells were seeded into methylcellulose media (containing IL-3, IL-6, SCF, and erythropoietin) and cultured with or without 4-HT for 12 days. As shown in Fig. 4B, 4-HTactivated Myc/ERT increased the number of CFU-Mix with a statistically significant difference p Ͻ 0.05. Also, Myc/ERT significantly increased burst-forming unit-E (p Ͻ 0.05) and colony forming units-granulocyte/macrophage (p Ͻ 0.01). We also analyzed changes of telomerase activity in the Myc/ERTtransduced cells. Quiescent HSCs have been reported to exhibit a low level of telomerase activity, which is up-regulated in response to cytokine stimulation (44,45), whereas it declines during the maturation process (46). After a 7-day culture in the presence of SCF with or without the 4-HT treatment, we isolated cell extracts from Myc/ERT-transduced cells and evaluated telomerase activity by the stretch PCR method. As observed in a positive control, Hela, the 4-HT treatment induced telomerase activity in Myc/ERT-transduced cells, of which specificity was confirmed by its heat instability. In contrast, Myc/ERT-transduced cells cultured without 4-HT did not have telomerase activity (Fig. 4C). Together, these results suggest that c-Myc-induced cell growth does not spoil the immature characteristics of HSCs.
Reconstitution Assays Using HSCs Expanded by Myc/ERT-Next, we evaluated the in vivo function of Myc/ERT-expanded HSCs with reconstitution assays. After 12-day cultures, 4-HTtreated or untreated Myc/ERT-transduced cells (4 ϫ 10 6 cells) with Ly5.1 phenotype were transfused into lethally irradiated Ly5.2 mice in combination with 1 ϫ 10 5 Ly5.2 ϩ normal bone marrow cells. When 4-HT-treated cells were transplanted, 11.3% of the peripheral blood cells were Ly5.1 ϩ in the recipient mice after 4 weeks (Fig. 4D, upper left panel). In contrast, Ly5.1 ϩ cells were hardly detected in the mice transplanted with 4-HT-untreated cells after 4 weeks (Fig. 4D, upper right  panel). Furthermore, Ly5.1 ϩ cells were still detected in 7.9% of the peripheral blood cells in the mice transplanted with 4-HTtreated cells even after 12 weeks (Fig. 4D, lower left panel). In this Ly5.1 ϩ fraction, 15% of the cells were Mac1/Gr-1 ϩ , 37.1% were CD4/8 ϩ , and 50.7% were B220 ϩ , indicating that HSCs expanded by Myc/ERT has an ability of multilineage reconstitution. Also, these cells were able to supply hematopoietic cells in the recipient mice for more than 6 months (data not shown). Collectively, these results suggested that ectopically introduced c-Myc can induce self-renewal of HSCs.
The Effects of Notch Signaling and HOXB4 on c-myc Promoter Activities-Since c-Myc was considered to be important for self-renewal of HSCs, we next tried to clarify the mecha-  4. Characterization of the cells expanded and maintained by c-Myc for 28 days. A, the surface phenotype of Myc/ERT-transduced cells was examined by flow cytometric analyses before and after the 28-day culture with 4-HT in the presence of SCF or SCF ϩ FL ϩ IL-6. The antilineage Abs recognize Gr-1, B220, CD3e, Mac1 and Ter119. B, Myc/ERT-transduced murine Lin Ϫ Sca-1 ϩ cells were cultured with G418 for 3 days, and the living cells were seeded into methylcellulose media (containing IL-3, IL-6, SCF, and erythropoietin). The number of the indicated nism through which Notch1 and HOXB4 induce the c-myc mRNA expression. For this purpose, we performed luciferase assays with a reporter gene Ϫ1137/ϩ516-Luc containing the murine c-myc promoter (numbered from the transcription initiation site) with NotchIC (an activated form of Notch1), RBP-VP16 (a transcriptionally active derivative of RBP-J), and HOXB4 as effectors. In 293T cells, NotchIC, RBP-VP16, and HOXB4 activated Ϫ1137/ϩ516-Luc by 2.9-, 6.9-, and 4.7-fold, respectively, whereas these effector genes hardly stimulated its backbone vector ϩ141/ϩ516-Luc (Fig. 5A). Similar results were obtained from a murine fibroblast cell line NIH3T3 (data not shown). To determine which region is responsive to Notch in the c-myc promoter, we generated several reporter genes and performed luciferase assays with RBP-VP16 as an effector gene. As shown in Fig. 5B, Ϫ195/ϩ24-Luc was most responsive to RBP-VP16 (18-fold induction). Also, the similar level (16.1fold) of activation was induced by RBP-VP16 in Ϫ195/Ϫ31-Luc. To further identify which element is responsive to RBP-VP16, we isolated nuclear extracts from 293T cells transfected with FLAG-RBP-VP16 or a mock vector and performed EMSA with seven overlapping probes spanning between Ϫ195 and Ϫ11 of the c-myc promoter (Fig. 6A). As compared with DNA binding patterns formed from mock-transfected cells, nuclear extracts isolated from FLAG-RBP-VP16-transfected cells bound to the colonies was counted after 12 days. The results are shown as mean Ϯ S.D. of quintupled cultures. BFU-E, burst-forming unit-E. CFU-GM, colony forming units-granulocyte/macrophage. C, after 7-day cultures, the cultured cells were subjected to the stretch PCR method for the evaluation of telomerase activities. probes 1, 6, and 7 (Fig. 6B, upper panel, lanes 2, 12, and 14, DNA-binding complexes are indicated by arrows), whereas a significant difference was not detected with the other probes. These binding complexes were abolished by non-labeled wildtype competitors (Fig. 6B, lower panel, lanes 3, 7, and 11) but not by mutated competitors (Fig. 6B, lower panel, lanes 4, 8,  and 12), indicating that these complexes were formed in a sequence-specific manner. To further characterize the role of these sites in Notch signals, we performed luciferase assays with reporter genes each containing the sequence of the probe 1 (Ϫ195/Ϫ161-Luc) and probes 6 -7 (Ϫ70/Ϫ11-Luc). As shown in Fig. 6C, RBP-VP16 activated Ϫ195/Ϫ161-Luc by 4.1-fold, whereas it was hardly effective on Ϫ70/Ϫ11-Luc, implying that the protein that binds to the probe 1 would contribute to the activation of the c-myc promoter by Notch signaling. However, this complex reacted neither to the anti-FLAG Ab nor to the anti-RBP-J Ab (data not shown). Also, the probe 1 did not contain the typical RBP-J-binding sequence. These results suggest that RBP-VP16 would activate this element indirectly. In addition, since the reactivity of Ϫ195/Ϫ161-Luc was not as prominent as that of Ϫ195/Ϫ31-Luc, it was speculated that an additional element would be necessary for the c-myc promoter to achieve the maximal response to Notch signals.

DISCUSSION
With regard to the roles for Notch and HOXB4 in cell cycle regulation, Notch1 activation was reported to shorten the G 1 phase of cell cycle, thereby delaying the differentiation of human HL-60 cells and primary CD34 ϩ cells (47). Also, Ronchini and Capobianco (48) showed that the intracellular domain of Notch directly activated cyclin D1 promoter in a human kidney cell line, 293T. In addition, Krosl and Sauvageau (49) reported that overexpression of HOXB4 led to the induction of Jun-B and Fra-1 expression and subsequent up-regulated expression of cyclin D1 in Rat-1 fibroblast. To identify a critical target molecule in Notch-and HOXB4-induced self-renewal of HSCs, we analyzed their effects on the expression of cell cycle regulatory molecules and found that both molecules up-regulated the expression of c-myc, cyclin D2, cyclin D3, cyclin E, and E2F1. However, neither Notch 1 nor HOXB4 up-regulated the expression of cyclin D1, whereas its expression was up-regulated by these molecules in other cell types (as described above). These findings were consistent with the previous report that cyclin D1 was not detected in the proliferating normal hematopoietic stem/progenitor cells (50). Although cyclin D1 expression did not accompany the proliferation of normal hematopoietic stem/progenitor cells, our results suggest that G 1 /S transition in self-renewal of HSCs is not so unique but rather common to other cell types. In addition, we found that Myc/ ERT efficiently induced the expression of cyclin D2, cyclin D3, and cyclin E and promoted the growth of HSCs. In addition, previous reports suggested that c-Myc would transcriptionally repress the expression of p21 WAF1 , which plays a crucial role in the quiescence of HSCs (19,51), and of p15 INK4b , which is induced by transforming growth factor-␤ signaling (52,53). Together, these results imply that the c-myc could be a master regulator of the cell cycle in both Notch1-and HOXB4-induced self-renewal of HSCs.
In this study, promoter assays revealed that both Notch signaling and HOXB4 activated the c-myc promoter. However, the element most responsive to RBP-VP16 in the c-myc promoter did not contain the putative RBP-J-binding sequence. Also, we confirmed that the nuclear protein, which bound to this element, did not contain RBP-VP16 itself in EMSA with supershift assays, suggesting RBP-VP16 could activate the c-myc promoter indirectly. To understand the mechanism of Notch-induced self-renewal of HSCs, we are now trying to purify this molecule using nuclear extracts from RBP-VP16transfected 293T cells.
In this study, reconstitution assays revealed that cells expanded by Myc/ERT could contribute to hematopoiesis for more than 6 months, indicating that these cells still possess characteristics of HSCs in terms of long term reconstitution. In addition, these cells developed Mac1/Gr-1 ϩ cells, CD4/8 ϩ cells, and B220 ϩ cells, implying that the cells expanded by Myc/ERT have a multilineage reconstitution ability. When we compare the reconstitution efficiency of Myc/ERT-expanded cells with that of Notch1-expanded cells (43), Myc/ERT-expanded cells showed similar or a little more potent reconstitution abilities than Notch-expanded cells. (120-fold excess of Notch1-expanded cells (i.e. transplantation of 120 ϫ 10 5 Notch1-expanded cells in combination with 1 ϫ 10 5 normal BM cells) contributed to hematopoiesis in about 23% of BM cells after transplantation versus 40-fold excess of Myc/ERT-expanded cells (i.e. transplantation of 40 ϫ 10 5 Myc/ERT-expanded cells in combination with 1 ϫ 10 5 normal BM cells) participated in hematopoiesis in about 10% of the peripheral blood cells.) However, even if sufficient numbers (1 ϫ 10 7 ) of Myc/ERT-expanded cells were transplanted, co-injection of normal supporting BM cells was required for radioprotection in the recipient mice (data not shown). These results indicate that the cells expanded by Myc/ ERT lack some ability to fully or rapidly reconstitute hematopoiesis in vivo. The similar defect was observed in HSCs expanded by Notch1 (43). Like Myc/ERT-expanded cells, Notch1expanded cells alone could not reconstitute hematopoiesis in lethally irradiated mice, whereas these cells could participate in hematopoiesis for more than 6 months in combination with normal supporting BM cells. One possible explanation is that the ex vivo culture of HSCs or modulation of cell cycle by Notch1 or c-Myc might result in the reduced homing abilities after intravenous transplantation. For example, Szilvassy et al. (54) showed that HSCs expanded by growth factors in vitro expressed little or no ␤ 1 integrin, and this change was associated with a failure of radioprotection. Alternatively, HSCs induced to proliferate by Notch or c-Myc in vitro might not be able to rapidly undergo terminal differentiation after transplantation, resulting in the failure to supply a sufficient number of mature cells required for radioprotection.
In summary, our results indicate that c-Myc is, at least in part, capable of supporting self-renewal of HSCs as a downstream mediator of Notch and HOXB4. Further analyses on cell cycle regulation would undoubtedly provide more informative findings to expand HSCs in vitro.