C/EBPβ isoforms sequentially regulate regenerating mouse hematopoietic stem/progenitor cells

Proliferation, self-renewal, and differentiation of hematopoietic stem cells (HSCs) is tightly controlled.^1,2  Upon exposure to hematological stress, such as infection, inflammation, treatment with chemotherapeutic agents, or transplantation into irradiated recipients, HSCs undergo drastic changes characterized by rapid cell-cycle entry and accelerated differentiation, especially toward the myeloid lineage, at the expense of self-renewal capacity, to meet the increasing hematopoietic demand.^3-6  Activation and recruitment of myeloid-biased HSCs underlie the elevated level of myeloid-biased differentiation under stress conditions.⁷  A recent study showed that all lineage-biased multipotent progenitor (MPP) subsets are independent progenies of long-term HSCs (LT-HSCs) or short-term HSCs (ST-HSCs).⁸  In response to hematopoietic stresses, replenishment of myeloid-biased MPPs (megakaryocyte-biased MPP2s and granulocyte/macrophage-biased MPP3s) from HSCs is upregulated, and the fate of lymphoid-biased MPP4s is reprogrammed toward the myeloid lineage. Together, these findings suggest that myeloid transcription factors are involved in the altered behavior of hematopoietic stem/progenitor cells (HSPCs) under stress conditions.

CCAAT enhancer-binding proteins (C/EBPs) are a family of leucine zipper transcription factors. C/EBPα accelerates differentiation and inhibits proliferation of HSPCs, and thus is indispensable to the maintenance of steady state granulopoiesis.^9-11  In contrast, C/EBPβ plays critical roles in stress-induced granulopoiesis, which requires the differentiation and proliferation of granulopoietic precursors.^12-16  We and others demonstrated that C/EBPβ is required at the level of HSPCs in stress conditions.^17-20  However, the roles of C/EBPβ in the regulation of HSPCs, as well as its molecular mechanisms of action, remain largely unclear.

The Cebpb gene contains a single exon, and 3 isoforms (liver-enriched activating protein* [LAP*], LAP, and liver-enriched inhibitory protein [LIP]) are translated from its unique mRNA.^21,22  These isoforms have distinct functions, and the ratio of LIP to LAP*/LAP (termed the LIP/LAP ratio) is an important determinant of C/EBPβ function.^21-23  However, technical limitations have hampered intensive analysis of these isoforms in extremely rare cells such as HSCs. To overcome this obstacle, we developed an intracellular flow cytometric technique capable of detecting distinct domains of C/EBPβ. We then used this method to reveal the distinct functions of these isoforms and the critical roles of C/EBPβ in the regulation of HSPCs under stress conditions.

After surface marker staining, bone marrow (BM) cells (1 × 10⁶ to 3 × 10⁶) were fixed and permeabilized using the Transcription Factor Buffer Set (BD Biosciences, Franklin Lakes, NJ), and stained with primary and secondary antibodies. For double staining of C/EBPβ isoforms, samples were stained with a goat anti-C–terminus antibody (C-19, sc-150G; Santa Cruz Biotechnology, Dallas, TX) and a rabbit anti-N–terminus antibody (#3807; Cell Signaling Technology, Danvers, MA), followed by an Alexa Fluor 647–conjugated donkey antigoat IgG antibody and a BV421-conjugated donkey antirabbit IgG antibody.

Statistical analyses were performed using GraphPad Prism or Microsoft Excel. Statistical significance was calculated and determined by 2-tailed Student t test, except in Figures 6B and 7A, in which the 2-tailed paired Student t test was used. Values of P < .05 were considered statistically significant.

Information regarding mice, cell lines, retrovirus vectors, BM transplantation, flow cytometric analysis, cell sorting, intracellular Ki67/7-amino actinomycin D staining, short-term BrdU incorporation assay, reverse transcription polymerase chain reaction (RT-PCR), immunoblotting, and chromatin immunoprecipitation–qualitative polymerase chain reaction (ChIP-qPCR) can be found in the supplemental Materials and methods.

To assess the significance of C/EBPβ in the regulation of HSPCs in vivo, we compared the numbers of HSPCs in the BM between wild-type (WT) and Cebpb^−/− mice at steady state. The frequencies of KL (c-kit⁺Sca-1⁻lineage⁻) and KSL (c-kit⁺Sca-1⁺lineage⁻) cells corresponding to myeloid progenitors and HSPCs, respectively,²⁴  were almost identical in WT and Cebpb^−/− mice (Figure 1A). Within KSL cells, ST-HSCs, MPP2s, MPP3s, and MPP4s were defined on the basis of surface expression of CD135, CD48, and CD150 (supplemental Figure 1A).^25,26  LT-HSCs were defined as CD135⁻CD150⁺CD48⁻ KSL cells, or CD150⁺CD48⁻ KSL cells when CD135 was not stained. The frequencies of these HSPCs subsets were also comparable between WT and Cebpb^−/− mice (Figure 1B), suggesting that C/EBPβ is dispensable for the development and maintenance of HSPCs at steady state.

Our previous findings suggest that there is a specific requirement for C/EBPβ in stress conditions.^{12,16,17,18,20,27}  To further characterize this requirement, we first measured Cebpb and Cebpa mRNA expression in HSPCs isolated from WT mice treated with 5-fluorouracil (5-FU). The gating strategy for LT-HSCs after 5-FU treatment was verified by expression of 2 other HSC markers, CD201 and endothelial cell-selective adhesion molecule^28-30  (supplemental Figure 1B). Cebpb mRNA expression in LT-HSCs and KSL cells was not significantly altered by 5-FU treatment, whereas expression of Cebpa was significantly downregulated (Figure 1C). In contrast, intracellular flow cytometry revealed that C/EBPβ protein was significantly upregulated in LT-HSCs and KSL cells on days 3 and 6 after 5-FU treatment (Figure 1D). Taken together, these findings demonstrate that C/EBPβ has little impact on HSPCs at steady state but is upregulated at the protein level under stress conditions.

To assess the requirement for C/EBPβ in HSPCs, we performed a competitive repopulation assay (Figure 2). In these experiments, BM cells from WT or Cebpb^−/− mice (CD45.2⁺) mixed with an equal number of competitor BM cells from CD45.1⁺ WT mice were transplanted into lethally irradiated recipient WT mice (CD45.1⁺). Chimerism of donor-derived CD45.2⁺ cells among 3 lineages (granulocytes, monocytes, and lymphocytes, as defined in supplemental Figure 1C) in the peripheral blood (PB) of the recipients was monitored monthly (Figure 2A). As we reported previously,^31,32  the contribution of Cebpb^−/− cells to monocytes was severely impaired. Moreover, chimerism of Cebpb^−/− cells was significantly lower than that of WT donor cells among granulocytes and lymphocytes at 1 month after transplantation. At 6 months after transplantation, these differences were still observed in granulocytes and monocytes, but became less in total leukocytes or lymphocytes. At this time point, chimerism among LT-HSCs and KSL cells did not significantly differ between recipients of WT and Cebpb^−/− cells (Figure 2B).

Next, we evaluated the self-renewal capacity of Cebpb^−/− HSCs by performing secondary transplantations (Figure 2C). Six months after primary transplantation, 2 × 10⁶ BM cells harvested from the primary recipients were transplanted into lethally irradiated CD45.1⁺ WT recipients. Chimerism of Cebpb^−/− cells in the secondary recipients was initially lower than that of WT cells but gradually increased to equal that of WT cells among all lineages except for monocytes. At 4 months, chimerism of Cebpb^−/− cells was slightly higher than that of WT cells among LT-HSCs and KSL cells in the secondary recipients (Figure 2D), although this difference was not statistically significant. These results suggest that C/EBPβ is required for early engraftment and is possibly involved in the exhaustion of HSPCs in response to stress induced by transplantation.

We further investigated the impact of C/EBPβ on the differentiation and exhaustion of HSPCs under enhanced stress (Figure 2E). One month after competitive BM transplantation, recipient mice were subjected to monthly 5-FU treatment. Consistent with the results shown in Figure 2A,C, chimerism of Cebpb^−/− cells was significantly lower than that of WT cells among all lineages at 1 month after transplantation. With repetitive 5-FU treatment, chimerism of Cebpb^−/− cells among all lineages in PB increased over time and exceeded that of WT cells by 5 months after transplantation. In accordance with these changes in PB, chimerism of Cebpb^−/− cells was significantly higher than that of WT cells among LT-HSCs and KSL cells (Figure 2F). Taken together, these data clearly suggest that C/EBPβ promotes the exhaustion of HSPCs under stress conditions.

The differences between WT and Cebpb^−/− cells during early engraftment could be because of a difference in the homing ability or early expansion of HSPCs. After confirming that the homing ability of Cebpb^−/− KSL cells was similar to that of WT KSL cells (supplemental Figure 2), we assessed the role of C/EBPβ in the initial expansion of HSPCs (Figure 3A). A total of 1000 purified KSL cells from WT or Cebpb^−/− mice (CD45.2⁺) mixed with an equal number of competitor KSL cells from WT mice (CD45.1⁺) were transplanted into lethally irradiated CD45.1⁺CD45.2⁺ WT recipient mice. Four weeks after transplantation, chimerism of CD45.2⁺ test cells and the ratio of CD45.2⁺ test cells to CD45.1⁺ competitor cells among KSL cells were significantly lower in the recipients of Cebpb^−/− cells (Figure 3B). This suggests that C/EBPβ is cell-intrinsically involved in the initial expansion of HSPCs after transplantation.

We next examined the involvement of C/EBPβ in cell-cycle regulation in HSPCs. At steady state, the cell-cycle status of LT-HSCs was identical in Cebpb^−/− and WT mice (Figure 3C-E). Thirty-six hours after transplantation of BM cells from WT or Cebpb^−/− mice (supplemental Figure 3), a significantly higher percentage of Cebpb^−/− LT-HSCs remained in G0 phase. This suggests that C/EBPβ plays an important role in cell-cycle entry after transplantation.

In parallel with the transplantation model described above, we analyzed the roles of C/EBPβ using a 5-FU treatment model. To exclude the cell-extrinsic contribution of C/EBPβ, we transplanted CD45.2⁺ BM cells from WT or Cebpb^−/− mice into lethally irradiated CD45.1⁺ WT mice (Figure 4A). Twelve weeks later, >99% of BM cells in the recipients were replaced by CD45.2⁺ donor cells, and the cell-cycle statuses and abundances of LT-HSCs, KSL cells, and KL cells did not significantly differ between the recipients of WT and Cebpb^−/− cells (Figure 4B-C, day 0). These BM-replaced mice were treated once with 5-FU (supplemental Figure 4).

We first evaluated the impact of C/EBPβ on the cell-cycle status of LT-HSCs (Figure 4B). After 5-FU treatment, WT LT-HSCs immediately entered the cell cycle, as revealed by a decrease in the proportion of cells in G0 phase and an increase in the proportion of cells in S/G2M phase. The ratio of cells in G1 phase to cells in G0 phase, indicating entry of dormant cells into the cell cycle, also increased and peaked at day 5. All of these parameters of cell-cycle entry were significantly attenuated in recipients of Cebpb^−/− cells. Consistent with this, short-term BrdU incorporation assays revealed that significantly fewer Cebpb^−/− HSPCs were cycling 6 days after 5-FU administration, whereas no difference was detected between Cebpb^−/− and WT HSPCs at steady state (supplemental Figure 5A). Collectively, these data suggest that C/EBPβ helps to accelerate cell-cycle entry of LT-HSCs during the early phase of regeneration.

Next, we assessed the numbers of LT-HSCs, KSL cells, and KL cells (Figure 4C). After the nadir induced by 5-FU, recovery of LT-HSCs was followed by recovery of KSL and KL cells. In recipients of Cebpb^−/− cells, the recovery of KSL and KL cells, but not the recovery of LT-HSCs, was more delayed than in recipients of WT cells, indicating an impairment in the differentiation of Cebpb^−/− LT-HSCs into KSL or KL cells during regeneration. Accordingly, LT-HSCs were more enriched in Cebpb^−/− BM-reconstituted mice than in WT BM-reconstituted mice on days 10 and 18 after 5-FU treatment (Figure 4D). In addition, there were significantly fewer CD11b⁺ myeloid cells in the BM of Cebpb^−/−-reconstituted mice on day 18 (supplemental Figure 5B). These results suggest that C/EBPβ promotes differentiation of LT-HSCs into myeloid lineages at a later phase of regeneration.

After identifying the phase-specific roles of C/EBPβ in the regulation of HSPCs under stress conditions, we sought to determine the distinct functions of 3 isoforms of C/EBPβ. To monitor expression of these isoforms in small numbers of HSPCs, we developed an intracellular double-staining method for flow cytometric analysis. An antibody against the C terminus of C/EBPβ recognized all 3 isoforms, whereas an antibody against the N terminus only recognized LAP* and LAP (Figure 5A-B). Thus, simultaneous staining with both antibodies allowed us to distinguish cells that predominantly expressed LIP (C-term⁺N-term⁻) from those that expressed all 3 isoforms (C-term⁺N-term⁺).

To test this strategy, we first applied it to the mouse HSPC line EML cells transduced with one of the C/EBPβ isoforms (Figure 5C). As expected, LIP-transduced cells were identified as N-term⁻C-term⁺ cells and could be clearly distinguished from LAP*- and LAP-transduced cells, which were C-term⁺N-term⁺ (Figure 5C). In quantitative terms, the mean fluorescence intensities (MFIs) of staining with the anti-C–terminus and anti-N–terminus antibodies reflected the amounts of total C/EBPβ protein (LIP, LAP, and LAP*) and the long isoforms (LAP* and LAP), respectively. Hence, the ratio of the MFI^C-terminus to the MFI^N-terminus, termed the C/N ratio, is a surrogate for the LIP/LAP ratio.

Using this method, we monitored the expression patterns of these isoforms in LT-HSCs treated with 5-FU. BM cells obtained from mice at steady state were stained in parallel and used as controls (supplemental Figure 6). As shown in Figure 5D, LT-HSCs first became C-terminus single-positive (days 1-4), and subsequently changed to C-term⁺N-term⁺ (after day 6). During the interval when total C/EBPβ protein was upregulated (days 1-8; Figure 5E), the C/N ratio increased during the early phase (days 1-4; Figure 5F), peaked at day 3, and decreased to steady-state level during the later phase (days 6-8). Similarly, after transplantation, LIP was upregulated prior to LAP*/LAP in LT-HSCs (supplemental Figure 7A-C). These data suggest that LIP is upregulated prior to LAP*/LAP during regeneration.

Mammalian/mechanistic target of rapamycin (mTOR) signaling is involved in the translation of C/EBPβ.³³  We confirmed that the mTOR pathway was activated in LT-HSC after 5-FU treatment (supplemental Figure 8). To explore the role of mTOR, we treated mice with rapamycin, an allosteric mTOR inhibitor, prior to 5-FU treatment. Rapamycin significantly delayed and attenuated the upregulation and alteration of C/EBPβ isoforms (green lines in Figure 5E-F), and the hematopoietic recovery was significantly delayed (supplemental Figure 9A-B). This finding suggests that mTOR is involved in the translational regulation of C/EBPβ in HSPCs during regeneration.

LT-HSCs with high CD150 expression (CD150^high LT-HSCs) exhibit higher self-renewal activity and more robust myeloid-biased reconstitution at steady state.^34,35  Given that C/EBPβ plays an important role in emergency granulopoiesis, we hypothesized that C/EBPβ is closely associated with CD150^high LT-HSCs. Indeed, at day 3 after 5-FU administration, when the upregulation of LIP peaked (Figure 5E-F), C/EBPβ expression in LT-HSCs correlated well with surface expression of CD150 (Figure 5G). Furthermore, expression levels of C/EBPβ and LIP were higher in CD150^high LT-HSCs than in CD150^low LT-HSCs at the early phase of regeneration (Figure 5H-I). These findings were reproduced in the BM transplantation model (supplemental Figure 10), suggesting that stress-induced LIP-dominant upregulation is specific to CD150^high LT-HSCs during regeneration.

We next assessed the function of each C/EBPβ isoform. For in vitro evaluation, we retrovirally transduced EML cells³⁶  with one of the C/EBPβ isoforms. LIP-transduced EML cells were slightly, but significantly, more proliferative and actively cycling than those transduced with the control vector, whereas proliferation and cell cycling were markedly suppressed in LAP*- and LAP-expressing EML cells (Figure 6A-B). LIP-expressing cells remained undifferentiated (c-kit^highCD11b⁻) for more than 2 weeks, whereas LAP*- and LAP-expressing cells rapidly differentiated into c-kit^lowCD11b⁺ myeloid cells and eventually stopped proliferating within 2 weeks (Figure 6C; supplemental Figure 11; data not shown).

For in vivo examination, primary BM cells from WT or Cebpb^−/− mice were retrovirally transduced with one of the C/EBPβ isoforms or the control vector, and then transplanted into irradiated WT mice (Figure 6D). After 2 (PB) or 3 (BM) weeks, LAP*- and LAP-expressing cells disappeared from the recipients (supplemental Figure 12A-B; data not shown), indicating that they underwent rapid differentiation and exhaustion. In contrast, LIP-transduced cells stably engrafted for 12 weeks after transplantation and gave rise to significantly more KSL cells at 3 and 16 weeks after transplantation than cells transduced with the control vector (Figure 6E; supplemental Figures 12C and 13A). These time points represented stress condition and steady state, respectively. In addition, cell-cycle analyses revealed that LT-HSCs expressing LIP were less quiescent than those transduced with the control vector (Figure 6F; supplemental Figure 12C). These data were consistent with those obtained in vitro. However, further subdivision of KSL cells revealed that LIP-transduced WT cells gave rise to significantly fewer LT-HSCs and ST-HSCs and significantly more lymphoid-primed MPP4s at both 3 and 16 weeks (Figure 6G; supplemental Figure 13A). Interestingly, transplantation of LIP-transduced cells resulted in myeloid-biased output at 3 weeks, but lymphoid-biased output at 16 weeks after transplantation (Figure 6H). When the recipient mice were treated with 5-FU at 16 weeks after transplantation to induce stress condition again, LIP-transduced cells gave rise to more myeloid cells than control vector-transduced cells. Consistently, transfer of LIP-transduced MPP4s gave rise to more myeloid cells under a stress condition induced by irradiation, but less myeloid cells at steady state (supplemental Figure 13B-C). Taken together, these results suggest that LIP negatively regulates quiescence of LT-HSCs and triggers differentiation to MPP4 and that LIP modulates further differentiation in a condition-dependent manner. In contrast, LAP*/LAP induces rapid myeloid differentiation and exhaustion of HSPCs.

To explore the machinery downstream of C/EBPβ, especially the factors acting downstream of the initial elevation of LIP in HSPCs during regeneration, we focused on c-Myc, which is encoded by Myc. c-Myc promotes exit of quiescent HSCs from the niche and their differentiation into progenitors.³⁷  Therefore, we hypothesized that LIP regulates regenerating HSCs by upregulating c-Myc.

To test this idea, we first assessed the effect of LIP overexpression. Retroviral transduction of LIP increased Myc expression by 1.5-fold and twofold in EML cells and LT-HSCs, respectively (Figure 7A-B). When the recipients of WT or Cebpb^−/− BM cells were analyzed at 36 hours after transplantation, the time point at which LIP was predominantly expressed, Myc expression was significantly higher in WT than Cebpb^−/− LT-HSCs (Figure 7C). After 5-FU treatment, Myc expression was twofold higher in WT than Cebpb^−/− LT-HSCs only on day 3, when LIP was predominantly expressed. These results suggest that LIP positively regulates Myc expression in HSCs under stress condition in vivo. (Figure 7D). Changes in Myc expression in LT-HSCs induced by LIP or Cebpb-deficiency were accompanied with changes in expression of putative Myc target genes (supplemental Figure 14A-B, respectively). These results suggest that LIP positively regulates Myc expression in HSCs under stress condition.

A recent study demonstrated that C/EBPβ binds to distal enhancers of the Myc gene.³⁸  To confirm this in the context of HSPCs, we performed ChIP-qPCR analysis of EML cells transduced with C/EBPβ fused to the estrogen receptor. At 24 and 72 hours after nuclear translocation of C/EBPβ, C/EBPβ was remarkably enriched at the E3 enhancer of the Myc gene (Figure 7E). To identify the isoforms bound to the enhancer, we retrovirally transduced NIH3T3 cells with vectors expressing one of the isoforms, and then subjected the cells to ChIP-qPCR. The data clearly showed that all isoforms were bound to the enhancer (Figure 7F). In functional terms, LAP or LAP* overexpression significantly downregulated Myc in EML cells, whereas overexpression of LIP upregulated Myc (Figure 7A,G). Collectively, these findings suggest that upregulation of LIP in HSPCs positively regulates Myc transcription, possibly by antagonistically inhibiting LAP/LAP*-mediated downregulation at the Myc enhancer.

C/EBPα and C/EBPβ are involved in the regulation of steady state and emergency granulopoiesis, respectively. C/EBPα restricts the proliferation of HSCs during steady state.^10,11  In contrast, as revealed by this study, sequentially upregulated C/EBPβ isoforms regulate the proliferation and differentiation of regenerating HSCs.

Using an intracellular flow cytometric technique that we developed, we showed that LIP was upregulated before the longer isoforms, LAP and LAP*, in LT-HSCs after 5-FU treatment. Cell-cycle entry in Cebpb^−/− LT-HSCs was impaired at an early phase, exactly when LIP was predominantly expressed. In contrast, differentiation toward the myeloid lineage in the recipients of Cebpb^−/− cells was delayed during the period when the long isoforms were upregulated. These results suggest that phase-specific upregulation of LIP and LAP/LAP* is strongly associated with phase-specific functions of C/EBPβ in cell-cycle activation and differentiation, respectively.

Our in vitro experiments using EML cells clearly showed that LIP accelerated cell growth and entry into the cell cycle, whereas LAP and LAP* inhibited cell proliferation. Indeed, transduction of LIP decreased the number of quiescent LT-HSCs in vivo. In mechanistic terms, LIP transduction upregulated Myc in LT-HSCs. In addition, expression of Myc was significantly lower in Cebpb^−/− LT-HSCs only when LIP was predominantly upregulated. These results, together with the ChIP-qPCR results demonstrating that all isoforms of C/EBPβ directly bound to an enhancer of Myc, suggest that LIP accelerates the cell-cycle entry under stress conditions by positively regulating Myc.

Our experiments using EML cells revealed that LIP has little myeloid differentiation-inducing ability when compared with LAP and LAP*. In contrast, in vivo transplantation of LIP-transduced BM cells or LIP-transduced MPP4s resulted in myeloid-biased output under stress conditions, but more lymphoid-biased output at steady state. We hypothesize that LIP may indirectly induce the differentiation of LT-HSCs into MPP4s by inhibiting their interaction with the niche through positively regulating Myc,³⁷  irrespective of the situation. Under stress condition, LIP may collaborate other myeloid-inducing machineries activated by cell-extrinsic factors. At steady state, LIP may inhibit further myeloid differentiation by antagonizing LAP and LAP* or other C/EBP family members in a dominant-negative manner. Because overexpression of C/EBPβ isoforms affected the behavior of Cebpb^−/− LT-HSCs readily at steady state, the detailed mechanisms involved in the endogenous C/EBPβ-mediated lineage determination under stress condition should be addressed using isoform-specific knockout mice in the future.

By subdividing regenerating LT-HSCs on the basis of their surface expression of CD150, we found that stress-induced upregulation of C/EBPβ was remarkably specific to the CD150^high subset, which represents myeloid-biased HSCs at steady state.^34,35  Furthermore, a detailed review of our cell-cycle analyses revealed that C/EBPβ specifically activated the cell cycle in CD150^high LT-HSCs under stress conditions (data not shown). In addition, close observation of mice that underwent competitive transplantation followed by repetitive 5-FU treatment (Figure 2E) revealed that Cebpb^−/− cells outcompeted WT cells more effectively in myeloid lineages than in lymphocytes, indicating that C/EBPβ acts specifically in myeloid-biased HSCs, and induces their differentiation and exhaustion. These findings add to recently accumulated knowledge about the specific regulation of myeloid-biased HSCs in under stress conditions.

Our data indicate that C/EBPβ expression is regulated post-transcriptionally in HSPCs during stress. Such mechanisms enable rapid adjustment to meet an increased protein demand and are consequently of great significance when the cell is subjected to stress.³⁹  Furthermore, we found that upregulation and alteration of the translation of C/EBPβ isoforms in LT-HSCs in vivo were dependent on mTOR. mTOR preferentially upregulates LIP by accelerating the initial translation of the upstream open reading frame located between start codons for LAP/LAP* and LIP, as well as translation reinitiation from the downstream LIP-AUG codon.³³  Given that mTOR is activated in response to a variety of stress cues, this alteration in the translation of C/EBPβ isoforms may represent a general mechanism that regulates the behavior of HSCs under various kinds of stress.^40,41  In contrast, we and others have shown that interferons and BCR-ABL can upregulate Cebpb mRNA expression in HSCs in vivo.^18-20  The type, duration, or strength of the upstream signals may determine whether upregulation occurs at the transcriptional or post-transcriptional level. This issue should be investigated in future studies.

Taken together, our results indicate that stress-induced upregulation of C/EBPβ is critical to allow regenerating HSPCs to meet increasing myeloid demand. Upregulation of LIP in the early phase of regeneration likely amplifies the reservoir of HSPCs by facilitating the proliferation of LT-HSCs and their differentiation into MPPs, whereas LAP*/LAP induce subsequent myeloid differentiation at a later phase of regeneration. These findings provide insight into the pathophysiology of infection, inflammation, and regenerating hematopoiesis in response to myeloablative chemotherapies or HSC transplantation, which increase hematopoietic demand.

The authors thank Toshio Kitamura (University of Tokyo) for providing Plat-E cells and Schickwann Tsai (University of Utah) for providing EML cells. The authors also thank Shigekazu Nagata (Osaka University) for providing CD45.1 mice. The authors gratefully acknowledge Naoko Watanabe-Okochi and Mineo Kurokawa (University of Tokyo) for kindly providing the pGCDNsam-IRES-Kusabira-Orange (KuO), pGCDNsam-LIP-KuO, pGCDNsam-LAP-KuO, and pGCDNsam-LAP*-KuO vectors. The authors are also grateful to Yoko Nakagawa and Yoshiko Manabe for their excellent technical assistance.

This work was partly supported by KAKENHI Grants-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (25461415 and 18K08354 [H.H.], 25430149 [A.Y.], and 25112706 and 16K07171 [T.M.]); the Extramural Collaborative Research Grant from the Cancer Research Institute, Kanazawa University (H.H.); the Highway Program for Realization of Regenerative Medicine (JP17bm0504008) (H.H.); the Acceleration of Transformative Research for Medical Innovation from the Japan Agency for Medical Research and Development (H.H.); a grant from the Kyoto University Research Development Program “Ishizue” (H.H.); and a grant from the Kyoto University Foundation.

Contribution: A.S., N.K., and A.Y. designed and performed experiments, analyzed data, and wrote the manuscript; Y.H. performed experiments, analyzed data, and wrote the manuscript; A.T. performed the experiments and analyzed data; Y.M. performed experiments; T.M. supervised the project; and H.H. conceived, designed, and performed experiments; analyzed data; and wrote the manuscript.

Conflict-of-interest disclosure: H.H. received research funding from Kyowa Kirin, Bristol-Myers K.K., CSL Behring, Mitsubishi Tanabe Pharma, Sumitomo Dainippon Pharma, and Novartis Pharma. T.M. received research funding from Bristol-Myers K.K. The remaining authors declare no competing financial interests.

Correspondence: Hideyo Hirai, Laboratory of Stem Cell Regulation, School of Life Sciences, Tokyo University of Pharmacy and Life Sciences, 1432-1 Horinouchi, Hachioji, Tokyo 192-0392, Japan; e-mail: hirai@toyaku.ac.jp.