NF-κB Family Proteins Participate in Multiple Steps of Hematopoiesis through Elimination of Reactive Oxygen Species*

To examine the roles for NF-κB family proteins in hematopoiesis, we first expressed dominant negative Rel/NF-κB(IκBSR) in a factor-dependent cell line, Ba/F3. Although IκBSR neither affected thrombopoietin-dependent nor gp130-mediated growth, it suppressed interleukin-3- and erythropoietin-dependent growth at low concentrations. In addition, IκBSR enhanced factor-deprived apoptosis through the accumulation of reactive oxygen species (ROS). When expressed in normal hematopoietic stem/progenitor cells, IκBSR induced apoptosis even in the presence of appropriate cytokines by accumulating ROS. We also expressed IκBSR in an inducible fashion at various stages of hematopoiesis using the OP9 system, in which hematopoietic cells are induced to develop from embryonic stem cells. When IκBSR was expressed at the stage of Flk-1+ cells (putative hemangioblasts), IκBSR inhibited the development of primitive hematopoietic progenitor cells by inducing apoptosis through the ROS accumulation. Furthermore, when IκBSR was expressed after the development of hematopoietic progenitor cells, it inhibited their terminal differentiation toward erythrocytes, megakaryocytes, and granulocytes by inducing apoptosis through the ROS accumulation. These results indicate that NF-κB is required for preventing apoptosis at multiple steps of hematopoiesis by eliminating ROS.

To examine the roles for NF-B family proteins in hematopoiesis, we first expressed dominant negative Rel/NF-B (IBSR) in a factor-dependent cell line, Ba/F3. Although IBSR neither affected thrombopoietindependent nor gp130-mediated growth, it suppressed interleukin-3-and erythropoietin-dependent growth at low concentrations. In addition, IBSR enhanced factordeprived apoptosis through the accumulation of reactive oxygen species (ROS). When expressed in normal hematopoietic stem/progenitor cells, IBSR induced apoptosis even in the presence of appropriate cytokines by accumulating ROS. We also expressed IBSR in an inducible fashion at various stages of hematopoiesis using the OP9 system, in which hematopoietic cells are induced to develop from embryonic stem cells. When IBSR was expressed at the stage of Flk-1 ؉ cells (putative hemangioblasts), IBSR inhibited the development of primitive hematopoietic progenitor cells by inducing apoptosis through the ROS accumulation. Furthermore, when IBSR was expressed after the development of hematopoietic progenitor cells, it inhibited their terminal differentiation toward erythrocytes, megakaryocytes, and granulocytes by inducing apoptosis through the ROS accumulation. These results indicate that NF-B is required for preventing apoptosis at multiple steps of hematopoiesis by eliminating ROS.
The Rel/NF-B family consists of five members (c-Rel, p65 (RelA), RelB, p50 (NF-B1), and p52 (NF-B2)) (1,2). Each subunit shares a conserved N-terminal domain that encompasses sequences required for DNA binding, dimerization, and nuclear localization that is called the Rel homology domain. c-Rel, p65, and RelB each possess distinct transactivation domains in the C terminus, whereas p50 and p52, consisting of the only Nterminal domain, lack these intrinsic transactivation domains. Although these members form homodimers or heterodimers in various combinations, a heterodimer, p65/p50, is the most common form in mammalian cells and is specifically called NF-B. Under the unstimulated condition, the major proportion of Rel/NF-B is retained in cytoplasm by its inhibitor IB proteins as an inactive complex. Various stimuli such as pro-inflammatory cytokines and viruses activate the IB kinase complex and phosphorylate IB. Upon phosphorylation, IB is degraded by the ubiquitin/proteasome pathway. The released Rel/NF-B enters the nucleus, binds to target DNA elements, and initiates transcription. Until now, Rel/NF-B has been reported to promote the expression of over 150 target genes that regulate immune response, stress response, cell growth, or survival.
Mice homozygous for null mutations in genes encoding Rel/ NF-B subunits have revealed the unique roles of each family member protein (3,4). Although p65 is not essential for the generation of mature hematopoietic cells, p65 Ϫ/Ϫ mice are embryonic lethal at embryonic day 15 because of the massive apoptosis in the liver (5). Also, the cells from p65 Ϫ/Ϫ mice are susceptible to apoptosis induced by various reagents. In contrast, c-Rel, RelB, p50, and p52 are dispensable for embryogenesis. c-Rel is specifically expressed in lymphocytes, monocytes, granulocytes, and erythroid cells. Although the number of hematopoietic cells is normal in c-Rel Ϫ/Ϫ mice, both T-and Blymphocytes have defects in proliferative responses to various mitogens, isotype switching, and cytokine production (6 -9). Also, mutant mice lacking RelB, which is specifically expressed in dendric cells and B-lymphocytes, exhibit defects in acquired and innate immunity (10,11). Although p50 is ubiquitously expressed, p50 is not essential for hematopoiesis. However, p50 Ϫ/Ϫ mice reveal multiple defects in the immune system (12). Mature quiescent p50 Ϫ/Ϫ B-lymphocytes are susceptible to apoptosis and show poor proliferative responses to the CD40 ligand and lipopolysaccharide (8,12). Also, knockout mice for p52, of which expression is restricted to the epithelium of the stomach and selected areas of hematopoietic organs such as the thymic medulla and the marginal zone of the spleen, show abnormalities in splenic and lymph node structure (13,14). These lines of evidence suggest that Rel/NF-B family proteins are solely important for immune responses mediated by B-and T-lymphocytes. However, the redundancy among these family members was supposed to veil certain phenotypes in the single mutant mice. This hypothesis was supported by the subsequent findings that mice lacking two Rel/NF-B subunits exhibit novel phenotypes or exaggerated versions of those seen in the single mutants. Examples include the blockage of lymphocyte development in p50 Ϫ/Ϫ p65 Ϫ/Ϫ mice (15), greater severity of the inflammatory disease in p50 Ϫ/Ϫ RelB Ϫ/Ϫ mice than observed in RelB Ϫ/Ϫ mice (16), and impaired osteoclast and B-cell development in p50 Ϫ/Ϫ p52 Ϫ/Ϫ mice (14,17). Furthermore, the combined loss of c-Rel and p65 was shown to result in impaired erythropoiesis and deregulated expansion of granulocytes by transplantation experiments, indicating that, besides lymphopoiesis, Rel/NF-B is also required for normal hematopoiesis in the other lineages (18). However, at present, the precise roles of NF-B in hematopoiesis in the respective lineage remain undetermined. Also, molecular mechanism through which Rel/ NF-B family proteins regulate hematopoiesis is still unknown.
* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Reactive oxygen species (ROS) 1 including superoxide anions (O 2 . ), H 2 O 2 , organic peroxides, and hydroxyl radicals are byproducts of oxidative phosphorylation and constantly generated in all aerobic cells during normal metabolism (19,20). Although ROS are required for the physiologic function of the cells, excessive ROS cause apoptosis through several mechanisms such as activation of c-Jun N-terminal kinase, disruption of mitochondrial membrane potential (⌬⌿ m ), and/or direct activation of caspase cascades (21). Under normal circumstances, ROS are eliminated by antioxidant enzymes. The scavenger enzyme, superoxide dismutase (SOD) (MnSOD or Cu/ZnSOD), converts superoxide anions to H 2 O 2 . H 2 O 2 is subsequently detoxified by catalase or glutathione peroxidase (19,20). This redox regulation is essential for protecting cells from apoptosis.
In the present study, we expressed IBSR, which can inhibit the function of Rel/NF-B family proteins as a dominant negative mutant, in IL-3-dependent cell line Ba/F3 and normal hematopoietic stem/progenitor cells and assessed its effects on cytokine-dependent growth and survival. We also evaluated the roles for NF-B by expressing IBSR in an inducible manner at various stage of hematopoiesis using the OP9 system, in which hematopoietic cells are induced to develop from embryonic stem (ES) cells. We found here that Rel/NF-B family proteins play critical roles in preventing apoptosis at multiple steps of hematopoiesis by eliminating ROS.
Plasmids-The expression vector for IBSR was a gift from Dr. D. W. Ballard (Vanderbilt University School of Medicine, Nashville, TN); Mn-SOD was from K. Scharffetter-Kochanek (University of Cologne, Hamburg, Germany); and TRX was from Dr. J. Yodoi (Kyoto University, Kyoto, Japan).
Immunoblot Analysis-Isolation of total cellular lysate, electrophoresis, and immunoblotting were performed as described previously (26). Immunoreactive proteins were visualized with the chemiluminescence detection system (PerkinElmer Life Sciences).
Measurement of the Intracellular Glutathione-The amount of intracellular glutathione was measured using the glutathione assay kit (Cayman Chemical) according to the manufacturer's instruction.
Stable and Inducible Expression of IBSR in Ba/F3 Cells-A murine IL-3-dependent cell line Ba/F3 was cultured in RPMI1640 supplemented with 10% fetal bovine serum and 1 ng/ml of murine IL-3. To stably express IBSR in Ba/F3 cells, we transfected 30 g of pcDNA3.1 (Invitrogen) containing IBSR by electroporation (300 V, 960 microfarads). After the culture with 1.0 mg/ml of G418, several single clones and a mixed clone were subjected to further analyses. To inducibly express IBSR, we used a LacSwitch™ II inducible expression system (Stratagene) as described previously (27). In short, we initially transfected an expression vector for Lac repressor into Ba/F3 cells. After the culture with hygromycin, we selected one clone in which Lac repressor was most intensively expressed. We further transfected pOPRSVI containing IBSR into this clone and cultured with G418. In this system, the expression of the target cDNA is initiated when isopropyl ␤-D-thiogalactopyranoside (IPTG) is added to the culture medium. We selected several clones in which IBSR was effectively induced by IPTG and performed further experiments. We further introduce the EPO receptor, TPO receptor, and G-CSFR/gp130 consisting of the extracellular domain of G-CSFR and cytoplasmic domain of gp130 (28) into a Lacinducible IBSR clone using the puromycin (Puro)-resistant plasmid.
OP9 System to Develop Hematopoietic Cells from ES Cells-E14tg2a ES cells and OP9 stromal cells were maintained as described previously (29,30). To induce differentiation toward hematopoietic cells, ES cells were deprived of leukemia inhibitory factor and seeded onto confluent OP9 cells on six-well plates at a density of 10 4 cells/well in ␣-minimum essential medium supplemented with 20% fetal bovine serum. After 4.5 days, the cells were harvested by 0.25% trypsin-EDTA, and Flk-1 ϩ cells were purified by fluorescence-activated cell sorter sorting. The sorted cells were replated onto OP9 cells at a density of 10 5 cells/well of 6-well plate or 7-8 ϫ 10 5 cells/10-cm dish and cultured under the indicated conditions.
Tetracycline-regulated Inducible Expression of IBSR in ES Cells-To inducibly express IBSR in ES cells, we utilized a Tet-off system as reported previously (31), in which transcription of the target mRNA is initiated in response to the removal of Tet. Briefly, we initially introduced pCAG20 -1-tTA and pUHD10 -3-puro by electroporation (800 V, 3 microfarads). pCAG20 -1-tTA stably expresses the Tet-regulated transactivator under the control of the modified chicken ␤-actin promoter, and pUHD10 -3-puro contains the Puro-resistant gene downstream of the Tet-off-cytomegalovirus promoter. Thus, if the Tet-off system efficiently works in the transfected cells, these cells are expected to be Puro-resistant in the Tet Ϫ medium and Puro-sensitive in the Tet ϩ medium. According to this method, we selected one ES clone designated E14 by the culture with 1 g/ml of Puro and/or 1 g/ml of Tet, in which the Tet-regulatory system works most effectively. We further transfected pUHD10 -3-IBSR-IRES-GFP, which can inducibly express IBSR and GFP as a single mRNA through internal ribosome entry site in response to the removal of Tet, together with the neomycin-resistant plasmid pcDNA3.1-neo. After the culture with 0.2 mg/ml of G418 in the Tet ϩ medium, we selected several clones that can inducibly express GFP in response to the Tet deprivation. Subsequently, we examined the Tet-regulated expression of IBSR in the Tet ϩ and Tet Ϫ medium in these clones, and one clone was subjected to further analyses.
Preparation of Conditioned Media Containing High Titer Retrovirus Particles-The conditioned media containing high titer retrovirus particles were prepared as described previously (25). Briefly, pMSCV-IBSR-neo or an empty pMSCV-neo (Clontech) was transfected into an ecotropic packaging cell line Plat-E. After 12 h, the cells were washed and cultured for 48 h. Then the supernatant containing virus particles was collected, centrifuged, and concentrated by 50-fold in volume.
Retrovirus Transfection into Murine Hematopoietic Stem/Progenitor Cells-Bone marrow cells were harvested from 9 -12-week-old Balb/c mice pretreated with 150 mg/kg of 5-fluorouracil for 2 days. Lin Ϫ Sca-1 ϩ cells were isolated with MACS TM (Miltenyi Biotec) and cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum in the presence of murine (m) IL-3 (10 ng/ml), mSCF (50 ng/ml), mFLT3 ligand (L) (10 ng/ml), and human (h) IL-6 (10 ng/ml) for 48 h. Then the cells were cultured with 10% volume of conditioned media containing high titer retrovirus and 8 mg/ml of polybrene. After 24 h, the cells were washed and cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, the above described cytokines, and 1 mg/ml of G418 for 48 h.
Statistical Analysis-Statistical analysis was performed with Student's t test.

IBSR Partially Suppressed Cytokine-dependent Growth of
Ba/F3 Cells-To examine the roles for Rel/NF-B in the growth of hematopoietic cells, we stably expressed DN-NF-B (IBSR) in a murine IL-3-dependent cell line, Ba/F3; we designated two single clones as Ba/F3/IBSRCl1 and Ba/F3/IBSRCl2 and a mixed clone as Ba/F3/IBSRMix, respectively. We also prepared several Lac-inducible clones from Ba/F3 cells, in which IBSR was inducibly expressed by the IPTG treatment. As shown in Fig. 1A (left panel), IBSR was intensively expressed in Ba/F3/IBSRCl1, Ba/F3/IBSRCl2, and Ba/F3/IBSRMix, but not in Ba/F3/Mock. Also, the expression of IBSR was effectively induced by the IPTG treatment in Lac-inducible clones (Fig. 1A, right panel). To examine the effects of IBSR on cytokine-dependent growth of Ba/F3 cells, we further introduced EPO receptor, TPO receptor, and G-CSFR/gp130. As shown in Fig. 1B, these clones dose-dependently proliferated in response to IL-3, EPO, TPO, and the gp130 ligand (gp130L) (we here denote G-CSF as gp130L when it acts on G-CSFR/gp130), respectively. Although IPTG-induced IBSR did not affect TPO-or gp130L-dependent growth, it significantly inhibited IL-3-dependent growth at low concentrations (from 0.001 to 0.03 ng/ml) with a statistical significance (p Ͻ 0.05). Also, IBSR suppressed EPO-dependent growth at low concentrations (from 0.005 to 0.15 unit/ml) with a statistical significance (p Ͻ 0.05) (Fig. 1B).
IBSR Enhanced Factor-deprived Apoptosis in Ba/F3 Cells-Next, we examined the effects of IBSR on factor-deprived apoptosis in Ba/F3 cells. In DNA content analysis, a subdiploid fraction formed from apoptotic cells was hardly detectable in Ba/F3/Mock cells under IL-3-deprived conditions up to 24 h (Fig.  1C). In contrast, 55.2% of Ba/F3/IBSRMix cells led to apoptosis after 24-h IL-3 deprivation, indicating that IBSR enhanced fac- tor-deprived apoptosis in Ba/F3 cells. We also analyzed the changes of ⌬⌿ m during these cultures by flow cytometric analysis, in which the cells with the decreased ⌬⌿ m are detected as those with the increased green fluorescent intensity and decreased red fluorescent intensity in the lower right area (Fig. 1D). After 18 h of IL-3 depletion, ⌬⌿ m decreased in only 6.2% of Ba/F3/Mock, whereas it decreased in a substantial fraction (31.9%) of Ba/F3/IBSRMix. As for this mechanism, we found that the expression of Bcl-2 and Bcl-XL mRNA declined to an undetectable level in both Ba/F3/Mock and Ba/F3/IBSR after IL-3 deprivation (Fig. 1E). However, we did not detect an apparent difference in their expression levels between these clones that can explain IBSR-enhanced apoptosis.
IBSR Accumulated ROS in Ba/F3 Cells-Because Rel/ NF-B is involved in the metabolism of ROS (24) (Fig. 2B), suggesting that ROS may be involved in the IBSR-enhanced apoptosis.
IBSR Inhibited Cytokine-dependent Survival of Normal Hematopoietic Cells-We next analyzed the effects of IBSR on the growth and survival of normal hematopoietic cells. We introduced IBSR into murine Lin Ϫ Sca-1 ϩ cells with the retrovirus system. After the selection with G418, we further cultured these cells in the presence of IL-3, SCF, IL-6, and FLT3L for 7 days. Although only 11.9% of Mock-transfected cells led to apoptosis, 75.9% of IBSR-transfected cells underwent apoptosis (Fig. 3A). We also examined the effects of IBSR on hematopoietic stem/progenitor cells with colony assays. As shown in Fig. 3B, IBSR drastically reduced the number of CFU-Mix, BFU-E, and CFU-GM. These results indicated that although IBSR showed little effect on survival of Ba/F3 cells cultured with an appropriate cytokine, it disrupted cytokinedependent survival of normal hematopoietic cells. As for this mechanism, we found that although ⌬⌿ m decreased in only 18.7% of Mock-transfected cells, ⌬⌿ m was disrupted in 85.7% of IBSR-transfected cells after 5 days (Fig. 3C). Furthermore, flow cytometric analysis showed that total ROS, O 2 . , and H 2 O 2 accumulated in IBSR-transfected cells under the culture with SCF, IL-3, IL-6, and FLT3L (Fig. 3D, upper panel). Similarly, IBSR induced the ROS accumulation under the cultures with different cytokine combinations (Fig. 3D, lower panel). Regarding the mechanisms of IBSR-induced ROS accumulation, we found that the intracellular glutathione was reduced in IBSRtransfected cells (Fig. 3E). Also, the semiquantitative reverse transcriptase-PCR analysis showed that the expression of several ROS scavenger enzymes was suppressed in IBSR-trans-  (29). In this system, ES cells, which are deprived of leukemia inhibitory factor and cultured on OP9 cells, develop into Flk-1 ϩ cells (so-called hemangioblasts that have an ability to differentiate into both endothelial cells and hematopoietic cells) after 4.5-day cultures (Fig. 4A). To develop hematopoietic cells with high purity, we sorted Flk-1 ϩ cells by fluorescence-activated cell sorter and further cultured on OP9 cells (Fig. 4A). In this system, we inducibly expressed IBSR with the Tet-off system, in which the expression of IBSR is induced by the Tet removal from the culture medium (31). In addition, because our Tet-off vector is a bicistronic vector, it can express IBSR and GFP through internal ribosome entry site as a single mRNA in response to the Tet removal. We cultured ES clones each transfected with IBSR and a Mock vector (designated as E14/ IBSR and E14/Mock, respectively) without leukemia inhibitory factor on OP9 cells for 4.5 days. At this point, about 35% of the cultured cells were Flk-1 ϩ (data not shown), and we sorted these cells and cultured on OP9 cells with or without Tet for 48 h. As shown in Fig. 4B, the expression of GFP was induced in most of E14/IBSR cells under the culture without Tet. Also, immunoblot analysis showed that IBSR protein was effectively induced by Tet deprivation in E14/IBSR cells.
After the culture with Tet for 14 days, 29.2% of E14/IBSR cells became to be positive for an erythroid marker Ter119, 30.5% for a macrophage marker Mac1, and 18.7% for a Blymphoid marker B220 (Fig. 4C, left panel), indicating that our OP9 system could effectively yield erythroid cells, macrophages and B-lymphocytes. In contrast, when Tet was washed out from the culture medium, Ter119 ϩ , Mac ϩ , and B220 ϩ cells scarcely developed from E14/IBSR cells (Fig. 4C, right panel). In addition, we found that 74.7% of E14/IBSR cells led to apoptosis after 12-day culture without Tet, whereas only 19.7% of the cells underwent apoptosis in the presence of Tet (Fig. 4D). Also, we found that total ROS accumulated under the culture without Tet at days 6 and 10 (Fig. 4E). These results suggested that IBSR might inhibit the development of hematopoietic cells by inducing apoptosis through the ROS accumulation. Rel/NF-B Was Required for the Development of Primitive Hematopoietic Progenitors-In an attempt to determine at which step IBSR inhibited the development of hematopoietic cells, we initially analyzed the effect of IBSR on the development of primitive hematopoietic progenitors. We cultured Flk-1 ϩ cells in the presence or absence of Tet for 4 days and performed flow cytometric analysis at day 8.5 (Fig. 5A). After the culture with Tet, 66.4% of E14/IBSR cells were positive for CD45, 17.9% were positive for CD34, and 28.8% were positive for c-Kit (Fig. 5B), indicating that a substantial fraction of the cultured cells developed into primitive hematopoietic progenitors. Next, we cultured E14/IBSR cells without Tet in the presence or absence of MCI and evaluated the effects of IBSR in the GFP-positive fraction (GFP-PF), because IBSR was supposed to be effectively induced in this fraction. In the absence of MCI, CD45 ϩ , CD34 ϩ , and c-Kit ϩ fractions were severely reduced in the GFP-PF as compared with those after the culture with Tet; the relative percentages in the GFP-PF were: CD45 ϩ , 17.1%; CD34 ϩ , 7.6%; and c-Kit ϩ , 14.3% (shown in parentheses in Fig. 5B). However, when MCI was added, MCI significantly restored these fractions in the GFP-PF: CD45 ϩ , 54.1%; CD34 ϩ , 28.6%; and c-Kit ϩ , 21.8%. Regarding this mechanism, we found that although only 5.39% of the cultured cells were annexin V ϩ after the culture with Tet, 57.1% of the cells were Annexin V ϩ in the GFP-PF after the culture without Tet and MCI, which was reduced to 12.0% by MCI. These results indicated that IBSR inhibited the development of primitive hematopoietic progenitors by inducing apoptosis, largely through the ROS accumulation.
Rel/NF-B Was Also Required for the Subsequent Maturation of Hematopoietic Cells from Primitive Hematopoietic Progenitors-Next, we examined the roles for Rel/NF-B in the subsequent maturation stage of hematopoiesis. For this purpose, we cultured Flk-1 ϩ cells with Tet for 4.5 days and confirmed the development of primitive hematopoietic progenitors from E14/ IBSR cells by flow cytometry: CD45 ϩ , 51.6%; CD34 ϩ , 10.1%; and c-Kit ϩ , 27.2% (flow cytometric data not shown). Then we further cultured these E14/IBSR cells under the indicated conditions up to day 12 (Fig. 6A). As shown in Fig. 6B, substantial fractions of E14/IBSR cells cultured with Tet developed into Gr-1 ϩ (myeloid), Ter119 ϩ (erythroid), and CD41 ϩ (megakaryocytic) cells after the respective cultures: the relative proportion of Gr-1 ϩ cells in the GFP-negative fraction, where IBSR was not expressed, 37.2%; Ter119 ϩ , 49.2%; and CD41 ϩ , 77.6% (shown in parentheses). In contrast, when E14/ IBSR cells were cultured without Tet and MCI, the relative proportions of these mature cells severely decreased in the GFP-PF: Gr-1 ϩ , 23.7%; Ter119 ϩ , 14.1%; CD41 ϩ , 29.8% (shown in parentheses). However, MCI almost completely restored these fractions: Gr-1 ϩ , 43.8%; Ter119 ϩ , 52.4%; and CD41 ϩ , 75.6% (shown in parentheses). As for this reason, we found that IBSR induced apoptosis in 51.8% of E14/IBSR cells cultured without Tet and that this apoptotic fraction was reduced to 14.1% by MCI (Fig. 6B). These results indicated that Rel/ NF-B was also necessary for hematopoietic cells to undergo terminal differentiation toward various lineages. DISCUSSION Rel/NF-B family proteins play important roles in various biologic phenomena such as immune responses, stress responses, and inflammation. Also, knockout mice for each subunit revealed that these factors crucially control cell growth and survival of B-and T-lymphocytes (8,9,12). In addition, recent studies have demonstrated that these factors are constitutively activated in various types of hematologic malignancies, including lymphomas (Hodgkin's disease, adult T-cell leukemia/lym- FIG. 4. Effects of IBSR on the development of hematopoietic cells. A, experimental design 1 using the OP9 system. ES cells were deprived of leukemia inhibitory factor and cultured on OP9 cells for 4.5 days. Then Flk-1 ϩ cell were sorted, replated onto OP9 cells, and cultured with IL-3, EPO, and SCF with or without Tet for the time indicated. B, Flk-1 ϩ cell were cultured for 48 h. Then the expression of GFP was examined by flow cytometric analysis, and the induction of IBSR was examined by immunoblot analysis in the indicated clones. C, after 14-day cultures with or without Tet, the expression of Ter119, Mac1, and B220 was examined by flow cytometric analysis. The percentage of the positive fraction is indicated. D, DNA content of the cells was analyzed by propidium iodide staining at days 4.5 and 12. The percentage of apoptotic cells is indicated. E, total ROS were detected with RedCC-1 by flow cytometric analysis at days 4.5, 6, and 10. phoma, Burkitt's lymphoma, and anaplastic lymphoma), multiple myeloma, acute myeloid leukemia, and acute lymphoblastic leukemia, thereby causing these diseases and/or affecting their pathophysiologic aspects (32). These results suggest that the appropriate regulation of NF-B activity is required for normal hematopoiesis.
In a previous study, IL-3 and granulocyte-macrophate colony-stimulating factor were reported to activate NF-B in hematopoietic cells (33). Also, EPO was shown to activate NF-B through the activation of Jak2 (34). In these studies, NF-B was considered to mainly mediate anti-apoptotic functions of these growth factors. On the other hand, NF-B has been shown to regulate cell growth through the induction c-myc and cyclin D1 in other cell types (35,36). Supporting these results, B-and/or T-lymphocytes obtained from c-Rel Ϫ/Ϫ or p50 Ϫ/Ϫ mice showed the decreased proliferative response to various mitogens (8,9,12). Moreover, constitutively activated NF-B was reported to plays a crucial role in regulating the growth of Hodgkin's lymphoma cells (37,38). However, in the present study, we found that IBSR neither influenced TPO-nor gp130Ldependent growth and showed only a limited degree of growth inhibitory effect on IL-3-and EPO-dependent growth of Ba/F3 cells. Also, when we prevented IBSR-induced apoptosis by MCI, cytokine-dependent growth of normal hematopoietic cells was hardly suppressed by IBSR (data not shown). Therefore, it was speculated that, although NF-B is important for the growth regulation in some aspects of hematopoiesis such as inflammatory responses, immune responses, and oncogenesis, it may be dispensable for cytokine-dependent growth of normal hematopoietic cells.
As for the mechanisms of Rel/NF-B-mediated cell survival, Rel/NF-B has been shown to transcriptionally regulate the expression of anti-apoptotic Bcl-2 family members Bcl-XL and A1/Bfl1 and caspase inhibitors c-IAP1 and c-IAP2 in various cell types (1,2). Also, Rel/NF-B promotes the expression of Bcl-2 in EB virus-transformed B lymphocytes, probably through the indirect mechanism (39). In addition, NF-B has been reported to prevent apoptosis by reducing the oxidative stress through the induction of the ROS scavenger enzyme, MnSOD (24,40). In agreement with these findings, we found here that IBSR down-regulated the expression of Bcl-XL, A1, Bcl-2, and MnSOD in normal hematopoietic cells. Furthermore, regarding the mechanism of the ROS accumulation in IBSR-introduced cells, we found that the amount of the intracellular glutathione and the expression of glutathione peroxidase and TRX were reduced by IBSR in these cells as well as that of MnSOD, suggesting that NF-B would regulate the expression of these molecules. In addition, our experiments showed that catalytic antioxidants efficiently cancelled IBSRenhanced apoptosis. This result suggests that the antioxidant activity of NF-B would be more important to protect hematopoietic cells from apoptosis, whereas we should be still aware that Bcl-2 family members also play a certain role in NF-Bmediated cell survival.
In this study, MCI, NAC, and TRX but not MnTBAP efficiently cancelled IBSR-enhanced apoptosis, suggesting that NO Ϫ and H 2 O 2 were more toxic to normal hematopoietic stem/ progenitors cells than O 2 . . However, based on the fact that MnSOD Ϫ/Ϫ mice exhibited severe anemia caused by ineffective erythropoiesis, O 2 . was also assumed to be toxic to erythroid progenitor cells (41,42). Thus, further studies are required to assign the roles of O 2 . in erythroid progenitor cells, especially in both normal and pathologic states of erythropoiesis. As for the roles for Rel/NF-B in hematopoietic stem cells, Pyatt et al. (33) showed that Rel/NF-B was latent under physiologic conditions and that p65, c-Rel, and p50 but not p52 or RelB were activated and bound to the target DNA in 12-O-tetradecanoylphorbol-13-acetate-stimulated human CD34 ϩ CD19 Ϫ cells using gel shift assays. In addition, they demonstrated that anti-apoptotic activity of Rel/NF-B was indispensable for the colony formation from CD34 ϩ cells using the membrane-permeable peptide that can inhibit Rel/NF-B activity. Furthermore, we found that Rel/NF-B plays a crucial role in the development of primitive hematopoietic progenitors by preventing apoptosis through ROS elimination. Together, these results suggest that Rel/NF-B activity is indispensable for both development and maintenance of hematopoietic stem/progenitors. With regard to the function of Rel/NF-B in hematopoietic progenitor cells, Grossman et al. (18) previously reported that the number of hematopoietic progenitor cells that can generate CFU-spleen in the recipient mice was reduced in the fetal liver of c-Rel Ϫ/Ϫ p65 Ϫ/Ϫ mice. However, in their analysis, fetal liver cells obtained from c-Rel Ϫ/Ϫ p65 Ϫ/Ϫ mice yielded the similar number, size, and appearance of various colonies (CFU-granulocyte, CFU-granulocyte macrophage, CFU-macrophage, CFU-eosinphil, CFU-megakaryocyte, CFU-erythroid, CFU-E/Meg, and CFU-Mix) to control cells in colony assays. This result seems to be inconsistent with our result that IBSR drastically reduced the number of CFU-Mix, BFU-E, and CFU-granulocyte macrophage. Thus, further studies are required to define the roles of NF-B in clonogenic growth and survival of hematopoietic progenitor cells.
Regarding the roles for Rel/NF-B in terminal differentiation of hematopoietic cells, Zhang et al. (43) previously showed that p50, p52, and p65 were highly expressed in erythroid progenitors and bound to the target DNA in the promoters of c-myc and c-myb genes, suggesting a possibility that Rel/NF-B would play some role in erythropoiesis. Also, Grossman et al. (18) demonstrated that mice transplanted with c-Rel Ϫ/Ϫ p65 Ϫ/Ϫ hematopoietic cells revealed severe anemia and granulocytosis as compared with those transplanted with normal cells, suggesting that Rel/NF-B is a positive regulator of erythropoiesis and a negative regulator of granulopoiesis. However, we found here that Rel/NF-B activity was required for terminal differentiation of erythrocytes, granulocytes, and megakaryocytes. Although our result on the function of NF-B in granulopoiesis was at variance with their finding, we speculate that these conflicting data may originate from the difference in the experimental systems. That is, our result was obtained from the in vitro experiments using the OP9 system, whereas their finding was from the in vivo transplantation assays. Alternatively, it is also possible that functional roles for Rel/NF-B in granulopoiesis in vivo might be essentially different according to the state of hematopoiesis, such as the steady state of hematopoiesis, a recovery period after transplantation, and the acute phase of inflammation.
In summary, we demonstrated here that NF-B family proteins play crucial roles at multiple stages in hematopoiesis by preventing apoptosis through the elimination of ROS. Further studies focusing on the regulation of cell survival and death by NF-B/ROS system would undoubtedly provide useful information to understand both normal and ineffective hematopoiesis FIG. 6. Effects of IBSR on terminal differentiation of hematopoietic cells. A, experimental design 3. Sorted Flk-1 ϩ cells were cultured with IL-3, EPO, and SCF in the presence of Tet on OP9 cells for 4 days. Then the cells were washed, replated onto OP9 cells, and cultured under the indicated conditions up to day 12. B, after 12-day cultures under the indicated conditions, the cells were subjected to flow cytometric analysis on the expression Gr-1, Ter119, and CD41, respectively. Also, the intensity of GFP and reactivity to phycoerythrinlabeled annexin V were analyzed by flow cytometry. The percentage of the each fraction is indicated. The relative frequency of the upper panel in the GFP-negative fraction (GFP-NF) or GFP-PF is shown in parentheses. and to construct useful therapeutic strategies against hematologic malignancies.