Keap1‐Nrf2 system regulates cell fate determination of hematopoietic stem cells

Save this PDF as:

Size: px
Start display at page:

Download "Keap1‐Nrf2 system regulates cell fate determination of hematopoietic stem cells"


1 Keap1-Nrf2 system regulates cell fate determination of hematopoietic stem cells Shohei Murakami1, Ritsuko Shimizu2, Paul-Henri Romeo3, Masayuki Yamamoto1* and Hozumi Motohashi4* 1 Department of Medical Biochemistry, Tohoku University Graduate School of Medicine, 2-1 Seiryo-machi, Aoba-ku, Sendai, Miyagi , Japan 2 Department of Molecular Hematology, Tohoku University Graduate School of Medicine, 2-1 Seiryo-machi, Aoba-ku, Sendai, Miyagi , Japan 3 CEA, ircm, Laboratoire de Recherche sur la R eparation et la Transcription dans les Cellules Souches, 18 route du Panorama, Fontenay-aux-Roses Cedex, BP , France 4 Department of Gene Expression Regulation, Institute of Development, Aging and Cancer, Tohoku University, 4-1 Seiryo-machi, Aoba-ku, Sendai, Miyagi , Japan Nrf2 is a major transcriptional activator of cytoprotective genes against oxidative/electrophilic stress, and Keap1 negatively regulates Nrf2. Emerging works have also suggested a role for Nrf2 as a regulator of differentiation in various cells, but the contribution of Nrf2 to the differentiation of hematopoietic stem cells (HSCs) remains elusive. Clarifying this point is important to understand Nrf2 functions in the development and/or resolution of inflammation. Here, we established two transgenic reporter mouse lines that allowed us to examine Nrf2 expression precisely in HSCs. Nrf2 was abundantly transcribed in HSCs, but its activity was maintained at low levels due to the Keap1-mediated degradation of Nrf2 protein. When we characterized Keap1-deficient mice, their bone marrow cells showed enhanced granulocyte-monocyte differentiation at the expense of erythroid and lymphoid differentiation. Importantly, Keap1-null HSCs showed lower expression of erythroid and lymphoid genes than did control HSCs, suggesting granulocyte-monocyte lineage priming in Keap1-null HSCs. This abnormal lineage commitment was restored by a concomitant deletion of Nrf2, demonstrating the Nrf2-dependency of the skewing. Analysis of Nrf2-deficient mice revealed that the physiological level of Nrf2 is sufficient to contribute to the lineage commitment. This study unequivocally shows that the Keap1-Nrf2 system regulates the cell fate determination of HSCs. Introduction The Keap1-Nrf2 system plays a central role in cellular protection against xenobiotics and oxidative stress. Nrf2 is a transcriptional activator for the genes encoding phase II detoxification and antioxidative stress enzymes, such as NAD(P)H quinone oxidoreductase 1 (Nqo1), thioredoxin reductase 1 (Txnrd1) and catalytic and regulatory subunits of glutamate-cysteine ligase (Gclc and Gclm, respectively; Itoh et al. 1997; Chan & Kwong 2000; Sakurai et al. 2005). Under homeostatic conditions, Keap1 binds to Communicated by: Shunsuke Ishii *Correspondence: or the Neh2 domain of Nrf2 and suppresses Nrf2 activity by promoting its poly-ubiquitination and subsequent degradation (Itoh et al. 1999; Wakabayashi et al. 2003). Upon exposure to oxidative or electrophilic stress, Keap1 inactivation allows Nrf2 to bind to antioxidant/electrophile response elements (AREs/EpREs) and to activate cytoprotective genes (Dinkova-Kostova et al. 2002; Motohashi & Yamamoto 2004; Kobayashi et al. 2009). Nrf2-deficient (Nrf2 / ) mice are sensitive to diverse cytotoxic stresses, such as acetaminophen, benzo[a]pyrene and lipopolysaccharide (Enomoto et al. 2001; RamosGomez et al. 2001; Thimmulappa et al. 2006). Nrf2 in hematopoietic cells has been shown to play important roles in the protection against acute inflammation (Ishii et al. 2005; Kong et al. 2011). Nrf2 / DOI: /gtc

2 S Murakami et al. mice are susceptible to elastase-induced pulmonary inflammation, and the mice can be rescued by the bone marrow (BM) transplantation of wild-type cells (Ishii et al. 2005). The myeloid cell-specific deletion of Nrf2 augments the susceptibility to sepsis (Kong et al. 2011). These studies suggest that macrophage activity is regulated by Nrf2, and of the various hematopoietic cells harboring Nrf2, macrophages play critical roles in the protection against inflammation. In addition to the changes in the characteristics of macrophages by Nrf2 deletion, the macrophage influx into inflammatory sites is delayed in Nrf2 / mice, leading to the persistence of neutrophils (Itoh et al. 2004). Although the critical contributions of Nrf2 in differentiated hematopoietic cells have been addressed, it remains to be clarified how Nrf2 participates in the differentiation of hematopoietic cells, especially hematopoietic stem cells (HSCs). Understanding this process is particularly important, as the Keap1-Nrf2 system has been emerging as a regulator of the development and differentiation of various tissues and cell types. For instance, loss of Nrf2 activity affects the differentiation of erythroid cells, megakaryocytes and dendritic cells (Motohashi et al. 2010; Kawatani et al. 2011; Aw Yeang et al. 2012). Importantly, the intracellular redox balance controlled by the Keap1-Nrf2 system appears to be strongly associated with the regulation of differentiation in these cases. Another intriguing observation is that Nrf2 appears to regulate differentiation program by directly activating the expression of genes involved in tissue-specific signal transduction. For instance, Nrf2 promotes the regeneration of the liver by inducing Notch1 (Beyer et al. 2008; Wakabayashi et al. 2010). Nrf2 also regulates adipocyte differentiation by enhancing the expression of Peroxisome proliferator-activated receptor c (Pparg), CCAAT/enhancer binding protein a (Cebpa) and b (Cebpb) (Pi et al. 2010; Hou et al. 2012; ). Increasing attention has been paid as to how the Keap1-Nrf2 system determines the cell differentiation. To clarify the functional contribution of the Keap1-Nrf2 system to the differentiation of HSCs in this study, we generated two new transgenic mouse lines that allowed us to examine the specific expression profile of Nrf2 and the unique contribution of Nrf2 to hematopoietic lineages. Analyses of these transgenic mice provided firm lines of evidence that the Keap1-Nrf2 system is highly functional in hematopoietic stem and progenitor cells (HSPCs) and granulocyte-monocyte lineage cells. We also examined the Nrf2 function in hematopoiesis by exploiting three loss-of-function lines of mice, that is, Nrf2 knockout mice, conditional Keap1 knockout mice and Nrf2::Keap1 double-mutant mice. Through the analyses of these lines of mice, we delineated that Nrf2 activation skews hematopoietic differentiation toward the granulocyte-monocyte lineage, which appears to emerge as early as at the HSC stage. Our results thus show that Nrf2 activation in HSCs supports the production of granulocytes and monocytes, particularly under stressed conditions, such as inflammation. Results Nrf2 is abundantly expressed in granulocytes, monocytes and HSPCs To determine which hematopoietic lineages exhibit the highest dependency on Nrf2, we initially examined the expression level of Nrf2 in each cell lineage defined by specific surface markers (Fig. 1A). To monitor the mrna level of Nrf2, we established a transgenic mouse line with the tdtomato fluorescence gene as a reporter using a bacterial artificial chromosome (BAC) clone, RP23-374O4 (Fig. 1B). The cdna encoding tdtomato was inserted into the first exon of the Nrf2 gene in the BAC clone such that the Neh2 domain of Nrf2 was eliminated in this construct. Two independent lines were obtained for the Nrf2-tdTomato BAC transgenic (Nrf2-BAC-TG) mouse, and the one line with the highest reporter gene expression was analyzed in this study. We confirmed the reproducibility of the analyses by using the other mouse line. In mature stages, Mac-1 + and Gr-1 + cells showed relatively strong tdtomato fluorescence, whereas only weak tdtomato fluorescence was detected in Ter119 + (Fig. 1C). Intriguingly, HSC fractions (Lin Sca-1 + c-kit + (LSK) cells) and hematopoietic progenitor cell (HPC) fractions (Lin Sca-1 c-kit + (LK) cells) also exhibited intense tdtomato fluorescence compared with Ter119 + cells. The fluorescent intensity was well correlated with the mrna abundance of Nrf2 in these cell fractions (Fig. 1D). In contrast, B220 + cells expressed significantly lower level of tdtomato fluorescence than Ter119 + cells, but Nrf2 mrna in B220 + cells was higher than that in Ter119 + cells. These results suggest that the regulatory regions contained in RP23-374O4 are sufficient for Nrf2 expression in hematopoietic cells except for lymphocytes, particularly B cells, and indicate that Nrf2 is richly expressed in granulocytemonocyte lineage cells and HSPCs. 240

3 Keap1-Nrf2 system in hematopoiesis (A) (B) (C) (D) Figure 1 Nrf2 is predominantly expressed in Lin Sca-1 + c-kit +, Lin Sca-1 c-kit + and myeloid cells. (A) Definition of hematopoietic lineages and HSPC fractions used in this study. (B) Schematic structure of the murine Nrf2 gene locus and the strategy of bacterial artificial chromosome (BAC) mutagenesis for generating the Nrf2-BAC-tdTomato transgene. The genomic region contained in the BAC clone, RP23-374O4, is represented with a double-ended arrow, which spans from 104 kbp to +54 kbp relative to the transcription start site (+0). The murine Nrf2 locus is shown at the top, and the final product, Nrf2-BAC-TG, is shown at the bottom. The white boxes represent Nrf2 exons, and the black box indicates the Hnrpa3 gene. The red box and the triangles indicate tdtomato cdna and the FLP recombination target sequences, respectively. (C) tdtomato expression in hematopoietic cell fractions. The bar graphs indicate the differences in the fluorescence intensities between Nrf2-BAC-TG and wild-type (background) cells, presented as the means SD (left panel, n = 10; right panel, n = 4). (D) Nrf2 mrna expression in hematopoietic cell fractions. The expression levels were normalized to rrna expression. The expression level in Ter119 + cells was set to one. Values are presented as the means SD (n = 5-6). *P < 0.01; **P < 0.005; ***P < Keap1 regulates Nrf2 activity in HSCs As the Nrf2 protein is constantly degraded in a Keap1-dependent manner under homeostatic conditions (Itoh et al. 1999; Motohashi & Yamamoto 2004), the protein level of Nrf2 was expectedly maintained at low levels even in cell fractions expressing abundant Nrf2 mrna, such as LSK and LK cells (Fig. 1C,D). Thus, we next investigated the Nrf2 accumulation in hematopoietic cells, particularly in HSCs. To this end, we generated a second transgenic reporter mouse, MGRD-Neh2-TG, for 241

4 S Murakami et al. monitoring the Keap1-dependent regulation of Nrf2 accumulation. In this mouse, EGFP and Neh2- tdtomato fusion proteins were expressed under the control of the MafG regulatory domain (MGRD; (A) Fig. 2A). The EGFP protein is constitutively stable irrespective of oxidative stress and thus used to monitor the MGRD activity. In contrast, as Neh2 is an N-terminal domain of Nrf2 and serves as a degron (C) (B) (D) (E) (F) Figure 2 Nrf2 activation status in hematopoietic cells. (A) Schematic structure of the MafG regulatory domain (MGRD)-Neh2-td Tomato-IRES-EGFP (MGRD-Neh2-TG) transgene. The murine MafG allele is shown at the top. The EGFP and Neh2-tdTomato fusion proteins are expressed under the control of the murine MafG regulatory element. The white boxes represent the MafG exons, and the brown, dark red, black and green boxes represent the cdna for the Neh2 domain, tdtomato cdna, IRES sequence and EGFP cdna, respectively. (B) Illustration of Keap1-dependent stabilization of the Neh2-tdTomato protein. Under homeostatic conditions, Neh2-tdTomato is degraded in a Keap1-dependent manner, but on exposure to oxidative stress, stabilized concomitantly with Keap1 inactivation. In MGRD-Neh2-TG mice, the cells are constantly labeled with EGFP, and those with Keap1 inactivation are also labeled with tdtomato. (C) tdtomato and EGFP fluorescence in Lin Sca-1 + c-kit + (LSK) cells from MGRD-Neh2-TG (TG) and wild-type (WT) mice. Lineage-negative cells were cultured in the presence of DMSO (0.01%) or DEM (100 lm) with DMSO (0.01%) for 9 h and analyzed using flow cytometry. (D) Expression of Nrf2 target genes in LSK cells of MGRD-Neh2-TG and wild-type mice (top) as well as Nrf2 / and Nrf2 +/+ mice (bottom). The expression of Nqo1, Gclc and Txnrd1 was normalized to rrna levels. The representative data of three (top) and two (bottom) independent experiments are shown, and values are presented as the means SD of triplicate measurements in quantitative RT-PCR. (E) tdtomato and EGFP fluorescence in LSK Flt3 CD34 cells from Keap1 CKO::MGRD-Neh2-TG (Keap1 CKO::TG) and Control::MGRD-Neh2-TG (Control::TG) mice. (F) Expression of Keap1 and Nrf2 target genes in LSK Flt3 cells from Keap1 CKO or Control mice. The expression levels were normalized to Gapdh expression. The expression levels in Control mice were set to one. Values are presented as the means SD (n = 3 4). # P = 0.05; *P < 0.05; **P <

5 Keap1-Nrf2 system in hematopoiesis for Keap1-mediated degradation (Itoh et al. 1999), the Neh2-tdTomato protein is degraded in a Keap1- dependent manner under unstressed conditions but is stabilized upon Keap1 inactivation by oxidative/electrophilic stress, thereby labeling cells with tdtomato fluorescence (Fig. 2B). Using the MGRD-Neh2-TG mice, we examined the tdtomato fluorescence in HSC fractions (i.e., LSK cells). As expected, EGFP fluorescence was constantly detected regardless of treatment with an Nrf2 inducer, diethyl maleate (100 lm DEM in 0.01% DMSO), or vehicle (DMSO, 0.01%), but tdtomato fluorescence was only detected upon treatment with DEM. Importantly, the tdtomato fluorescence was accompanied by increased expression of Nqo1, Gclc, and Txnrd1 (Fig. 2D). We examined whether the increased expression of the cytoprotective enzyme genes in the LSK fraction was dependent on Nrf2 using Nrf2 / mice. The results clearly showed that the expression of these three genes was barely induced by DEM in the LSK fraction of Nrf2 / mice, confirming that tdtomato fluorescence in MGRD-Neh2-TG cells accurately reflects Nrf2 activity. These results suggest that HSCs possess the functional Keap1-Nrf2 system. To further validate the presence of the functional Keap1-Nrf2 system in HSCs, we examined the effect of Keap1 gene deletion. We crossed MGRD-Neh2- TG mice with Keap1 conditional knockout (Keap1 F/F ::Mx1-Cre; Keap1 CKO) mice and Keap1 F/+ or Keap1 F/+ ::Mx1-Cre (Control) mice to obtain Keap1 CKO::TG (Keap1 F/F ::Mx1-Cre::MGRD-Neh2-TG) mice and Control::TG (Keap1 F/+ ::MGRD-Neh2-TG or Keap1 F/+ ::Mx1-Cre::MGRD-Neh2-TG) mice, respectively. We subsequently analyzed tdtomato fluorescence in their BM cells after polyinosinic-polycytidylic acid (pipc) treatment. The tdtomato fluorescence was not observed in the LSK Flt3 CD34 cells, a long-term HSC (LT-HSC) fraction, of Control::TG mice but was detectable in the LSK Flt3 CD34 cells of Keap1 CKO::TG mice (Fig. 2E). We obtained the similar results in other fractions, including whole LSK cells, LK cells and Mac-1 + Gr-1 + cells. Thus, Keap1 regulates Nrf2 activation in granulocyte-monocyte lineage cells and HSPCs, even in the most immature stage, LT-HSCs. Conditional disruption of Keap1 gene in BM cells As tdtomato fluorescence was undetectable in BM cells of MGRD-Neh2-TG and Control::TG mice without stimuli (Fig. 2C,E), the steady-state level of Nrf2 activity was low in hematopoietic cells. Hence, to clarify the Nrf2 contribution to hematopoiesis, we attempted to conduct genetic activation of Nrf2 by using the Keap1 CKO mice, in whose hematopoietic cells Nrf2 is constitutively active. As almost all LSK Flt3 CD34 cells of Keap1 CKO::TG mice expressed an increased level of tdtomato fluorescence (Fig. 2E), we judged that the treatment with pipc efficiently deleted the Keap1 gene in the Keap1 CKO mice. The expression analysis of the Keap1 and Nrf2 target genes, Nqo1 and Txnrd1, also supported the efficient Keap1 deletion (Fig. 2F). Therefore, we investigated hematopoiesis in Keap1 CKO mice compared with Control mice. Keap1 deletion enhances granulocyte-monocyte differentiation To assess the effects of Nrf2 activation on hematopoietic differentiation, we first examined hematopoietic indices of Keap1 CKO mice. Through the analysis of peripheral blood, we observed a significant reduction in the red blood cell count, hemoglobin content, and hematocrit in Keap1 CKO mice (Table 1). We also analyzed the BM cells by flow cytometry and found that the whole BM cellularity was comparable between Keap1 CKO and Control mice (Fig. 3A), but that Mac-1 + Gr-1 + cells were increased and the other lineage cells were reduced in Keap1 CKO mice (Fig. 3B). Of note, this skewing was also observed in the lineage-committed progenitor fraction. The LK cells in Keap1 CKO mice contained more granulocyte-monocyte progenitors (GMPs) and fewer megakaryocyte- Table 1 Peripheral blood counts in Control and Keap1 CKO mice Control Keap1 CKO P value WBC (10 2 /ll) NS RBC (10 4 /ll) P < HGB (g/dl) P < HCT (%) P < MCV (fl) NS MCH (pg) NS PLT (10 4 /ll) NS Values are presented as the means SD of Control (n = 10) and Keap1 CKO (n = 8) mice. WBC, white blood cell; RBC, red blood cell; HGB, hemoglobin; HCT, hematocrit; MCV, mean corpuscular volume; MCH, mean corpuscular hemoglobin; PLT, platelet; NS, not significant. 243

6 S Murakami et al. (A) (B) (C) (D) (E) (F) (G) Figure 3 Keap1 deletion enhances granulocyte-monocyte differentiation. (A C) Whole bone marrow (WBM) cellularity (A, n = 5-7), ratios of each lineage in BM cells (B, n = 4-8) and ratios of each lineage-committed progenitor in the LK fraction (C, 4 5) in Keap1 CKO and Control mice. Values represent means SD. CMP (LK FccR CD34 + ), common myeloid progenitor; GMP (LK FccR + CD34 + ), granulocyte-monocyte progenitor; MEP (LK FccR CD34 ), megakaryocyte-erythroid progenitor. (D) Lineage skewing toward granulocytes and monocytes in Keap1 CKO mice. (E-G) Noncompetitive BM transplantation assay of BM cells ( ) from Keap1 CKO::TG and Control::TG mice. BM cells of the recipients were analyzed at weeks post-transplantation. (E) tdtomato and EGFP fluorescence in the LSK Flt3 CD34 fractions of the recipients. (F) Ratios of each lineage in the BM cells. Values represent means SD (n = 8). (G) Ratios of each lineage-committed progenitor in donor-derived LK cells. Values represent means SD (n = 3). *P < 0.05; **P < 0.005; ***P < erythroid progenitors (MEPs) than those in Control mice (Fig. 3C). These observations suggest that Keap1 deficiency preferentially skews the hematopoietic differentiation toward the granulocytemonocyte lineage at the expense of the erythroid and lymphoid lineages and, importantly, that the cell fate determination due to the absence of Keap1 occurs at a more immature stage than HPCs (Fig. 3D). To determine whether the Keap1 deletion intrinsically causes these phenotypes, we carried out a noncompetitive BM transplantation assay. We confirmed that the LT-HSC fraction (i.e., LSK Flt3 CD34 cells) was almost completely replaced with donor-derived EGFP-positive cells in the recipients, indicating that all hematopoietic cells in the recipients were derived from the donor cells (Fig. 3E). The lineage skewing of BM cells and HPCs in Keap1 CKO mice was reproducible in the recipients (Fig. 3F,G), demonstrating that the effects of Keap1 deficiency occur in a cell-autonomous manner. 244

7 Keap1-Nrf2 system in hematopoiesis Keap1-deficient HPCs possess enhanced granulocyte-monocyte differentiation capacity To further verify the distorted differentiation in Keap1-deficient HPCs, we carried out an in vitro colony-forming cell (CFC) assay using LK cells. While the total number of colonies did not change significantly between Keap1-deficient and control cells, Keap1-deficient LK cells formed more CFU-M colonies than did control cells (Fig. 4A, asterisk; Control, ; Keap1 CKO, ). The number of CFU-G colonies also tended to increase in the Keap1-deficient LK cells. (A) (C) (B) We carried out an expression analysis of the lineage-specific genes. As the segregation of the megakaryocyte-erythroid and granulocyte-monocyte lineages occurs in the LK cell levels (Akashi et al. 2000), Gata1 expression was examined as a marker of the erythroid lineage, and Cebpa and Cebpb were examined as markers of the granulocyte-monocyte lineage. The Keap1-deficient LK cells showed downregulation of Gata1 and up-regulation of Cebpa and Cebpb, although the difference of Cebpa expression did not reach the statistical significance (Fig. 4B). Thus, Keap1 deficiency skews the HPC differentiation toward the granulocyte-monocyte lineage in a cell-autonomous manner. We then wished to assess the differentiation potential of Keap1-deficient HPCs in vivo. For this purpose, we conducted a colony-forming unit-spleen (CFU-S) assay in lethally irradiated mice. As recipients transplanted with MEPs and/or CMPs develop spleen colonies, but those with GMPs do not, the CFU-S assay using HPCs reflects the MEP production and/ or function (Na Nakorn et al. 2002; Forsberg et al. 2006). The recipients with Keap1-deficient cells gave rise to fewer and smaller colonies than those with control cells (Fig. 4C). This is consistent with the results of the in vitro CFC assay (Fig. 4A) and the gene expression analysis (Fig. 4B). These results show that Keap1 deficiency stimulates HPC differentiation toward the granulocyte-monocyte lineage but suppresses its differentiation toward the erythroid lineage. Figure 4 Skewed lineage commitment of hematopoietic progenitor cells toward granulocyte-monocyte lineage. (A) Colony-forming cell assay using LK cells ( ) of Keap1 CKO and Control mice. Values indicate means SD (n = 5). BFU-E, burst-forming unit-erythroid; CFU-G, colony-forming unit-granulocyte; CFU-M, colony-forming unit-monocyte; CFU-GM, mixed colonies including CFU-G and CFU-M; CFU-Mix, mixed colonies including BFU-E, CFU- G and CFU-M. (B) Expression of Gata1, Cebpa and Cebpb in LK cells from Keap1 CKO and Control mice. The expression levels were normalized to Gapdh expression. The expression levels in Control mice were set to one. Values are presented as the means SD (n = 6). (C) CFU-S assay using BM cells ( ) of Keap1 CKO and Control mice. The spleen colonies were counted on day 8 post-transplantation. (left) Values indicate means SD (n = 9). (right) The representative spleen colonies are shown. *P < 0.05; **P < Keap1-deficient HSCs show granulocyte-monocyte lineage priming Given that Nrf2 is highly expressed in HSCs (Fig. 1C,D), we hypothesized that Nrf2 activation due to Keap1 deficiency might influence the lineage priming of HSCs. To examine whether the skewing of the differentiation potential due to Keap1 deficiency occurs in HSC fractions, we conducted a gene expression analysis in the LSK fraction. The LSK fraction was separated into two subfractions using the Flt3 surface marker: LSK Flt3 + for the multipotent progenitor (MPP) fraction and LSK Flt3 for the short-term (ST-) and LT-HSC fraction. We found that Gata1 expression was suppressed in the Keap1- deficient LSK Flt3 + cells, whereas Cebpa and Cebpb expression levels were almost comparable between Keap1-deficient and control cells (Fig. 5A, upper panel). Although the reduction of Gata1 in the LK cells was reproducible in the MPP fraction, the 245

8 S Murakami et al. (A) (B) (C) Figure 5 Skewed differentiation potential in hematopoietic stem cells fractions. (A) Expression of Gata1, Cebpa, Cebpb and Ikaros in Lin Sca-1 + c-kit + (LSK) Flt3 + and LSK Flt3 cells from Keap1 CKO and Control mice. The expression levels were normalized to Gapdh expression. The expression levels in Control mice were set to one. Values are presented as the means SD (n = 5 7). (B) Colony-forming cell assay using LSK Flt3 + cells ( ; left) and LSK Flt3 cells ( ; right) of Keap1 CKO and Control mice. Values indicate the means SD (n = 4 5). (C) CFU-S assay using LSK Flt3 + ( ) and LSK Flt3 cells ( )ofkeap1 CKO and Control mice. The spleen colonies were counted on day 12 post-transplantation. (left) Values indicate means SD (n = 5 7). (right) The representative spleen colonies are shown. # P = 0.05; *P < 0.05, **P < 0.005, ***P < increases in the Cebpa and Cebpb expression levels were not apparent. We also examined expression levels of Ikaros as a lymphoid-related gene, which is required for lymphoid lineage priming of HSCs (Ng et al. 2009). Ikaros expression was significantly downregulated in Keap1-deficient LSK Flt3 + cells, in good agreement with the reduction of lymphoid cells in Keap1 CKO mice (see Fig. 3B). Consistent with previous studies (Mansson et al. 2007; Ng et al. 2009), we observed low but steady expression of Gata1 and Ikaros in LSK Flt3 cells. Notably, Gata1 and Ikaros were down-regulated in Keap1-deficient LSK Flt3 cells (Fig. 5A, lower panel). These results support our hypothesis that the differentiation potential toward the erythroid and lymphoid lineages is limited in Keap1-null hematopoietic cells even in the HSC stages. From these data, we consider that the Nrf2 activation in HSC fractions elicits skewed lineage commitment as early as at the ST/LT-HSC stages toward granulocytes and monocytes. We executed two experiments to validate this contention: in vitro CFC assays and in vivo CFU-S assays using both LSK Flt3 + and LSK Flt3 cells. First, in the CFC assays, CFU-M colonies were increased in Keap1-deficient cells [Fig. 5B, asterisk (Control, ; Keap1 CKO, ) and square (Control, ; Keap1 CKO, )], whereas the total number of colonies did not significantly differ between the Keap1-deficient and control cells. This observation is consistent with the result of the gene expression analysis in that Keap1-deficient HSCs possess granulocyte-monocytebiased differentiation potential. Second, in the CFU-S assays, Keap1-deficient HSCs exhibited significantly fewer and smaller colonies than did control cells (Fig. 5C). When we compared the results with those of whole BM cells (Fig. 4C), we found that the differences were more striking in LSK Flt3 + and LSK Flt3 cells, indicating that Keap1 deletion in HSCs substantially limits MEP production. Thus, these results in combination strongly argue that 246

9 Keap1-Nrf2 system in hematopoiesis the Keap1 disruption in the HSC fractions induces lineage priming toward the granulocyte-monocyte lineage as early as at ST/LT-HSC stages. Nrf2 activation in ST/LT-HSCs promotes the differential preference for granulocyte-monocyte lineage To prove the Nrf2 dependency of the lineage skewing in the context of Keap1 deficiency, we analyzed the Nrf2 and Keap1 double knockout (Keap1 F/F ::Mx1-Cre::Nrf2 / ; Keap1 CKO::Nrf2 / ) mice in comparison with Control (Keap1 F/+ :: Nrf2 / or Keap1 F/+ ::Mx1-Cre::Nrf2 / ; Control:: Nrf2 / ) mice. While the BM cellularity was comparable in the Keap1- and/or Nrf2-deficient mice (Fig. 6A), the distorted differentiation observed in the absence of Keap1 was abrogated by the concomitant Nrf2 deletion (Fig. 6B). A similar result was also obtained in the lineage-committed progenitor fraction (Fig. 6C). These results clearly show that Nrf2 activation is responsible for the differential (A) (B) (C) (D) Figure 6 Distorted lineage commitment in Keap1 CKO mice depends on Nrf2 activation. (A,B) Whole bone marrow (WBM) cellularity (A) and ratios of each lineage in BM cells (B) of Control, Keap1 CKO, Control::Nrf2 / and Keap1 CKO::Nrf2 / mice. Values indicate means SD (n = 4 6). (C) Ratios of each lineage-committed progenitor in the LK fraction from mice of the four genotypes. (right) Values indicate means SD (n = 4 6). (left) The representative plots are shown. (D) CFU-S assay using Lin Sca-1 + c-kit + Flt3 cells ( ) from mice of the four genotypes. The spleen colonies were counted on day 12 post-transplantation. (left) Values are presented as means SD (n = 3-6). (right) The representative spleen colonies are shown. *P < 0.05; **P < 0.005; ***P <

10 S Murakami et al. preference of granulocytes and monocytes in the absence of Keap1. In the CFU-S assay using LSK Flt3 cells, the additional Nrf2 deficiency also recovered the reduction in the number of spleen colonies in the Keap1-deficient cells (Fig. 6D). In addition, it is noteworthy that Nrf2-deficient cells gave rise to larger colonies than Control::Nrf2 +/+ cells regardless of Keap1 genotype. This result suggests that the physiological amount of Nrf2, which is maintained at a low level due to the Keap1-mediated degradation (Fig. 2C,D), contributes to suppression of erythroid differentiation in Control::Nrf2 +/+ cells. In aggregate, Nrf2 contributes to cell fate determination at ST/LT-HSC levels, particularly by diverting the differentiation from the erythroid to the granulocyte-monocyte lineage. Phenotype of Nrf2 deficiency opposes that of Keap1 deficiency in the segregation of erythroid and myeloid lineages To evaluate the contribution of Nrf2 to hematopoiesis under steady-state conditions, we investigated Nrf2 / mice. Nrf2 / mice did not present significant differences in BM cellularity (Fig. 7A) but showed an increase in Ter119 + cells and a reduction in Mac-1 + Gr-1 + cells in the BM compared with wild-type mice (Fig. 7B). This phenotype is the opposite of that in Keap1 CKO mice. However, the lymphoid lineages, B220 + and CD3e + cells, were not affected by Nrf2 deficiency (Fig. 7B), indicating that under unstressed conditions, the contribution of Nrf2 to lymphoid differentiation is marginal. In the lineage-committed progenitors, GMPs were significantly reduced in Nrf2 / mice, whereas the ratio of MEPs was almost comparable in the two genotypes (Fig. 7C). In the CFC assay using LK cells, Nrf2-deficient cells established more BFU-E (Nrf2 +/ +, ; Nrf2 /, ) and fewer CFU-M (Nrf2 +/+, ; Nrf2 /, ) colonies than did wild-type cells (Fig. 7D). Furthermore, Nrf2-deficient BM cells gave rise to more colonies in the CFU-S assay (Fig. 7E), and the size of colonies was larger than that of wild-type colonies, which is similar to the result shown in Figure 6D. These results indicate that Nrf2 deficiency promotes megakaryocyte-erythroid differentiation at the expense of granulocyte-monocyte differentiation. Taken together, Nrf2 contributes to the efficient production of granulocytes and monocytes, even under homeostatic conditions when Nrf2 activation is maintained at very low levels (Fig. 2C,D). (A) (B) (C) (D) (E) Figure 7 Nrf2 deletion promotes erythroid differentiation. (A,B) Whole bone marrow (WBM) cellularity (A) and ratios of each lineage in BM cells (B) of Nrf2 / and wild-type mice. Values indicate means SD (n = 4 5). (C) Ratios of each lineage-committed progenitor in the LK fraction. Values indicate means SD (n = 4 5). (D) Colony-forming cell assay using LK cells ( ) from Nrf2 / and wild-type mice. Values indicate means SD (n = 6 7). (E) CFU-S assay using BM cells ( ) from Nrf2 / and wild-type mice. The spleen colonies were counted on day 8 post-transplantation. (left) Values indicate means SD (n = 13 17). (right) The representative spleen colonies are shown. *P < 0.05; **P <

11 Keap1-Nrf2 system in hematopoiesis Discussion This study has revealed a novel function of Nrf2 in the lineage commitment of hematopoietic cells. Our in vivo system using two types of transgenic reporter mouse lines enabled us to conduct single-cell measurements of Nrf2 expression at the mrna and protein levels. Capitalizing on this system, we unraveled that Nrf2 is abundantly transcribed in HSPC fractions and granulocyte-monocyte lineage cells. We also verified that the Keap1-Nrf2 system is functional in these cells. Consistent with the expression analysis, Nrf2 makes a substantial contribution to the differentiation of HSPCs by enhancing the lineage commitment toward granulocytes and monocytes at the expense of erythroid and lymphoid differentiation. Importantly, we discovered for the first time that the granulocyte-monocyte lineage priming due to Nrf2 activation emerges in the stage of ST/LT-HSCs. Our findings provide lines of evidence for crosstalk between the stress response and cell differentiation and add a new layer of complexity to Nrf2 function. Our expression analysis demonstrates that Nrf2 activity is almost entirely suppressed in HSCs under steady-state conditions (see Fig. 2). Therefore, we consider that Nrf2 contributes substantially to the function of HSCs under stressed conditions, especially during inflammation, because multiple inflammationrelated molecules activate Nrf2 (Kobayashi et al. 2013). Recent studies demonstrated that Nrf2 regulates HSC maintenance (Merchant et al. 2011; Tsai et al. 2013). Nrf2-deficient HSCs are easily driven into the proliferative state from quiescence, and their long-term reconstitution ability is impaired in BM transplantation assays. As inflammatory signals activate HSCs in a dormant state and drive them into proliferation, differentiation and eventually exhaustion (Essers et al. 2009; Baldridge et al. 2010; Esplin et al. 2011; Takizawa et al. 2012), we surmise that Nrf2 contributes to the protection and maintenance of HSCs from exhaustion under inflammatory conditions. We show in this study that Nrf2 activation skews the differentiation potential of HSCs toward the granulocyte-monocyte lineage, suggesting that in response to inflammatory signaling, Nrf2 regulates not only the maintenance but also the differentiation program of HSCs to enhance the production of granulocytes and monocytes. Notably, Nrf2 in macrophages has been shown to regulate the production of proinflammatory cytokines and to accelerate the resolution of inflammation (Thimmulappa et al. 2006; Kong et al. 2011). In addition to controlling the property of macrophages, our current study proposes that Nrf2 promotes the production of macrophages and confers resistance against inflammatory disorders. Recent studies demonstrate the presence of functionally distinct subsets within HSCs, that is, myeloid-biased HSCs and lymphoid-biased HSCs (Challen et al. 2010; Benz et al. 2012). The balance between myeloid- and lymphoid-biased HSCs determines the myeloid versus lymphoid differentiation potential of the BM during aging and other exogenous stresses, including inflammation (Beerman et al. 2010; Ergen et al. 2012). Intriguingly, a recent study identifies platelet-biased HSCs, which are selectively maintained by thrombopoietin and preferentially reconstitute platelets at the expense of other lineages (Sanjuan-Pla et al. 2013). This result implies the existence of multiple lineage-primed HSCs that are maintained by specific signals for each lineage. Given that the clearest and most solid observation from the effects of Nrf2 activation in HSCs is the enhancement of granulocyte-monocyte differentiation at the expense of erythroid differentiation rather than lymphoid differentiation, Nrf2 activation might selectively activate HSC subsets specifically primed for granulocyte-monocyte lineages and simultaneously suppress other lineage-biased HSCs. Other reports on HSCs have suggested the presence of heterogeneity at a single multipotent stem cell level (Hayashi et al. 2008; Pina et al. 2012). Multipotent cells transiently exhibit distinct biases for lineage commitment, which generates cellular heterogeneity. Therefore, Nrf2 activation in HSCs may reduce the probability of the stochastic expression of key regulatory factors for erythroid and lymphoid lineage commitments, resulting in the enhanced differentiation toward granulocytes and monocytes in progeny cells. Further investigations are needed to determine which model, lineage-biased HSC model and stochastic heterogeneous HSC model, is more appropriate to understand the Nrf2 function in HSCs. The intracellular ROS level is one of critical factors affecting the proliferation and differentiation of HSCs (Jang & Sharkis 2007; Sardina et al. 2012). The ROS accumulation induced by the deletion of Foxo3a, Atm, Tsc1 or Prdm16 leads to HSC exhaustion (Chen et al. 2008; Chuikov et al. 2010; Ito et al. 2004; Miyamoto et al. 2007), indicating that low ROS levels are essential to maintain HSCs quiescent (Jang & Sharkis 2007; Sardina et al. 2012). 249

12 S Murakami et al. A previous study demonstrated that in Drosophila intestine, constitutive Nrf2 activation maintains intestinal stem cells quiescent by suppressing ROS levels (Hochmuth et al. 2011). However, the impact of Nrf2 on the maintenance of murine HSCs has been suggested to act via ROS-independent pathways (Merchant et al. 2011; Tsai et al. 2013). An attenuation of cytokine and chemokine signaling impairs HSC activity in Nrf2-deficient mice. Consistently, we found that Keap1-deficient HSCs exhibit normal ROS levels (SM and MY, unpublished data). Therefore, we assume that Nrf2 regulates HSC differentiation through directly activating the target genes associated with cytokine and/or chemokine signaling pathways rather than modulating intracellular ROS levels. In summary, we clearly demonstrate that the Keap1-Nrf2 system affects the cell fate determination of HSCs. We recently encountered a similar observation that Nrf2 affects the cell-type specificity in liver; the enhanced activity of Nrf2 alters the differentiation of hepatoblasts and/or provokes transdifferentiation of hepatocytes to cholangiocytes (KT and MY, unpublished data). Considering the wideranging expression of Nrf2, it seems reasonable to expect that Nrf2 regulates cell fate determination in the other multipotent stem cell systems. Therefore, it becomes extremely important to clarify the molecular mechanisms underlying how Nrf2 regulates cell fate determination and how the Nrf2-mediated stress response is involved in this novel activity of Nrf2. Experimental procedures Generation of Nrf2-BAC-TG and MGRD-Neh2- TG mice To generate the transgene for the Nrf2-BAC-TG mice, a BAC clone (RP23-374O4) containing the murine Nrf2 gene locus was homologously recombined with a targeting vector constructed by ligation of a 0.7-kbp genomic fragment containing the 5 region and a part of the first exon of the murine Nrf2 gene, tdtomato cdna, polya fragment, neomycin resistance cassette flanked by FRT sites, and a 1.8-kbp genomic fragment containing the first intron of the murine Nrf2 gene. The targeting vector was electroporated into Escherichia coli strain EL250 with the BAC clone RP23-374O4 and subjected to prophage-based homologous recombination for BAC mutagenesis (Lee et al. 2001). The neomycin resistance cassette was deleted through FLP recombinase induced with L-arabinose. The properly recombined BAC clones were identified by Southern blot analysis and used for the generation of transgenic mice. To construct the transgene for the MGRD-Neh2-TG mice, a murine Nrf2 cdna fragment containing the 188-bp 5 UTR and a coding region for the Neh2 domain (1-99 amino acids) was excised from pneh2-gfp (Kobayashi et al. 2002) and ligated to tdtomato cdna. The resultant DNA fragment, encoding the fusion protein Neh2-tdTomato, was inserted into the 5 end of IRES in pires2-egfp (Clontech) to generate a plasmid harboring the Neh2-tdTomato-IRE- S-EGFP fragment. LacZ cdna in MGRD-LacZ (Yamazaki et al. 2012) was replaced with the Neh2-tdTomato-IRES- EGFP fragment resulting in MGRD-Neh2-tdTomato-IRES- EGFP, which was used for the generation of transgenic mice. The purified transgene constructs were injected into BDF1 fertilized eggs using the standard method. Two and three independent transgenic mouse lines were obtained for Nrf2-BAC-TG mice and MGRD-Neh2-TG mice, respectively. In both the transgenic mice, the single line with the highest level of fluorescence was used for the analyses in this study. Mice Keap1 F/F mice (C57BL/6; Okawa et al. 2006) were mated with Mx1-Cre transgenic mice (C57BL/6; Kuhn et al. 1995) to generate Keap1 F/F ::Mx1-Cre. The Keap1 F/F ::Mx1-Cre mice were mated with MGRD-Neh2-TG (mix background) mice, and the offspring were backcrossed to C57BL/6 background at least five times before being used for BM transplantation assays. To obtain Keap1 F/F ::Mx1-Cre::Nrf2 / mice, Keap1 F/ F ::Mx1-Cre mice were mated with Nrf2 / mice (C57BL/6; Itoh et al. 1997). pipc (25 mg/kg; Sigma) was intraperitoneally injected into 5- to 8-week-old mice every other day for 9 14 days. At 6 9 weeks after the last injection, the mice were analyzed. Nrf2 / mice were used for analyses at 7 13 weeks old. Keap1, Mx1-Cre and Nrf2 genotypes were determined using tail genomic DNAs. Peripheral blood was collected from the buccal region and analyzed using a hemocytometer (Nihon Kohden Co.). All animals were housed in specific pathogen-free conditions, according to the regulations of The Standards for Human Care and Use of Laboratory Animals of Tohoku University and Guidelines for Proper Conduct of Animal Experiments by the Ministry of Education, Culture, Sports, Science, and Technology of Japan. Flow cytometry Bone marrow cells were collected from femurs and tibias with phosphate-buffered saline supplemented with 3% heatinactivated fetal bovine serum and were used for cytometric analysis and cell sorting (FACS AriaII and LSR Fortessa, BD Biosciences). For analyses of lineage-positive cells, the BM cells were stained with antibodies against Ter119 (TER-119), Mac-1 (M1/70), Gr-1 (RB6-8C5), B220 (RA3-6B2) and CD3e (145-2C11). For analyses of HSPCs, the BM cells were reacted with biotin-conjugated lineage antibodies, and 250

13 Keap1-Nrf2 system in hematopoiesis lineage-positive cells were removed using streptavidincoupled M280 Dynabeads (Invitrogen). The lineage-negative cells were stained with antibodies against c-kit (2B8), Sca-1 (D7), FccR (93), Flt3 (A2F10) and CD34 (RAM34). The antibodies were purchased from BD Biosciences or ebioscience. Bone marrow cellularity was defined as the number of total cells obtained from the femurs and tibias of one mouse. FlowJo software (Tree star) was used to analyze the tdtomato fluorescence in the Nrf2-BAC-TG mice, and BD FACSDiva software (BD Biosciences) was used for the remaining analyses. Non-competitive BM transplantation assays For BM transplantation assays of Keap1-deficient cells, MGRD-Neh2-TG mice were used to mark donor-derived cells with EGFP expressed from the transgene. Keap1 F/F :: Mx1-Cre::MGRD-Neh2-TG and Keap1 F/+ ::MGRD-Neh2- TG or Keap1 F/+ ::Mx1-Cre::MGRD-Neh2-TG mice were treated with pipc at 5 to 8 weeks of age, and 6-13 weeks later, their BM cells were used for BM transplantation. BM cells ( ) were transplanted into lethally irradiated mice (C57BL/6), and the recipients were analyzed weeks after transplantation. Quantitative real-time PCR analysis RNA was isolated from sorted cells by using RNAiso (TaKaRa), and cdna was synthesized with random primers (Invitrogen) and SuperScript III reverse transcriptase (Invitrogen). Quantitative PCR was carried out on an Applied Biosystems 7300 sequence detector system, using qpcr Mastermix Plus for SYBR Green I (Nippon Gene) or Power SYBR Green PCR Master Mix (Applied Biosystems) for SYBR Green system and qpcr Mastermix (Nippon Gene) for Taqman probe system. The primer sequences are shown in Table S1, in Supporting Information. CFC assay The cells sorted through flow cytometry were seeded and cultured in duplicate in methylcellulose media supplemented with cytokines, MethoCult GFM3434 (StemCell Technologies), and the number of colonies was counted after 7 days. CFU-S assay Bone marrow cells ( ), LSK Flt3 + cells ( ) and LSK Flt3 cells ( ) were transplanted into lethally irradiated mice. The spleens with colonies were fixed using Tellesniczky solution (375 ml 70% ethanol, ml glacial acetic acid, and 37 ml formalin). The number of colonies was counted on day 8 or 12 post-transplantation. Statistical analysis The quantitative data are presented as the means SD and were analyzed using Student s t-test. P values of <0.05 were considered statistically significant. Acknowledgements We thank Dr T. Moriguchi for transgene construct design and Dr F. Katsuoka for the MGRD-LacZ plasmid. We also thank Y. Kawatani, H. Suda, E. Naganuma and the Biomedical Research Core of the Tohoku University for technical support. This work was supported by JSPS KAKENHI Grant Numbers (MY), (HM), MEXT KA- KENHI Grant Numbers (HM), Naito Foundation (MY), Takeda Scientific Foundation (HM and MY) and the Core Research for Evolutional Science and Technology from the JST (HM and MY). References Akashi, K., Traver, D., Miyamoto, T. & Weissman, I.L. (2000) A clonogenic common myeloid progenitor that gives rise to all myeloid lineages. Nature 404, Aw Yeang, H.X., Hamdam, J.M., Al-Huseini, L.M., Sethu, S., Djouhri, L., Walsh, J., Kitteringham, N., Park, B.K., Goldring, C.E. & Sathish, J.G. (2012) Loss of transcription factor nuclear factor-erythroid 2 (NF-E2) p45-related factor-2 (Nrf2) leads to dysregulation of immune functions, redox homeostasis, and intracellular signaling in dendritic cells. J. Biol. Chem. 287, Baldridge, M.T., King, K.Y., Boles, N.C., Weksberg, D.C. & Goodell, M.A. (2010) Quiescent haematopoietic stem cells are activated by IFN-c in response to chronic infection. Nature 465, Beerman, I., Bhattacharya, D., Zandi, S., Sigvardsson, M., Weissman, I.L., Bryder, D. & Rossi, D.J. (2010) Functionally distinct hematopoietic stem cells modulate hematopoietic lineage potential during aging by a mechanism of clonal expansion. Proc. Natl Acad. Sci. USA 107, Benz, C., Copley, M.R., Kent, D.G., Wohrer, S., Cortes, A., Aghaeepour, N., Ma, E., Mader, H., Rowe, K., Day, C., Treloar, D., Brinkman, R.R. & Eaves, C.J. (2012) Hematopoietic stem cell subtypes expand differentially during development and display distinct lymphopoietic programs. Cell Stem Cell 10, Beyer, T.A., Xu, W., Teupser, D., auf dem Kelller, U., Bugnon, P., Hildt, E., Thiery, J., Kan, Y.W. & Werner, S. (2008) Impaired liver regeneration in Nrf2 knockout mice: role of ROS-mediated insulin/igf-1 resistance. EMBO J. 27, Challen, G.A., Boles, N.C., Chambers, S.M. & Goodell, M.A. (2010) Distinct hematopoietic stem cell subtypes are differentially regulated by TGF-b1. Cell Stem Cell 6, Chan, J.Y. & Kwong, M. (2000) Impaired expression of glutathione synthetic enzyme genes in mice with targeted dele- 251

14 S Murakami et al. tion of the Nrf2 basic-leucine zipper protein. Biochim. Biophys. Acta 1517, Chen, C., Liu, Y., Liu, R., Ikenoue, T., Guan, K.L., Liu, Y. & Zheng, P. (2008) TSC-mTOR maintains quiescence and function of hematopoietic stem cells by repressing mitochondrial biogenesis and reactive oxygen species. J. Exp. Med. 205, Chuikov, S., Levi, B.P., Smith, M.L. & Morrison, S.J. (2010) Prdm16 promotes stem cell maintenance in multiple tissues, partly by regulating oxidative stress. Nat. Cell Biol. 12, Dinkova-Kostova, A.T., Holtzclaw, W.D., Cole, R.N., Itoh, K., Wakabayashi, N., Katoh, Y., Yamamoto, M. & Talalay, P. (2002) Direct evidence that sulfhydryl groups of Keap1 are the sensors regulating induction of phase 2 enzymes that protect against carcinogens and oxidants. Proc. Natl Acad. Sci. USA 99, Enomoto, A., Itoh, K., Nagayoshi, E., Haruta, J., Kimura, T., O Connor, T., Harada, T. & Yamamoto, M. (2001) High sensitivity of Nrf2 knockout mice to acetaminophen hepatotoxicity associated with decreased expression of ARE-regulated drug metabolizing enzymes and antioxidant genes. Toxicol. Sci. 59, Ergen, A.V., Boles, N.C. & Goodell, M.A. (2012) Rantes/ Ccl5 influences hematopoietic stem cell subtypes and causes myeloid skewing. Blood 119, Esplin, B.L., Shimazu, T., Welner, R.S., Garrett, K.P., Nie, L., Zhang, Q., Humphrey, M.B., Yang, Q., Borghesi, L.A. & Kincade, P.W. (2011) Chronic exposure to a TLR ligand injures hematopoietic stem cells. J. Immunol. 186, Essers, M.A., Offner, S., Blanco-Bose, W.E., Waibler, Z., Kalinke, U., Duchosal, M.A. & Trumpp, A. (2009) IFNa activates dormant haematopoietic stem cells in vivo. Nature 458, Forsberg, E.C., Serwold, T., Kogan, S., Weissman, I.L. & Passegue, E. (2006) New evidence supporting megakaryocyte-erythrocyte potential of Flk2/Flt3 + multipotent hematopoietic progenitors. Cell 126, Hayashi, K., Lopes, S.M., Tang, F. & Surani, M.A. (2008) Dynamic equilibrium and heterogeneity of mouse pluripotent stem cells with distinct functional and epigenetic states. Cell Stem Cell 3, Hochmuth, C.E., Biteau, B., Bohmann, D. & Jasper, H. (2011) Redox regulation by Keap1 and Nrf2 controls intestinal stem cell proliferation in Drosophila. Cell Stem Cell 8, Hou, Y., Xue, P., Bai, Y., Liu, D., Woods, C.G., Yarborough, K., Fu, J., Zhang, Q., Sun, G., Collins, S., Chan, J.Y., Yamamoto, M., Andersen, M.E. & Pi, J. (2012) Nuclear factor erythroid-derived factor 2-related factor 2 regulates transcription of CCAAT/enhancer-binding protein b during adipogenesis. Free Radic. Biol. Med. 52, Ishii, Y., Itoh, K., Morishima, Y., Kimura, T., Kiwamoto, T., Iizuka, T., Hegab, A.E., Hosoya, T., Nomura, A., Sakamoto, T., Yamamoto, M. & Sekizawa, K. (2005) Transcription factor Nrf2 plays a pivotal role in protection against elastase-induced pulmonary inflammation and emphysema. J. Immunol. 175, Ito, K., Hirao, A., Arai, F., Matsuoka, S., Takubo, K., Hamaguchi, I., Nomiyama, K., Hosokawa, K., Sakurada, K., Nakagata, N., Ikeda, Y., Mak, T.W. & Sudo, T. (2004) Regulation of oxidative stress by ATM is required for self-- renewal of haematopoietic stem cells. Nature 431, Itoh, K., Chiba, T., Takahashi, S., Ishii, T., Igrashi, K., Katoh, Y., Oyake, T., Hayashi, N., Satoh, K., Hatayama, I., Yamamoto, M. & Nabeshima, Y. (1997) An Nrf2/small Maf heterodimer mediates the induction of phase II detoxifying enzyme genes through antioxidant response elements. Biochem. Biophys. Res. Commun. 236, Itoh, K., Mochizuki, M., Ishii, Y., Shibata, T., Kawamoto, Y., Kelly, V., Sekizawa, K., Uchida, K. & Yamamoto, M. (2004) Transcription factor Nrf2 regulates inflammation by mediating the effect of 15-deoxy-D(12,14)-prostaglandin j (2). Mol. Cell. Biol. 24, Itoh, K., Wakabayashi, N., Katoh, Y., Ishii, T., Igarashi, K., Engel, J.D. & Yamamoto, M. (1999) Keap1 represses nuclear activation of antioxidant responsive elements by Nrf2 through binding to the amino-terminal Neh2 domain. Genes Dev. 13, Jang, Y.Y. & Sharkis, S.J. (2007) A low level of reactive oxygen species selects for primitive hematopoietic stem cells that may reside in the low-oxygenic niche. Blood 110, Kawatani, Y., Suzuki, T., Shimizu, R., Kelly, V.P. & Yamamoto, M. (2011) Nrf2 and selenoproteins are essential for maintaining oxidative homeostasis in erythrocytes and protecting against hemolytic anemia. Blood 117, Kobayashi, E., Suzuki, T. & Yamamoto, M. (2013) Roles Nrf2 plays in myeloid cells and related disorders. Oxid. Med. Cell. Longev. 2013, Kobayashi, M., Itoh, K., Suzuki, T., Osanai, H., Nishikawa, K., Katoh, Y., Takagi, Y. & Yamamoto, M. (2002) Identification of the interactive interface and phylogenic conservation of the Nrf2-Keap1 system. Genes Cells 7, Kobayashi, M., Li, L., Iwamoto, N., Nakajima-Takagi, Y., Kaneko, H., Nakayama, Y., Eguchi, M., Wada, Y., Kumagai, Y. & Yamamoto, M. (2009) The antioxidant defense system Keap1-Nrf2 comprises a multiple sensing mechanism for responding to a wide range of chemical compounds. Mol. Cell. Biol. 29, Kong, X., Thimmulappa, R., Craciun, F., Harvey, C., Singh, A., Kombairaju, P., Reddy, S.P., Remick, D. & Biswal, S. (2011) Enhancing Nrf2 pathway by disruption of Keap1 in myeloid leukocytes protects against sepsis. Am. J. Respir. Crit. Care Med. 184, Kuhn, R., Schwenk, F., Aguet, M. & Rajewsky, K. (1995) Inducible gene targeting in mice. Science 269, Lee, E.C., Yu, D., Martinez de Velasco, J., Tessaroll, L., Swing, D.A., Court, D.L., Jenkins, N.A. & Copeland, N.G. (2001) A highly efficient Escherichia coli-based 252