Author contributions: N.U.: conception and design, financial support, collection and/or assembly of data, data analysis and interpretation, manuscript writing, and final approval of manuscript; R.D.M.: administrative support, collection of data, provision of study material, and final approval of manuscript; T.F.: financial support, data analysis and interpretation, and final approval of manuscript; M.U.: financial support, conception and design, data analysis and interpretation, manuscript writing, and final approval of manuscript.
Disclosure of potential conflicts of interest is found at the end of this article.
First published online in STEM CELLSEXPRESS January 30, 2012.
Bone marrow (BM) microenvironment, which is regulated by hypoxia and proteolytic enzymes, is crucial for stem/progenitor cell function and mobilization involved in postnatal neovascularization. We demonstrated that NADPH oxidase 2 (Nox2)-derived reactive oxygen species (ROS) are involved in postischemic mobilization of BM cells and revascularization. However, role of Nox2 in regulating BM microenvironment in response to ischemic injury remains unknown. Here, we show that hindlimb ischemia of mice increases ROS production in both the endosteal and central region of BM tissue in situ, which is almost completely abolished in Nox2 knockout (KO) mice. This Nox2-dependent ROS production is mainly derived from Gr-1+ myeloid cells in BM. In vivo injection of hypoxyprobe reveals that endosteum at the BM is hypoxic with high expression of hypoxia-inducible factor-1α in basal state. Following hindlimb ischemia, hypoxic areas and HIF-1α expression are expanded throughout the BM, which is inhibited in Nox2 KO mice. This ischemia-induced alteration of Nox2-dependent BM microenvironment is associated with an increase in vascular endothelial growth factor expression and Akt phosphorylation in BM tissue, thereby promoting Lin− progenitor cell survival and expansion, leading to their mobilization from BM. Furthermore, hindlimb ischemia increases proteolytic enzymes membrane type 1-matrix metalloproteinase (MMP) expression and MMP-9 activity in BM, which is inhibited in Nox2 KO mice. In summary, Nox2-dependent increase in ROS plays a critical role in regulating hypoxia expansion and proteolytic activities in BM microenvironment in response to tissue ischemia. This in turn promotes progenitor cell expansion and reparative mobilization from BM, leading to postischemic neovascularization and tissue repair. STEM CELLS2012;30:923–934
Bone marrow (BM)-derived stem and progenitor cells have a critical role in maintaining vascular health and facilitating vasculogenesis/angiogenesis . The BM microenvironment, also termed the stem cell niche, provides essential cues that regulate stem and progenitor cell functions and their mobilization into the circulation involved in postnatal neovascularization . Hypoxia and proteolytic enzymes are key extrinsic factors that affect BM microenvironment regulating stem and progenitor cell function and their reparative mobilization. Stem/progenitor cells are originally shown to be located in two different BM microenvironments, including the BM interface (endosteal niche) that provides low oxygen environment for maintaining hematopoietic stem cells quiescence and the vascular niche that provides oxygenated environment for cell proliferation and differentiation [3–5]. Ischemic injury increases cytokines and growth factors such as vascular endothelial growth factor (VEGF) in BM and circulation, which in turn activates matrix metalloproteinase (MMP)-9 and releases soluble Kit ligand in the BM . MMPs including MMP-9 that is secreted mainly by neutrophils in BM  and MT (membrane type)1-MMP  that is anchored on the cell surface play a significant role in stem/progenitor cell mobilization and angiogenesis. Thus, understanding mechanisms by which ischemic injury regulates BM microenvironment and stem/progenitor cell function is essential for developing novel therapeutic strategies for treatment of various ischemic diseases.
Although excess amounts of reactive oxygen species (ROS) such as superoxide anion (O2•−) and hydrogen peroxide (H2O2) are detrimental to cells, ROS at appropriate level function as signaling molecules to mediate normal cell growth, migration, differentiation, apoptosis, and senescence . Evidence suggests that redox status and ROS affect stem/progenitor cell function including cell survival and proliferation [10–13] as well as interaction between stem cells and their niches . Self-renewal and differentiation of hematopoietic stem cells are regulated by intracellular ROS and niche microenvironment . One of the major sources of ROS involved in signaling is NADPH oxidase (Nox), which is expressed in phagocytes and nonphagocytic cells . Differentiated myeloid cells such as neutrophils and monocytes/macrophages generate a large amount of O2•− extracellularly via activation of Nox2 (also known as gp91phox). Nox isoforms are also expressed in stem/progenitor cells , and Nox-derived ROS are involved in differentiation, proliferation, senescence, or apoptosis [12, 17, 18]. We previously demonstrated that hindlimb ischemia increases Nox2-dependent ROS production in isolated BM-derived mononuclear cells (BM-MNCs) and that postischemic neovascularization and mobilization of BM cells (BMCs) are impaired in Nox2−/− mice . However, the mechanisms by which Nox2 regulate progenitor cell function and mobilization and the role of Nox2 in BM microenvironment remains unknown.
In this study, using a hindlimb ischemia model of Nox2−/− mice and in vivo injection of O2− reactive dye, we provide the direct evidence that ROS production is markedly increased in entire BM, in a Nox2-dependent manner, following ischemic injury in vivo. We also show that Nox2-derived ROS play an important role in hindlimb ischemia-induced expansion of hypoxia, hypoxia-inducible factor-1α (HIF-1α)/VEGF expression, as well as Akt and proteolytic enzyme (MT1-MMP and MMP-9) activation in BM in situ. Furthermore, these ROS-mediated, hypoxic BM microenvironment alterations induced by ischemic injury regulate progenitor cell survival and expansion, thereby promoting their mobilization from BM. Our study will provide new insights into role of Nox in redox regulation of BM microenvironment during postnatal neovascularization and tissue repair in response to injury.
MATERIALS AND METHODS
Ethics Statement for Animal Study
Study protocols were approved by the Animal Care and Institutional Biosafety Committee of University of Illinois at Chicago (A09-066), and the experiments were performed according to the Guide for the Care and Use of Laboratory Animals of the NIH.
Mouse Hindlimb Ischemia Model
Nox2−/− mice and their age- (range, 8–12 weeks) and sex-matched wild-type (WT) mice (C57BL/6J) were purchased from Jackson Laboratories (Bar Harbor, ME, http://www.jax. org) and were fed regular chow and water containing antibiotics, such as sulfamethoxazole and trimethoprim (Hi-Tech Pharmacal, Amityville, NY) under specific pathogen-free conditions at University of Illinois at Chicago. Mice were subjected to unilateral hindlimb surgery under anesthesia with ketamine (100 mg/kg) and xylazine (10 mg/kg). To induce hindlimb ischemia, left femoral arteries were ligated and removed from the branch point for profunda femoris to the proximal portion of the saphenous artery. All arterial branches between the ligations were obliterated using an electrical coagulator. Mice with major bleeding or signs of infections during observation were euthanized. To examine the BM, we used only long bones of contralateral legs that were not subjected to hindlimb surgery to exclude the direct effect of ischemia in the BM. Samples were collected immediately after euthanasia.
O2•− Detection in Mice
Dihydroethidium (DHE) (Invitrogen, Grand Island, NY, http://www.invitrogen.com) was prepared as a 1 mg/ml solution in 1% dimethyl sulfoxide and administered at 10 mg/kg b.wt. by i.v. injection. Mice were killed and immediately perfusion-fixed with 4% paraformaldehyde 1 hour after DHE injections. Frozen sections of the BM were prepared by the direct flushing of BM cavities with O.C.T. compound that allowed it to keep the structure of BM central regions and were observed by confocal microscope (Zeiss, Thornwood, NY, https://www.micro-shop.zeiss.com) with excitation at 510–550 nm and emission at >585 nm to detect oxidized ethidium. In some experiments, frozen sections were stained with rat-origin anti-CD45 or anti-Gr-1 (RB6-8C5) (eBioscience, San Diego, CA, http://www.ebioscience.com) and with DyLight 488-conjugated anti-rat IgG angibodies (Jackson ImmunoResearch Laboratories, West Grove, PA, http://www.jacksonimmuno.com). To quantify fluorescence on captured images, a parallelogram with size of approximately 50 × 500 μm was drawn from the endosteum into the BM and defined as the endosteal region. The same size of parallelogram was drawn in the middle of BM as the central region. The mean fluorescence intensity of the regions and background signals in the area with no cell were measured on ImageJ software (NIH). The fluorescent signal was calculated as the mean fluorescence intensity subtracted by the background signals in each sections.
O2•− Detection Using Isoluminol Enhanced Chemiluminescence Assay
O2•− production in isolated Gr1+ cells was measured using isoluminol enhanced chemiluminescent method in 6-mm diameter wells of 96-well, flat-bottom, white tissue culture plates (E&K Scientific, Campbell, CA). Freshly isolated BM Gr-1+ cells were resuspended in Krebs-HEPES buffer. The buffer with 100 μM isoluminol and 40 U/ml horseradish peroxidase (HRP) was incubated at room temperature for 5 minutes. An aliquot (200 μl) of the cells (1 × 106 cells) was added into the well and assayed for chemiluminescence (CL) at 37°C in a Wallac 1420 Multilabel Counter (PerkinElmer Life Sciences, Boston, MA, http://www.perkinelmer.com). The CL counts per second was continually recorded for 30 minutes. To isolate Gr-1+ cells from total BM cells, immunomagnetic cell sorting was performed using EasySep (Stem Cell Technologies, Vancouver, BC, http://www.stemcell.com) with biotinylated anti-Gr-1 antibodies (RB6-8C5; eBioscience). The measurement was started within 1 hour after sacrificing the mice.
Hypoxia Detection by Pimonidazole
Mice were injected 60 mg/kg pimonidazole hydrochloride (Hypoxyprobe) (Chemicon, Temecula, CA, http://www.chemicon.com) into peritoneal cavity 3 hours prior to tissue harvest. Preparation of paraffin sections of decalcified femurs and immunolabeling of pimonidazole were performed according to the literature . Briefly, deparaffinized sections were subjected to antigen retrieval in 10 mM citrate buffer (pH 6.0, at 90°C for 20 minutes), incubated in 50 mM glycine in phosphate buffered saline (PBS) (pH 3.5) for 5 minutes, and permeablized in 0.3% Triton X-100 for 15 minutes. After blocking, sections were incubated with mouse fluorescein isothiocyanate (FITC)-labeled anti-pimonidazole antibody (NPI, Burlington, MA) and labeled by anti-mouse IgG conjugated to HRP (Invitrogen). The signal was amplified using TSA™ biotin system (PerkinElmer, Waltham, MA) and labeled by Alexa488-conjugated streptavidin (Invitrogen) followed by mounting with Vectashield mounting medium with 4′,6-diamidino-2-phenylindole (Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com). Images were captured by a Zeiss LSM 510 Meta confocal microscope with appropriate laser and filter settings and processed by LSM510 software (Zeiss). The acquired images were measured, and the threshold was determined based on the histograms. The number of fluorescence positive cells was counted using ImageJ software (NIH).
Paraffin sections of decalcified femurs were labeled with anti-HIF-1α (Novus Biologicals, Littleton, CO), anti-VEGF, anti-MMP14 (MT1-MMP) (Abcam, Cambridge, MA, http://www.abcam.com), anti-phospho-Akt (Ser473) (Cell Signaling Technology, Danvers, MA, http://www.cellsignal.com), or rabbit IgG at matched concentration. Donkey biotinylated anti-rabbit IgG (Vector Laboratories) was used as secondary antibody followed by TSA biotin system (PerkinElmer) for signal enhancement, Alexa488-conjugated streptavidin (Invitrogen) for fluorescence, or 3,3′-diaminobenzidine (Vector Laboratories) for immunohistochemistry. Images for fluorescence were captured by a confocal microscopy and processed by LSM510 software (Zeiss). For immunohistochemistory, an Axio Scope with AxioVision software (Zeiss) was used. The acquired images were measured, and the threshold was determined based on the histograms. The number of fluorescence positive cells was counted or % of fluorescence positive area was analyzed using ImageJ software (NIH).
After euthanasia, legs were cut and attached muscles were stripped quickly. The femurs and tibiae were flushed with ice-cold Dulbecco's phosphate-buffered saline supplemented with 2% FCS (DPBS/FCS). Single cell suspension was obtained by gently passing through 23-gauge needles. Unfractionated cells were stained with Lineage cocktail antibodies (Gr-1, CD11b, Ter119, B220, and CD3e) (eBioscience) and MMP14 (MT1-MMP) (Abcam). The cells were then washed, resuspended in 2 μg/ml propidium iodide in DPBS/2% FCS, and analyzed by a flow cytometer (DAKO ADP Cyan) in Research Resources Center in University of Illinois at Chicago. Summit (DAKO North America, Carpinteria, CA) or FlowJo 7.6 (Tree Star, Ashland, OR, http://www.treestar.com/) software was used for population analysis. For apoptosis detection, FITC Annexin V Apoptosis Detection Kit (BD Biosciences, San Diego, CA, http://www.bdbiosciences. com) was used, as described in the manufacturer's instructions. Peripheral blood was drawn from retro-orbital cavities under anesthesia and was separated to form the buffy coat. The number of circulating leukocyte or total BM cells was counted manually in the cell suspension from the buffy coat or tibiae, respectively, using hemocytometer.
Flow Cytometric Cell Separation
Gr-1+, Lin−, or cKit+/Lin− cells were sorted by MoFlo fluorescence-activated cell sorter (DAKO Cytomation, Carpinteria, CA) located in the research resource center at University of Illinois at Chicago from BM-MNCs separated by density gradient with Histopaque1083 (Sigma-Aldrich, St. Louis, MO, http://www.sigmaaldrich.com). Labeling was performed with Pacific blue-conjugated anti-Gr-1 (RB6-8C5), biotinylated Lineage cocktail antibodies, and allophycocyanin (APC)-conjugated anti-cKit (2B8) antibodies (eBioscience). Streptavidin-Alexa488 was used for Lineage detection.
Real-Time Polymerase Chain Reaction Analysis
Total RNA was prepared from cells using TRI Reagent (Molecular Research Center, Cincinnati, OH) according to manufacturer's protocol. Reverse transcription was carried out using high-capacity cDNA reverse transcription kit (Applied Biosystems, Carlsbad, CA, http://www.appliedbiosystems. com) or SuperScript VILO cDNA Synthesis Kit (Invitrogen). The real-time polymerase chain reactions (PCRs) were run on an ABI Prism 7000 Sequence Detection System (Applied Biosystems) in the SYBR Green PCR kit and the QuantiTect Primer Assay (Qiagen, Valencia, CA, http://www1.qiagen. com) for specific genes. Expression of genes was normalized and expressed as fold changes relative to β-actin unless otherwise indicated.
Colony-Forming Unit Assay
Total BM cells (104 per well) were plated in six-well plates with 1.5 ml of methylcellulose-based medium (Methocult (M3434), Stem Cell Technologies) and colonies (>50 cells) were scored at day 12. Colony-forming unit (CFU) was normalized by total BM cell number counted manually using hemocytometer.
BM plasma was collected by flushing femurs with a total of 0.5 ml PBS. After spinning, supernatants were analyzed for metalloproteinase activity by gelatinolytic zymography, as described .
Comparison between groups was analyzed by unpaired Student's two-tailed t test (two groups) or analysis of variance (ANOVA) for experiments with more than two subgroups followed by Bonferroni post hoc analysis. p < .05 was considered as statistically significant.
Nox2 Is Involved in Hindlimb Ischemia-Induced Increase in O2•− Production in the Endosteal and Central BM In Situ
To determine the role of Nox2 in BM microenvironment during postnatal neovascularization, we examined the spatial distribution of O2•− in the BM tissue in situ in response to hindlimb ischemia in WT mice and Nox2−/− mice, which impair postischemic angiogenesis [19, 22]. To eliminate a postharvesting phenomenon that may influence the measurement of O2•− in the BM, we injected O2•− detecting dye, DHE, into the mice before sacrifice . Figure 1 shows that hindlimb ischemia significantly increased DHE staining in the central BM and to lesser extent in the endosteal region defined as within 50 μm from the bone surface in WT mice, which were almost completely inhibited in Nox2−/− mice. Colocalization analysis for DHE positive signals with hematopoietic maker CD45, myeloid maker Gr-1 in BM tissues shows that the majority of ROS producing cells in BM after hindlimb ischemia were differentiated myeloid cells (Fig. 2A). Hind limb ischemia also significantly increased Nox2 mRNA in both BM-MNCs and cKit+Lin− BM progenitor cells (Fig. 2B). The extent of Nox2 gene expression in Gr-1+ cells was much higher than that of cKit+Lin− cells (4.38-fold). Since we found that Gr-1+ cells are major ROS producing cells in BM tissue in response to hindlimb ischemia, we measured ROS production in isolated Gr-1+ BM cells from WT and Nox2−/− mice using isoluminol assay. Figure 2C shows that hindlimb ischemia significantly increased ROS production in WT Gr-1+ cells, which was almost completely abolished in Nox2−/− Gr-1+ cells with significant decrease in basal ROS levels.
Hindlimb Ischemia Promotes Hypoxia Expansion and Increases HIF-1α and VEGF Expression Throughout the BM, in a Nox2-Dependent Manner
Since oxygen is consumed to produce O2•− by activation of Nox, we next examined the role of Nox2 in the spatial distribution of hypoxia in BM in response to hindlimb ischemia. For this purpose, we injected a hypoxic bioprobe, pimonidazole, which detects low O2 tension less than 1.3% . Figure 3 shows that without hindlimb ischemia, a majority of pimonidazole-positive cells was preferentially incorporated in the hypoxic endosteal area adjacent to long bone surface while its staining is lower in the central BM area. In response to hindlimb ischemia, the hypoxic area stained by pimonidazole was expanded throughout the BM, which was almost completely abolished in Nox2−/− mice (Fig. 3).
Since HIF-1α is a downstream target of hypoxia, we examined the distribution and expression of HIF-1α in the BM after hindlimb ischemia. Immunofluorescence analysis of BM tissue revealed that HIF-1α protein was highly expressed at the endosteum in basal state (Fig. 4A, 4C). Hind limb ischemia increased expression of HIF-1α in both endosteal and central BM, which was almost completely inhibited in Nox2−/− mice (Fig. 4A, 4C). These results suggest that Nox2 activation may contribute to expansion of hypoxic BM microenvironment and increase in HIF-1α expression in response to tissue ischemia.
We then examined the expression of VEGF protein, one of the key HIF-α target genes inducing stem cell mobilization, using immunostaining in BM tissue. We found that VEGF expression was increased and expanded in whole BM following hindlimb ischemia, which was almost completely inhibited in Nox2−/− mice (Fig. 4B, 4D). Furthermore, VEGF mRNA expression in both BM-MNC and Lin− progenitor cells was significantly upregulated following hindlimb ischemia, which was abolished in Nox2−/− mice (Fig. 4E). Basal expression of VEGF protein and mRNA were rather increased in Nox2−/− BM tissue, BM-MNC, and progenitor cells presumably due to compensatory mechanism. These results suggest that there is causal relationship between Nox2-ROS and hypoxia/HIF1-α/VEGF pathway in BM microenvironment in response to hindlimb ischemia.
Nox2 Is Involved in Promoting Cell Expansion and Survival Following Hindlimb Ischemia, Thereby Stimulating Mobilization of Lin− Progenitor Cells
Since hypoxia and HIF-1α in BM microenvironment may regulate stem/progenitor cell function and mobilization , we next examined the role of Nox2 in Lin− progenitor cell mobilization after hindlimb ischemia. Figure 5A shows that the number of circulating Lin− progenitor cells in peripheral blood was increased following hindlimb ischemia, which was significantly inhibited in Nox2−/− mice (Fig. 5A). Under this condition, the pool size of Lin− progenitor cells in the BM, as measured by flow cytometry analysis (Fig. 5B) and CFU assay (Fig. 5C), was significantly increased following hindlimb ischemia. Although the basal progenitor pool was significantly higher in Nox2−/− BM cells, ischemia-induced enhancement of progenitor cell expansion was markedly inhibited (Fig. 5B, 5C). Thus, Nox2-mediated increase in hypoxic BM microenvironment may promote progenitor cell expansion, thereby stimulating their mobilization from BM after ischemic injury. To address the underlying mechanism, we examined the role of Nox2 in cell survival and proliferation of BM Lin− progenitor cells. Flow cytometry analysis with Annexin V and propidium iodide (PI) staining shows that apoptotic cells (Annexin V+/PI−) were significantly increased in Nox2−/− mice following hindlimb ischemia, as compared to those in WT mice (Fig. 5D).
Nox2 Is Involved in Akt Phosphorylation and MMPs Activation in BM Microenvironment in Response to Hindlimb Ischemia
Since Akt is shown to be involved in cell survival and HIF-1α induction [24, 25], we examined the role of Nox2 in Akt activation in BM tissue following hindlimb ischemia. As shown in Figure 6A, immunostaining with p-Akt antibody on BM section reveals that hindlimb ischemia markedly increased phosphorylation of Akt in entire BM, which was almost completely blocked in Nox2−/− BM. Since proteolytic enzymes such as MT1-MMP and MMP-9 also play an essential role in regulating BM microenvironment, we thus examined the role of Nox2 in these enzymes' expression or activity in BM following hindlimb ischemia. MT1-MMP expression in BM tissue was markedly enhanced by hindlimb ischemia, which was prevented in Nox2−/− BM (Fig. 6A). Flow cytometry analysis further demonstrates that hindlimb ischemia increased cell surface MT1-MMP expression in Lin− BM progenitor cells, which was markedly inhibited in Nox2−/− mice (Fig. 6B). In parallel, activity of MMP-9, a target of MT1-MMP  and a major regulator of stem and progenitor cells mobilization from BM , in BM plasma was increased by hindlimb ischemia, which was inhibited in Nox2−/− mice (Fig. 6C). Basal MT1-MMP expression (Fig. 6A) and MMP-9 activity (Fig. 6C) were rather increased in Nox2−/− mice, which may be due to compensatory mechanisms. Of note, Akt is involved in increase in MT1-MMP expression in hematopoietic cells . We also performed Hematoxylin and Eosin staining of BM tissues and found that there was no morphological difference between WT and Nox2−/− mice (data not shown). Thus, these results suggest that Akt and MMPs are important downstream targets for Nox2-derived ROS in BM microenvironment that regulate progenitor cell function and mobilization in response to ischemic injury.
BM microenvironment regulated by hypoxia and proteolytic enzymes  plays an important role in stem/progenitor cell function and mobilization involved in postnatal neovascularization . We previously reported that Nox2-derived ROS are involved in postischemic mobilization of BMCs and revascularization; however, the role of Nox2 in regulating BM microenvironment in response to ischemic injury is entirely unclear. We show here that hindlimb ischemia increases: (a) Nox2-dependent ROS production mainly derived from Gr-1+ myeloid cells in the entire BM in situ; (b) hypoxia expansion and its downstream HIF-1α and VEGF expression throughout the BM; (c) VEGF mRNA expression in differentiated myeloid cells and Lin− progenitor cells; (d) BM Lin− progenitor cell survival and expansion and mobilization to the circulation; (e) Akt phosphorylation, MT1-MMP expression, and MMP-9 activity in BM. All of these ischemia-induced responses in BM tissue or progenitor cells are inhibited in Nox2−/− mice. Thus, Nox2-derived ROS play a critical role in regulating hypoxic niche and proteolytic enzyme activity in BM microenvironment, which is required for progenitor cell expansion and reparative mobilization in response to ischemia (Fig. 7).
Stem and progenitor cells in the BM are localized in a microenvironment known as the stem cell “niche,” which includes the low oxygen endosteal niche mainly containing quiescent hematopoietic stem cells as well as the more oxygenated vascular niche containing differentiated hematopoietic progenitors [3, 4]. Previous studies used isolated stem or progenitor cells to measure intracellular ROS levels under ambient oxygen ex vivo [10, 11, 28], which might not reflect the in vivo status of ROS levels. In this study, in vivo injection of O2•− reactive probe, DHE, into the mice demonstrates that hindlimb ischemia robustly increases ROS production in BM tissues in situ predominantly at the central BM and at lesser extent at the endosteal regions, in a Nox2-dependent manner. Together with previous report  and as discussed below, it is conceivable that lower levels of ROS generation in endosteal niche may be associated with its hypoxic nature favoring primitive hematopoietic stem cells, while higher ROS production in central BM may link to high Nox2 expressing differentiated cells localized at vascular niche. We also found that ischemic injury-induced ROS in BM is mainly derived from differentiated myeloid cells including Gr-1+ cells and that Nox2 gene expression in Gr-1+ cells is approximately 4.4-fold higher than that in cKit+Lin− cells. Given the diffusible nature of ROS, these results suggest that hindlimb ischemia increases ROS mainly from Gr-1+ cells in BM in a Nox2-dependent manner, which may create oxidative microenvironment to regulate BM stem cell niche.
Hypoxic microenvironment is an important determinant for stem/progenitor function in the BM [29, 30]. Since oxygen is a source for O2•− induced by Nox activation, we investigated the role of Nox2 in the spatial distribution of hypoxia in the BM following hindlimb ischemia. This study using in vivo injection of hypoxic bioprobe pimonidazole, which cross-links to protein adducts at oxygen tension below 10 mm Hg (less than 1.3% O2) , reveals that the endosteum at the bone BM interface is more hypoxic than the central BM in mice in the basal state. This is consistent with previous reports that the endosteal niche in steady state is hypoxic, while the central BM is more oxygenated presumably due to the higher density of endothelial sinuses where oxygenated blood circulates. Hypoxic endosteal niche maintains hematopoietic stem cells in quiescent state, which requires low oxygen tension and ROS level, while vascular niche with high oxygen tension contain more proliferative and differentiated stem/progenitor cells . In this study, hindlimb ischemia expands low oxygen (hypoxic) area throughout the BM, in a Nox2-dependent manner. In line with our results, granulocyte colony stimulating factor (G-CSF) or cyclophosphamide (CY), which stimulates mobilization of stem/progenitor cells, has been shown to increase hypoxia in the BM . Of note, G-CSF- and CY-induced stem cell mobilization was shown to be blunted in Nox2−/− mice . In this study, the mechanism by which hindlimb ischemia increases hypoxia through Nox2 in contralateral legs remains unclear. It is possible that ischemic injury may increase cytokines, growth factors, neurotransmitters, and metabolites such as lactates, ROS themselves, and inflammatory cells including Gr-1+ cells that produce large amounts of ROS in circulation, which may regulate contralateral BM microenvironment. The reason why BM becomes hypoxia by activation of Nox2 may be presumably due to the increase in oxygen consumption. Piccoli et al.  reported that the half of the oxygen consumption in hematopoietic stem and progenitor cells is dependent on Nox. In addition, the consumption of respiratory burst by Nox in differentiated myeloid cells has been shown to contribute to local hypoxia . Thus, Nox2 seems to be responsible for increased oxygen consumption and hypoxia in progenitor cells and Gr-1+ cells in mobilizing or inflammatory condition after hindlimb ischemia. Contribution of ROS-mediated vasoconstriction in the BM vasculature in reducing oxygen supply also cannot be excluded. Taken together, it is likely that hind limb ischemia increases Nox2-dependent ROS-producing oxidative microenvironment, which in turn creates a hypoxic niche throughout the BM by increasing oxygen consumption.
To address the consequence of increasing hypoxia in BM after hindlimb ischemia, we examined the expression of HIF-1α, a key downstream target of hypoxia involved in cell survival, angiogenesis , and hematopoietic stem cell function . Immunofluorescence analysis of BM tissue sections shows that hindlimb ischemia increases expression of HIF-1α in entire BM tissue, in a Nox2-dependent manner. Consistent with our finding about the role of Nox in HIF-1α level, it has been shown that lipopolysaccharide increases HIF-1α expression through activation of Nox in myeloid cells . Intermittent hypoxia activates HIF-1α through Nox2, which is required for hypoxia-induced ROS generation , suggesting that positive feed-forward mechanism among Nox2, hypoxia, and HIF-1α may exist. Of note, HIF-1α is also regulated by ROS or kinase in a hypoxia-independent manner. It is reported that Nox-derived ROS promote HIF-1α stabilization in circulating hematopoietic stem cells by inhibiting von Hippel-Lindau under normoxic conditions . HIF-1α is also regulated by Akt as discussed below. In this study, we also found that hindlimb ischemia increases VEGF expression, a downstream target of HIF-1α involved in stem/progenitor cell mobilization and reparative neovascularization  in whole BM tissue, BM-MNCs, and Lin− progenitor cells, in a Nox2-dependent manner. Diffusible staining of VEGF may be due to the possibility that VEGF is produced and secreted from various cell types including endothelial cells, progenitor cells, and differentiated myeloid cells in BM following ischemic injury. Most recently, Rehn et al.  reported that hypoxic niche regulates hematopoietic stem cell function via the hypoxia-VEGF axis. Thus, these results suggest that Nox2-derived ROS play an important role for providing redox-sensitive BM microenvironment at least in part by increasing the hypoxia/HIF-1α/VEGF pathway in response to tissue ischemia.
ROS at physiological levels act as signaling molecules to mediate various biological responses including stem/progenitor cell function. Hypoxia and HIF-1α influence stem/progenitor cells expansion by regulating cell survival and proliferation . Stem cells can survive better under lower oxygen condition . In this study, we show that Nox2 is required for ischemia-induced Lin− progenitor cell expansion in BM, as measured by Lin− progenitor cell pool using flow cytometry analysis and CFU assay. This is at least in part through increasing progenitor cell survival, thereby promoting their mobilization to circulation after ischemic injury. Of note, Nox2−/− mice had more basal Lin− progenitor cell pool (higher CFU) in unfractionated BMCs, which is consistent with previous report . This may be in part through the compensatory increase in basal expression of VEGF in Nox2−/− BM (Fig. 4). We found that BM transplantation of Nox2−/− BMCs was less efficient than that from WT mice (unpublished observations), indicating that Nox2−/− hematopoietic stem cells have lower reconstitution capacity after BM transplantation. This may be in part due to homing defect of Nox2−/− BMCs to BM, since we observed that SDF-1-induced BMCs migration mediated through Akt is inhibited in Nox2−/− mice . Akt is shown to be activated by hypoxia or ROS  and involved cell survival and HIF-1α expression and activity [24, 25, 41] as well as MT1-MMP expression in hematopoietic stem cells . This study shows that hindlimb ischemia markedly increases phosphorylation of Akt in the BM, which is almost completely blocked in Nox2−/− BM. Thus, Akt is an important downstream target for Nox2-derived ROS in BM microenvironment regulating stem and progenitor function and mobilization in response to ischemic injury. It is also possible that Nox2-mediated increase in hypoxia/HIF-1α/VEGF pathway may also indirectly activate Akt in the BM. This may represent a novel mechanism by which Nox2-derived ROS modulate stem/progenitor function by regulating BM microenvironment.
To further define the mechanism underlying Nox2-dependent mobilization of stem/progenitor cells from BM, we also examined the major proteolytic enzymes' expression and activities in the BM in WT and Nox2−/− mice. Both MT1-MMP and MMP-9 play an important role in stem/progenitor cell migration or mobilization through inhibiting adhesive interactions necessary to retain stem/progenitor cells within the BM . This study demonstrates that hindlimb ischemia increases MT1-MMP expression in BM tissues and Lin− progenitor cells, in a Nox2-dependent manner. This is consistent with previous report that G-CSF increases MT1-MMP expression and decreases its endogenous inhibitor, reversion-inducing cysteine-rich protein with Kazal motifs (RECK), expression in human CD34+ progenitor cells through Akt-dependent pathway, thereby promoting their mobilization . We also found that hindlimb ischemia decreases RECK expression in BM tissue, in a Nox2-dependent manner (unpublished observations). Furthermore, this study shows that hindlimb ischemia increases MMP-9 activities in BM plasma, which is inhibited in Nox2−/− mice. Consistent with our data, low-dose irradiation that potentially induces ROS promotes progenitor mobilization through MMP-9 and VEGF production . Basal MT1-MMP expression in BM tissue, but not that in Lin− cells, as well as MMP-9 activity in BM plasma are increased in Nox2−/− mice, which is associated with compensatory increase in VEGF expression in BM, as described above. Of note, these steady state basal increases in proteolytic enzymes seem not to affect the number of circulating BM progenitors in Nox2−/− mice, while hindlimb ischemia-induced responses are blunted in these mice (Fig. 6A). Thus ischemic injury-induced acute increase in MT1-MMP and MMP-9 may play more important role in mobilizing progenitor cells to circulation. Our results are consistent with the notion that Nox2-derived ROS play an important role in regulating proteolytic activity in BM in response to hindlimb ischemia, which may facilitate the stem/progenitor cell mobilization.
In summary, this study demonstrates that ischemic injury increases ROS mainly derived from Gr-1+ myeloid cells through Nox2 activation in entire BM, which increases hypoxic niche and its downstream HIF-1α and VEGF expression as well as Akt and proteolytic enzyme activity in BM tissue. These Nox2-dependent alterations of BM microenvironment promote Lin− progenitor cell survival and expansion, thereby promoting their mobilization to circulation, leading to reparative neovascularization and tissue repair (Fig. 7). This study should provide insights into redox regulation of BM microenvironment as a potential therapeutic strategy for treatment of various ischemic diseases [2, 13, 43-45].
We thank Dr. Nihal Kaplan for technical assistance and helpful comments and discussions. This work was supported by funds from NIH R01 Heart and Lung (HL)077524, HL077524-S1 (to M.U.-F.), and HL070187 (to T.F.); American Heart Association (AHA) Grant-In-Aid 0755805Z (to M.U.-F.), AHA National Center Research Program (NCRP) Innovative Research Grant 0970336N (to M.U.-F), and AHA Post-doctoral Fellowship 09POST2250151 (to N.U.).
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
The authors indicate no potential conflicts of interest.