2023

Reactive oxygen species (ROS) are generated not only as by-products of mitochondrial metabolism but also by a

variety of cellular enzyme systems, including nicotinamide adenine dinucleotide phosphate (NADPH) oxidase (NOX), uncoupled endothelial nitric oxide synthase (eNOS), xan-thine oxidase, and arachidonic acid–metabolizing enzymes. When cellular production of ROS exceeds the antioxidant capacity of cardiovascular cells, proteins, lipids, and nucleic acids become damaged and may eventually contribute to the development of cardiovascular diseases such as atheroscle-rosis, hypertension, diabetic cardiovascular complications, and ischemic–reperfusion injury. Conversely, low concentra-tions of ROS play a critical role in regulating cardiovascular functions such as angiogenesis and tissue repair.1–3 ROS are required for vascular endothelial growth factor (VEGF)–induced endothelial migration, proliferation, and tube for-mation.4,5 During ischemia and reperfusion, ROS generation promotes capillary tube formation in human microvascular endothelial cells6 and the heart,7 whereas inhibiting ROS

through treatment with antioxidants or superoxide dismutases blocks vascularization and growth of tumors.8,9 The precise molecular mechanisms, however, by which ROS mediate angiogenic responses are incompletely understood.

NOX is an important enzymatic source of ROS. There are 7 Nox genes identified in mammalian organisms: Nox1–5 and dual oxidases 1 to 2. NOXs are expressed in endothelial cells and other cardiovascular cells and regulate various functions such as cell survival, growth, apoptosis, differentiation, angio-genesis, and contractility.10 The NOX enzymes are hetero-protein complexes (except NOX5) with different regulatory mechanisms, tissue distribution, and subcellular localization and downstream targets. A membrane regulatory subunit, p22phox, is associated with NOXs 1, 2, and 4 and is required for their activity.10 NOXs 1 and 2 share a common over-all structure with a short cytoplasmic N terminus, which is required for activation,11–13 and 6 transmembrane domains.14,15 In contrast, NOX4 is constitutively active and is regulated by gene expression.16 Interestingly, unlike other NOXs, NOX5

© 2014 American Heart Association, Inc.

Arterioscler Thromb Vasc Biol is available at http://atvb.ahajournals.org DOI: 10.1161/ATVBAHA.114.303733

Objective—Reactive oxygen species (ROS) act as signaling molecules during angiogenesis; however, the mechanisms used for such signaling events remain unclear. Stromal cell–derived factor-1α (SDF-1α) is one of the most potent angiogenic chemokines. Here, we examined the role of ROS in the regulation of SDF-1α–dependent angiogenesis.

Approach and Results—Bovine aortic endothelial cells were treated with SDF-1α, and intracellular ROS generation was monitored. SDF-1α treatment induced bovine aortic endothelial cell migration and ROS generation, with the majority of ROS generated by bovine aortic endothelial cells at the leading edge of the migratory cells. Antioxidants and nicotinamide adenine dinucleotide phosphate oxidase (NOX) inhibitors blocked SDF-1α–induced endothelial migration. Furthermore, knockdown of either NOX5 or p22phox (a requisite subunit for NOX1/2/4 activation) significantly impaired endothelial motility and tube formation, suggesting that multiple NOXs regulate SDF-1α–dependent angiogenesis. Our previous study demonstrated that c-Jun N-terminal kinase 3 activity is essential for SDF-1α–dependent angiogenesis. Here, we identified that NOX5 is the dominant NOX required for SDF-1α–induced c-Jun N-terminal kinase 3 activation and that NOX5 and MAP kinase phosphatase 7 (MKP7; the c-Jun N-terminal kinase 3 phosphatase) associate with one another but decrease this interaction on SDF-1α treatment. Furthermore, MKP7 activity was inhibited by SDF-1α, and this inhibition was relieved by NOX5 knockdown, indicating that NOX5 promotes c-Jun N-terminal kinase 3 activation by blocking MKP7 activity.

Conclusions—We conclude that NOX is required for SDF-1α signaling and that intracellular redox balance is critical for SDF-1α–induced endothelial migration and angiogenesis. (Arterioscler Thromb Vasc Biol. 2014;34:2023-2032.)

Key Words: angiogenesis ◼ migration ◼ NADPH oxidase ◼ reactive oxygen species ◼ stromal cell-derived factor-1α

Received on: November 19, 2012; final version accepted on: June 20, 2014.From the UNC McAllister Heart Institute (X.P., L.X., A.L.P., P.L., X.L., C.P.), Department of Medicine (X.P., L.X., A.L.P., P.L., X.L., C.P.), and

Department of Chemistry (S.K.), University of North Carolina, Chapel Hill.The online-only Data Supplement is available with this article at http://atvb.ahajournals.org/lookup/suppl/doi:10.1161/ATVBAHA.114.303733/-/DC1.Correspondence to Xinchun Pi, PhD, Division of Cardiology and UNC McAllister Heart Institute, 8200 Medical Biomolecular Research Bldg, Chapel

Hill, NC 27599-7126. E-mail pxinchun@med.unc.edu

NADPH Oxidase–Generated Reactive Oxygen Species Are Required for Stromal Cell–Derived

Factor-1α–Stimulated AngiogenesisXinchun Pi, Liang Xie, Andrea L. Portbury, Sarayu Kumar, Pamela Lockyer, Xi Li, Cam Patterson

Basic Sciences

by guest on July 18, 2016http://atvb.ahajournals.org/Downloaded from by guest on July 18, 2016http://atvb.ahajournals.org/Downloaded from by guest on July 18, 2016http://atvb.ahajournals.org/Downloaded from by guest on July 18, 2016http://atvb.ahajournals.org/Downloaded from by guest on July 18, 2016http://atvb.ahajournals.org/Downloaded from by guest on July 18, 2016http://atvb.ahajournals.org/Downloaded from by guest on July 18, 2016http://atvb.ahajournals.org/Downloaded from by guest on July 18, 2016http://atvb.ahajournals.org/Downloaded from by guest on July 18, 2016http://atvb.ahajournals.org/Downloaded from by guest on July 18, 2016http://atvb.ahajournals.org/Downloaded from by guest on July 18, 2016http://atvb.ahajournals.org/Downloaded from by guest on July 18, 2016http://atvb.ahajournals.org/Downloaded from by guest on July 18, 2016http://atvb.ahajournals.org/Downloaded from by guest on July 18, 2016http://atvb.ahajournals.org/Downloaded from

2024 Arterioscler Thromb Vasc Biol September 2014

possesses a longer cytoplasmic N terminus containing Ca2+-binding motifs, resulting in its activation by Ca2+ elevation.17 ROS generated by NOXs 1, 2, and 4 mediate angiogenic effects in response to angiogenic factors such as VEGF and angiotensin II.10 However, the exact function of these NOXs and the underlying mechanisms by which they mediate their actions remain unknown, partly because of inconsistent pub-lished observations and context-dependent efficacy.

Stromal cell–derived factor 1α (SDF-1α, also called C-X-C motif chemokine ligand 12) is one of the most potent angio-genic CXC chemokines. Our previous studies have shown that SDF-1α requires MKP7 S-nitrosylation to activate c-Jun N-terminal kinase (JNK) 3 and promote endothelial migration and angiogenesis.18 MKP7 belongs to a subgroup of protein tyrosine phosphatases, which are widely recognized as targets of ROS that can be oxidized at redox-sensitive cysteine resi-dues resulting in the inhibition of phosphatase activity after growth factor treatment.19–21 MKP7 possesses a critical cys-teine in its catalytic pocket that is highly sensitive to oxida-tion because of its low pKa.22 Therefore, we hypothesized that NOX-generated ROS may oxidize and inhibit MKP7, thereby regulating SDF-1α–dependent JNK3 activity and angiogen-esis. By performing a series of biochemical and cell biological assays, we have developed strong evidence that suggests that NOXs, including NOX5, are novel positive regulators of SDF-1α–induced endothelial migration and angiogenesis.

Materials and MethodsMaterials and Methods are available in the online-only Supplement.

ResultsSDF-1α Induces Transient Generation of ROS in BAECsOur previous studies demonstrated that MKP7 S-nitrosylation and subsequent inhibition are required for SDF-1α–induced endothelial migration and angiogenesis.18 MKP7 activity can also be modified and inhibited through oxidation of a criti-cal redox-sensitive cysteine in its catalytic pocket.22 NOXs 1, 2, and 4 are the major source of ROS in endothelial cells.23 In addition, they regulate angiogenesis in response to growth factors such as VEGF and angiotensin II.10 However, the exact function of these NOXs remains elusive, and whether NOX-generated ROS regulate SDF-1α–dependent signaling

is unknown. We hypothesized that NOX-generated ROS may be another mechanism by which MKP7 activity is inhibited during SDF-1α–dependent angiogenesis. To investigate the role of NOX-generated ROS in SDF-1α–dependent angio-genesis, we first examined ROS generation in bovine aor-tic endothelial cells (BAECs) after SDF-1α treatment. The fluorescent dyes CM-H

2DCFDA (the chloromethyl deriva-

tive of 2′,7′-dichlorodihydrofluorescein diacetate) and dihy-droethidium were used to monitor and determine the level of hydrogen peroxide and superoxide (O

2·−) production,

respectively, in BAECs after treatment with SDF-1α. When cells were treated with SDF-1α for 5 minutes, an increase in fluorescent dichlorofluorescein (DCF) signal was detected in BAECs by confocal microscopy (Figure 1A) and by flow cytometry (from 0.01% cells positive under control condi-tions to 15.45% cells in response to SDF-1α; Figure 1B; Figure IA in the online-only Data Supplement). Furthermore, this SDF-1α–induced DCF signal was inhibited by polyeth-ylene glycol catalase but not the inhibitors of NOS, L-NAME (L-NG-nitroarginine methyl ester), and that of cyclooxy-genase, diclofenac (Figure IB–IE in the online-only Data Supplement), suggesting that this detected DCF signal is because of hydrogen peroxide generation but not the forma-tion of peroxynitrites or arachidonate metabolites. SDF-1α treatment of BAECs also resulted in O

2·− generation, detected

by the appearance of fluorescent ethidium generated from oxidation of dihydroethidium. SDF-1α–induced superoxide production rapidly increased ≈2 fold, peaking at 5 minutes after the SDF-1α treatment began (Figure 1C). Furthermore, the superoxide level was induced in a dose-dependent man-ner (Figure 1D). To determine whether ROS were generated during SDF-1α–induced endothelial cell migration, a wound scratch assay was performed to visualize the active migra-tory process of BAECs after SDF-1α treatment. Consistent with the observations above, SDF-1α treatment of BAECs resulted in generation of a fluorescent DCF signal that was most prominent at the leading edge of the BAECs where active migration takes place (Figure 1E). These data sug-gest that ROS are generated in response to SDF-1α treatment in BAECs and that this ROS generation correlates with the SDF-1α–dependent migratory response.

Antioxidants Inhibit SDF-1α–Induced JNK3 Activation and Endothelial MigrationTo determine whether the ROS generation seen at the lead-ing edge of the migrating BAECs is necessary for SDF-1α–induced migration, the wound scratch assay was repeated in the presence of the 2 antioxidants N-acetyl-l-cysteine and ebselen. As shown in Figure 2A, SDF-1α–induced BAEC migration was significantly inhibited by pretreatment of cells with either N-acetyl-l-cysteine or ebselen, indicating that ROS generation is required for SDF-1α–induced cell migration.

NOXs Are required for SDF-1α–Induced AngiogenesisRecent evidence indicates that NOXs are the major sources of ROS in endothelial cells.23 Given our observation that

Nonstandard Abbreviations and Acronyms

BAEC bovine aortic endothelial cell

DCF dichlorofluorescein

DPI diphenylene iodonium

eNOS endothelial nitric oxide synthase

HO-1 heme oxygenase-1

JNK3 c-Jun N-terminal kinase 3

NOX NADPH oxidase

ROS reactive oxygen species

SDF-1α stromal cell–derived factor-1α

TNF tumor necrosis factor

VEGF vascular endothelial growth factor

by guest on July 18, 2016http://atvb.ahajournals.org/Downloaded from

Pi et al NOX Is an Angiogenic Regulator 2025

SDF-1α treatment of BAECs resulted in ROS generation that was associated with SDF-1α–induced cell migration, we next examined whether NOXs were also involved in these events. Using the Boyden chamber migration assay, we observed that SDF-1α treatment increased BAEC migration from 13±1 to 48±1 cells per field (Figure 2B) but that SDF-1α–induced migration could be completely inhibited by the 2 general NOX inhibitors diphenylene iodonium (DPI) and apocynin (Figure 2B), indicating that NOX enzyme–depen-dent ROS generation is essential for SDF-1α–mediated BAEC migration.

Four different NOX enzymes (NOXs 1, 2, 4, and 5) are expressed in vascular endothelial cells such as BAECs (Figure IIA in the online-only Data Supplement).10 NOXs 1, 2, and 4 have each been linked to angiogenic responses; however, the underlying molecular mechanisms behind these events are not completely understood. Furthermore, because previous experiments have been largely performed in rodent cells that do not express NOX5, the role of NOX5 in angiogenesis is yet to be determined. Therefore, to deter-mine which NOX was responsible for SDF-1α–induced angiogenesis, we first examined the effect of knockdown of NOX5 or p22phox, a crucial component of NOX1/2/4 active enzyme complexes,10 on SDF-1α–induced endothelial migration in BAECs. p22phox or NOX5 small-interfering RNA (siRNA) specifically decreased the RNA and protein levels of p22phox or NOX5, respectively, but not other NOXs in BAECs (Figure IIB and IIC in the online-only Data Supplement). Similar to antioxidants or NOX inhibi-tors, knockdown of either NOX5 or p22phox decreased superoxide generation (Figure IID in the online-only Data Supplement). More excitingly, their knockdown blocked endothelial migration induced by SDF-1α (Figure 2C),

suggesting that multiple NOXs, including NOX5, are involved in SDF-1α–induced endothelial migration. To assess the effect of NOX depletion on tube formation, BAECs were transfected with siRNA specific for either NOX5 or p22phox, and the effect on SDF-1α–induced tube formation in Matrigel was analyzed. Seventy-two hours after transfection with p22phox siRNA, NOX5 siRNA, or control siRNA, BAECs were plated on Matrigel in medium containing 100 ng/mL SDF-1α. SDF-1α treatment signifi-cantly enhanced tube formation in BAECs transfected with control siRNA, with the tube number increasing from 12±3 to 68±10 tubes per field (Figure 2D). In contrast, SDF-1α–induced tube formation was significantly blocked in BAECs transfected with either p22phox siRNA or NOX5 siRNA by 4.5- or 3.4-fold, respectively (Figure 2D). To confirm this antiangiogenic effect of p22phox and NOX5 siRNAs, we also performed a spheroid-sprouting angiogenesis assay using BAECs treated with SDF-1α. This assay measures the sprouting and network formation of gel-embedded aggregated endothelial cells. Endothelial spheroids were prepared with BAECs transfected with different siRNAs and then stimulated with SDF-1α. The total number of sprouts from each spheroid was counted and compared with control cells. As shown in Figure IIE in the online-only Data Supplement, SDF-1α–treated BAECs showed a 3.4-fold increase in sprouts/spheroids compared with con-trol BAECs. However, this increase in SDF-1α–induced sprout formation was inhibited in BAECs transfected with either p22phox or NOX5 siRNA. The data collected from the Matrigel and spheroid-sprouting angiogenesis assays strongly suggest that multiple NOXs, including NOX5 and probably NOX1/2/4, are critical mediators of SDF-1α–dependent angiogenesis in BAECs.

Figure 1. Reactive oxygen species (ROS) are generated after stromal cell–derived factor-1α (SDF-1α) treatment. Bovine aortic endothelial cells (BAECs) were treated with SDF-1α (100 ng/mL) for 5 minutes, and ROS production was detected by CM-H2DCFDA (the chloromethyl derivative of 2′,7′-dichlo-rodihydrofluorescein diacetate; A and B) or dihydroethidium staining (C and D). A, Representative images of DCFDA-stained cells. Scale bar, 10 μm. B, Representative flow cytometry data obtained for H2O2 measurement that was detected as green fluorescence after CM-H2DCFDA staining. This is an overlay picture of 2 samples: the control (blue) and 100 ng/mL SDF-1α treated for 5 minutes (gray). C and D, BAECs were treated with SDF-1α in 100 ng/mL for different time periods (C) or different concentrations for 5 minutes (D). O2·

− production was quantified as ethidium bromide (EB) fluorescence. *P<0.05, compared with control BAECs (n=3). E, Representative images of an endothelial cell monolayer with a scratch wound. Cells were treated with SDF-1α (100 ng/mL) for 5 minutes and fixed for CM-H2DCFDA staining. Scale bar, 50 μm.

by guest on July 18, 2016http://atvb.ahajournals.org/Downloaded from

2026 Arterioscler Thromb Vasc Biol September 2014

NOX5, But Not NOX1/2/4, Is Required for JNK3 Activation Induced by SDF-1αWe have previously demonstrated that JNK3 is an impor-tant component of SDF-1α–mediated endothelial migration and angiogenesis.18 Because the data above demonstrate that NOX-generated ROS also play an essential role in SDF-1α–dependent angiogenesis (Figure 2), we tested whether SDF-1α–induced JNK3 activation was influenced by NOX-dependent ROS generation. All 3 antioxidants, butylated hydroxyanisole, N-acetyl-l-cysteine, and ebselen, inhibited SDF-1α–induced JNK phosphorylation in BAECs (Figure 3A; Figure IIIA in the online-only Data Supplement). SDF-1α–induced JNK3 activation was also inhibited by the NOX inhibi-tors DPI and apocynin (Figure 3B). To test which NOX isoform was involved in SDF-1α–induced JNK3 activation, BAECs were transfected with p22phox siRNA, NOX5 siRNA, or con-trol siRNA. Activation of SDF-1α–induced JNK3 activation was dramatically decreased by knockdown of NOX5 but not by knockdown of p22phox (Figure 3C). Furthermore, ectopic expression of NOX5 in BAECs was sufficient to activate JNK3 at baseline and augmented JNK3 activation after SDF-1α stim-ulation (Figure 3D). The inhibitory effect of NOX inhibitors and NOX5 siRNAs on SDF-1α–induced JNK3 activation was further confirmed in human vascular endothelial cells (Figure IIIB–IIID in the online-only Data Supplement). Furthermore, we also observed that extracellular signal–regulated kinase (ERK) activation in response to SDF-1α was also blocked by NOX inhibitors and NOX5 siRNAs, but not p22phox siRNAs in both BAECs and human vascular endothelial cells (Figure IIIC, IIIE, and IIIF in the online-only Data Supplement). In

summary, these data suggest that NOX5, but not NOX1/2/4, is required and sufficient for SDF-1α–induced JNK3 and ERK activation. In addition to the effects seen by SDF-1α, we also observed that other cytokines, including tumor necrosis factor (TNF) 1α and interleukin-1β but not basic fibroblast growth factor, evoke an increase in superoxide generation in BAECs (Figure IVA–IVC in the online-only Data Supplement). Furthermore, the activation of JNK by TNF1α was inhibited by NOX5 siRNA but not p22phox siRNA (Figure IVD in the online-only Data Supplement). The activation of NOXs 1, 2, and 4 and ROS generation regulate VEGF-induced angiogen-esis.24 However, the involvement of NOX5 has not been tested. To assess the effect of NOX5 depletion on VEGF-induced tube formation, human vascular endothelial cells were transfected with NOX5 siRNA. VEGF increased the formed tube number in control cells. However, the knockdown of NOX5 inhibited tube formation in response to VEGF significantly (Figure IVE and IVF in the online-only Data Supplement). Taken together, we speculate that the NOX5-dependent signaling pathway may represent a relatively general mechanism for growth factors and cytokines.

ROS induce the expression of numerous genes including those that code for antioxidant proteins. One such antioxi-dants, heme oxygenase-1 (HO-1) has previously been linked to SDF-1–induced angiogenesis.25 Therefore, we were interested to test whether HO-1 expression in BAECs is also regulated by NOXs. Excitingly, when BAECs were treated with SDF-1α for 8 hours, the protein level of HO-1 increased dramati-cally, which is consistent with a previous report.25 However, the induction in HO-1 protein level was inhibited in cells that were transfected with NOX5 siRNAs, but not p22phox

Figure 2. Nicotinamide adenine dinucleotide phos-phate oxidases (NOXs) inhibit stromal cell–derived factor-1α (SDF-1α)–induced migratory and angio-genic responses in bovine aortic endothelial cells (BAECs). A, A wound healing assay was performed with BAECs. Cell monolayers were scratched, incu-bated with N-acetyl-l-cysteine (NAC; 10 mmol/L) or ebselen (Ebs; 40 μmol/L) for 30 minutes, and then treated with SDF-1α (100 ng/mL). Twenty-four hours later, cells were fixed and images were taken. The level of cell migration into the scratch wound was quantified as the recovered area. *P<0.05, com-pared with control cells; #P<0.01, compared with control cells that were treated with SDF-1α (n=3). B, Boyden chamber analysis of BAEC migration was performed using 100 ng/mL SDF-1α as the che-moattractant. *P<0.001, compared with control cells without SDF-1α. #P<0.001, compared with control cells with SDF-1α (n=3). C, Boyden chamber analy-sis of BAEC migration was performed using 100 ng/mL SDF-1α as the chemoattractant. BAECs were transfected with p22phox, NOX5, or control small interfering RNA (siRNA). *P<0.01, compared with control cells without SDF-1α. #P<0.01, compared with control cells with SDF-1α (n=3). D, In vitro Matrigel angiogenesis assays were performed with BAECs that were transfected with p22phox, NOX5, or control siRNAs. SDF-1α (100 ng/mL) was used to induce tube formation. Images of formed tubes were taken, and tube numbers were counted. *P<0.001, compared with control cells without SDF-1α. #P<0.05, compared with control cells with SDF-1α. Scale bar, 50 μm. Apo indicates apocynin; and DPI, diphenylene iodonium.

by guest on July 18, 2016http://atvb.ahajournals.org/Downloaded from

Pi et al NOX Is an Angiogenic Regulator 2027

siRNAs (Figure V in the online-only Data Supplement), indi-cating that HO-1 is likely one of the downstream mediators for the angiogenic effect of NOX5 in BAECs.

Hyperglycemia Disrupts SDF-1α–Dependent ROS Generation and JNK3 ActivationIncreased oxidative stress is associated with many chronic diseases, including diabetes mellitus and related cardiovascu-lar diseases.26 SDF-1α is a dominant angiogenic factor under multiple pathological conditions such as diabetic retinopa-thy.27 Therefore, we investigated the effect of elevated ROS level on SDF-1α–dependent signaling when BAECs were cultured under elevated glucose conditions. Comparing the cells cultured under normal glucose concentration (7 mmol/L glucose), the incubation of cells with ≈3.3-fold higher glucose level for 3 days significantly increased the intracellular ROS level (Figure 4A). Reports have demonstrated that NOX1/2/4, but not NOX5, are responsible for the elevation of ROS level in response to high glucose in vascular cells.26,28,29 Interestingly, when cells were cultured under normal glucose concentra-tions, SDF-1α increased JNK3 activation and the antioxidant ebselen blocked it. However, when cells were cultured with a higher concentration of glucose, the response of JNK3 acti-vation to SDF-1α was blunted (Figure 4B). Together, these data suggest that SDF-1α–dependent NOX5 activation and

its downstream signaling are inhibited under hyperglycemic conditions and that the regulatory roles of NOX1/2/4 and NOX5 are likely altered when cellular glucose homeostasis is disrupted.

NOX5 Is Associated With MKP7 in BAECs and Inhibits MKP7 ActivityThe preceding experiments indicate that NOX5 is the domi-nant NOX responsible for SDF-1α–induced JNK3 activation. However, the molecular mechanisms underlying this NOX5-dependent signaling event are not clear. S-nitrosylation of the catalytic cysteine of MKP7, which blocks its phospha-tase activity, is required for the SDF-1α–induced activation of JNK3.18 Interestingly, the labile cysteine at the catalytic center of MKPs is also susceptible to oxidation.22 Oxidation of this cysteine by TNFα-generated ROS is critical for acti-vation of JNK1 and the apoptotic effects of TNFα.22 This prompted us to examine whether ROS generated by NOX5 could result in oxidation of MKP7. In order for NOX5-generated ROS to efficiently oxidize MKP7, NOX5 and MKP7 would need to be in close proximity.30 To deter-mine the relative localization of MKP7 and NOX5, Flag-tagged MKP7 and green fluorescent protein (GFP)-tagged NOX5 were transfected into BAECs and their localization determined using confocal microscopy. Both flag-tagged

Figure 3. Nicotinamide adenine dinu-cleotide phosphate oxidase (NOX) 5 is required for stromal cell–derived factor-1α (SDF-1α)–induced c-Jun N-terminal kinase 3 (JNK3) activation in bovine aortic endothelial cells (BAECs). A, BAECs were incubated with butylated hydroxyanisole (BHA; 200 μmol/L), N-acetyl-l-cysteine (NAC; 10 mmol/L), or ebselen (Ebs; 40 μmol/L) for 30 minutes and then treated with SDF-1α (100 ng/mL) for 10 minutes to activate JNK3. Cell lysates were used for Western blotting to detect the activa-tion of JNK3 using the phospho-JNK antibody. The phosphorylation of JNK was interpreted as JNK3 phosphorylation because JNK3 is the major JNK moiety activated by SDF-1α in endothelial cells in our system.18 B, BAECs were incubated with NOX inhibitor diphenylene iodonium (DPI; 10 μmol/L) or apocynin (Apo; 5 mmol/L) for 30 minutes and then treated with SDF-1α (100 ng/mL) for 10 minutes to activate JNK3. Cell lysates were used for Western blotting to detect the activa-tion of JNK3. C, BAECs were transfected with p22phox small interfering RNA (siRNA), NOX5 siRNA, or control siRNA. Three days later, cells were treated with SDF-1α (100 ng/mL) for 10 minutes to activate JNK3. Cell lysates were used for Western blotting to detect the activation of JNK3. D, BAECs were transfected with Myc-tagged NOX5 or control vector. Two days later, cells were treated with SDF-1α (100 ng/mL) for 10 minutes to activate JNK3. Cell lysates were used for Western blotting to detect the activation of JNK3. *P<0.05, compared with control cells without SDF-1α; #P<0.05, compared with control cells with SDF-1α (n=3).

by guest on July 18, 2016http://atvb.ahajournals.org/Downloaded from

2028 Arterioscler Thromb Vasc Biol September 2014

MKP7 (Figure 5B) and GFP-tagged NOX5 colocalized to the plasma membrane and cytoplasm of BAECs (Figure 5A), indicating a spatial proximity that would be favorable for NOX5 oxidation of MKP7. To determine whether these 2 proteins associate directly, immunoprecipitation assays were performed on transfected cells. As shown in Figure 5C, Flag-tagged MKP7 was detected in a protein complex immuno-precipitated with Myc-tagged NOX5 but not Myc-p22phox, suggesting that MKP7 was associated with NOX5 spe-cifically in human embryonic kidney 293 cells. The reverse coimmunoprecipitation experiment confirmed the associa-tion of MKP7 with NOX5 (Figure VIA in the online-only Data Supplement). Coimmunoprecipitation of endogenous MKP7 and NOX5 from BAECs confirmed that, under basal conditions, MKP7 and NOX5 were associated together (Figure 5D). However, after SDF-1α treatment, the proteins dissociated, suggesting a possible mechanism for revers-ible SDF-1α regulation of MKP7 (Figure 5D and 5E). To identify which region of MKP7 was responsible for NOX5 binding, a series of MKP7 deletion mutants (Figure 5B) were engineered. Using coimmunoprecipitation experiments once again, we determined that MKP7 (318–665) but not MKP7 (1–317) was associated with NOX5 (Figure VIB in the online-only Data Supplement). Further mutations of the 318–665 fragment revealed that MKP7 (415–440; indicated by black arrow in Figure 5B) was required for association with NOX5 (Figure 5F). Taken together, these data demon-strate that MKP7 and NOX5 associate in BAECs and that this association is dynamically regulated by SDF-1α.

Having established that NOX5 and MKP7 do associate with one another in BAECs, we next sought to determine whether ROS generated from SDF-1α–induced activation of NOX5 oxidizes MKP7, resulting in the inhibition of MKP7 phosphatase activity and subsequent activation of JNK3. An in vitro MKP7 activity analysis was performed using active JNK3 protein and MKP7 concentrated from cell lysates after our established protocol.18 At baseline, Flag-tagged MKP7 protein demonstrated strong phosphatase activity as shown by the decrease in JNK3 phosphorylation (Figure 6A). Similar to our previous report,18 the treatment of SDF-1α

blocked 42.8% of MKP7 phosphatase activity (Figure 6A and 6B). However, when cells were treated with the NOX inhibitor DPI, MKP7 retained high phosphatase activity at both baseline and after SDF-1α treatment of cells (Figure 6A and 6B). Similarly, NOX5 knockdown via transfection of cells with NOX5-specific siRNAs relieved the inhibitory effect of SDF-1α on MKP7 phosphatase activity (Figure 6C and 6D). These data suggest that ROS generated by NOX5 inhibit MKP7 phosphatase activity, offering one plausible mechanism by which NOX5 promotes SDF-1α–induced JNK3 activation.

Our previous study has demonstrated that MKP7 can be nitrosylated by nitric oxide generated by eNOS after eNOS is activated in response to SDF-1α treatment.18 Similar to eNOS activation leading to the inhibition of MKP7 phos-phatase activity,18 NOX5 activation blocked MKP7 activ-ity and enhanced JNK3 activation. Therefore, we tested whether the blockage of both eNOS and NOX5 activities would result in synergistic inhibitory effects on JNK3 acti-vation. BAECs were transfected with both siRNAs of eNOS and NOX5, and the changes in JNK3 activity were deter-mined. Interestingly, the knockdown of NOX5 or eNOS blocked the activation of JNK3. However, the knockdown of both NOX5 and eNOS did not further significantly inhibit the activation of JNK3 (Figure VII in the online-only Data Supplement). It suggests that NOX5 and eNOS might work on the same pathway upstream of JNK3 activation. Reports have shown that NOX5 induces eNOS activity possibly by enhancing eNOS:hsp90 binding or increasing intracellular calcium level.31 Another regulatory mechanism for their interaction could be that both eNOS and NOX5 regulate MKP7 activation.

Taken all together, our data identify NOX5 as a novel com-ponent of SDF-1α–dependent JNK3 activation and support the hypothesis that NOX-generated ROS are important mediators of SDF-1α–dependent angiogenesis. More importantly, both nitrous and oxidative stresses, and therefore the intracellular redox status, are important for SDF-1α–dependent signaling and endothelial activation in response to pathophysiologic conditions such as hyperglycemia.

Figure 4. The response to stromal cell–derived factor-1α (SDF-1α) is blunted when bovine aortic endothelial cells (BAECs) were cultured in the presence of high glucose concentrations. BAECs were cultured in media containing either normal glucose concentration (7 mmol/L glucose+23 mmol/L mannitol) or high glucose concentra-tion (30 mmol/L glucose). Three days later, cells were incubated with ebselen (Ebs; 40 μmol/L) for 30 minutes and then treated with SDF-1α (100 ng/mL) for 5 minutes to determine reactive oxygen species generation and for 10 minutes to determine c-Jun N-terminal kinase 3 (JNK3) activation. A, O2·

− production was quantified as ethidium bromide fluorescence. B, JNK3 activation was determined by Western blotting. *P<0.05, compared with control BAECs with low glucose and without SDF-1α. #P<0.05, com-pared with control cells without Ebs (n=3). p-JNK indicates phospho-JNK; and NS, not significant.

by guest on July 18, 2016http://atvb.ahajournals.org/Downloaded from

Pi et al NOX Is an Angiogenic Regulator 2029

DiscussionNOXs are well accepted as being novel players in the regu-lation of angiogenesis. However, the underlying mechanisms by which they exert their actions have not been completely elucidated. Using NOX inhibitors and NOX5-specific siRNA, we have demonstrated a requisite role for NOX5 in the regula-tion of SDF-1α–dependent JNK3 activation and subsequent endothelial activation and angiogenesis. This is the first report demonstrating that NOX-generated ROS are required for the angiogenic effects of SDF-1α in endothelial cells. Furthermore, we observed that multiple cytokines including SDF-1α, TNF1α, and interleukin-1β induced superoxide gen-eration, and the activation of JNK by TNF1α was inhibited by NOX5 siRNA (Figure IV in the online-only Data Supplement). This recapitulates previous evidence demonstrating that NOX-generated ROS play an important role in angiogenic signaling in response to growth factors and cytokines.10 Given the lack

of NOX5 expression in rodents, the functional roles of NOX5 have not been well studied. However, our data hint at an unex-pected role for NOX5 in the regulation of human pathophysi-ology that is associated with angiogenesis.

Increased oxidative stress is associated with many chronic diseases, including diabetes mellitus and related cardiovascular diseases.26 Our observation that SDF-1α–dependent JNK3 acti-vation was blunted under hyperglycemic conditions (Figure 4) indicates that NOX-dependent ROS generation normally required for angiogenesis is dysregulated in a microenvironment where significantly higher levels of oxidative stress are present. Therefore, both beneficial and detrimental roles of ROS, whose generation might be because of the activation of different enzyme and nonenzyme systems, on endothelial pathophysiology need further clarification to improve the efficacy and safety of thera-peutic strategies involving use of antioxidants and NOX inhibi-tors. In addition, NOX5 is upregulated in some cancers, such as

Figure 5. Nicotinamide adenine dinucleotide phosphate oxidase (NOX) 5 is associated with MKP7. A, Bovine aortic endothelial cells (BAECs) were transfected with Flag-MAP kinase phosphatase 7 (MKP7) and green fluorescent protein (GFP)-NOX5. Two days later, cells were fixed and analyzed by confocal microscopy. Scale bar, 5 μm. B, A schematic representation of full-length Flag-MKP7 (1–665) and its deletion mutant constructs. The black arrow indicates the fragment of MKP7 responsible for its minimal association with NOX5. C, Human embryonic kidney (HEK) 293 cells were transfected with Flag-MKP7, Myc-NOX5, or Myc-p22phox constructs. One day later, total cell lysates (TCL) were used for immunoprecipitation for Myc-NOX5 and Myc-p22phox and immunoblotting for Flag-MKP7. D, BAECs were treated with stromal cell–derived factor-1α (SDF-1α; 100 ng/mL) for 5 minutes and harvested for immunoprecipitation for NOX5 and immunoblotting for MKP7. E, The MKP7 proteins associated with NOX5 in BAECs were quantified as the fold change, compared with the control cells. *P<0.05, compared with control cells without SDF-1α (n=3). F, HEK293 cells were transfected with different Flag-MKP7 mutants and Myc-NOX5 constructs. One day later, TCL were used for immunoprecipitation for Myc-NOX5 and immunoblotting for Flag-MKP7. JNK indicates c-Jun N-terminal kinase 3; and p-JNK, phosphor-JNK.

by guest on July 18, 2016http://atvb.ahajournals.org/Downloaded from

2030 Arterioscler Thromb Vasc Biol September 2014

prostate cancer, esophageal adenocarcinoma, and hairy cell leu-kemia.32–34 DPI, a general NOX inhibitor, potently inhibits the migration, proliferation, and invasion of prostate cancer cells by modulating the activity of growth signaling pathways, includ-ing ERK1/2 and p38, and causing cell cycle arrest.35 Based on our finding that NOX-generated ROS participate in angiogenic events, it is also possible that DPI blocks angiogenesis within the tumor, thereby increasing its usefulness as an antitumor therapy. Similarly, blockade of the SDF-1α/CXCR4 signaling axis, including NOX5, may represent a promising additional or alternative target for the treatment of some forms of cancer such as rectal carcinoma and colorectal cancer in which SDF-1α levels are seen to increase after recurrence of tumors after chemotherapy and anti-VEGF therapy.36,37 Targeting NOX5 in the SDF-1α pathway may also be useful in treatment against atherosclerosis. In the vascular system, NOX5 is predominantly expressed in endothelial cells but is also detected in smooth muscle cells. In atherosclerotic aortas, the expression level of endothelial NOX5 is increased.38 Given our data demonstrating that NOX5 is involved in angiogenic responses in endothelial cells, it is possible that the elevation in NOX5 expression in ath-erosclerotic lesions exacerbates plaque formation by promot-ing new vessel formation. Blockade of NOX5 may therefore be one mechanism by which the development of atherosclerotic plaques can be cured.

The data presented in this report are the first demonstra-tion that NOXs are activated by the SDF-1α/CXCR4 signaling axis. All forms of NOX (NOXs 1, 2, 4, and 5) are expressed in endothelial cells.10 Previous studies have demonstrated that p22phox is required for activity of NOX1, NOX2, and NOX4.10 Because knockdown of p22phox also blocked SDF-1α–induced endothelial migration and angiogenesis, we con-clude that NOX1, NOX2, and NOX4, in addition to NOX5, can be activated by SDF-1α in endothelial cells. That said, their activation and downstream effectors are likely to be different

based on their individual molecular structures. NOX5 pos-sesses additional cytosolic N-terminal EF hands for binding calcium, rendering it able to be activated by calcium, prob-ably because of the fact that the binding of calcium to these EF hands, a common Ca2+-binding motif, relieves the auto-inhibitory loop on the activity domain of NOX5.39 CXCR4, a G-protein–coupled receptor, is ligated by SDF-1α, and its activation initiates a series of signaling processes including an increase in calcium influx.40 Alternative calcium-independent mechanisms have also been reported to promote calcium sen-sitization and activation of NOX5.39 Whether calcium influx or other mechanisms of calcium sensitization is required for NOX5 activation by SDF-1α remains to be determined. In contrast to NOX5, the activation of NOX1, NOX2, and NOX4 is regulated differentially by small G proteins and multiple kinases. Rac GTPases are required for both NOX1 and NOX2 activity.41–44 Protein kinase C, phosphatidylinositide 3-kinase, and Src are kinases that regulate the assembly of the NOX2 oxidase complex by recruiting p47phox bound to p67phox and p40phox; Akt, protein kinase C, and protein kinase A negatively modulate NOX1 activity.10 Unlike NOX1 and NOX2, NOX4 exhibits constitutive activity. NOX4 agonists acutely increase its activity, probably through the fast induc-tion of protein expression, increasing the available NADPH pool, or by promoting NOX4’s association with its recep-tors.10 The diverse range of activators and repressors of NOX 1, 2 and 4 activity suggest that numerous upstream mediators could be activated in response to SDF-1α. Therefore, further experiments are needed to characterize the detailed molecular mechanisms by which NOX1, NOX2, and NOX4 are activated by SDF-1α.

The downstream mediators of NOX1, NOX2, NOX4, and NOX5 responsible for promoting the effects of SDF-1α–induced endothelial migration and angiogenesis are unknown. By using NOX inhibitors and specific siRNAs of

Figure 6. MKP7 activity is regulated by nicotin-amide adenine dinucleotide phosphate oxidase (NOX) 5. A, Results of an in vitro phosphatase assay using active c-Jun N-terminal kinase 3 (JNK3) protein as the substrate. Bovine aortic endothelial cells (BAECs) were transfected with Flag-MKP7 and then treated with diphenylene iodonium (DPI; 10 μmol/L) for 30 minutes and followed by stromal cell–derived factor-1α (SDF-1α; 100 ng/mL) treatment for 5 minutes. The cell lysates concentrated for Flag-tagged MKP7 pro-tein were used for in vitro phosphatase assays. B, Quantitative analysis of results from in vitro phosphatase assays of MKP7 in Figure 5A based on 3 independent experiments. *P<0.05, com-pared with control cells. C, Results of an in vitro phosphatase assay using active JNK3 protein as the substrate. BAECs were transfected with Flag-MKP7, and NOX5 or control small interfering RNA (siRNA) and then treated with SDF-1α for 5 min-utes. The cell lysates concentrated for Flag-tagged MKP7 protein were used for in vitro phosphatase assays. D, Quantitative analysis of results from in vitro phosphatase assays of MAP kinase phospha-tase 7 (MKP7) in Figure 5C based on 3 indepen-dent experiments. *P<0.05, compared with control cells. NS indicates not significant; and p-JNK3 indicates phosphor-JNK3.

by guest on July 18, 2016http://atvb.ahajournals.org/Downloaded from

Pi et al NOX Is an Angiogenic Regulator 2031

NOX5 and p22phox, we have demonstrated that JNK3 and ERK are specific mediators for NOX5 in the regulation of SDF-1α–induced angiogenesis (Figures 2 and 3; Figures II and III in the online-only Data Supplement). Furthermore, we observed that the induction of HO-1, a known mediator of SDF-1α–induced angiogenesis, is mediated by NOX5 (Figure V in the online-only Data Supplement). However, whether JNK3 and ERK act as upstream mediators in response to SDF-1α to induce HO-1 expression still needs further investigation. Other than the downstream mediators that are mentioned above, multiple other pathways have been characterized as downstream mediators of NOX1/2/4 in the regulation of cell proliferation and migration. For example, Akt is reported to mediate NOX-promoted cell proliferation responses,6,45–47 whereas the PPARα (peroxisome proliferator-activated receptor-alpha) pathways have been implicated as downstream effectors for NOX1’s effect on cell migration.48 The interaction of NOX2 and the scaffold protein IQGAP1 (IQ motif containing GAP1) leads to the accumulation of NOX2 at the leading edge of migrating endothelial cells,49 and NOX4 is reported to regulate eNOS expression in endo-thelial cells.50 Whether these pathways are involved in cell migration and proliferation associated with SDF-1α activa-tion of these NOXs still needs further investigation.

The data presented in this report are a continuation of our recent observations that JNK3 is a major mediator of SDF-1α–dependent endothelial migration and that MKP7, the JNK3 phosphatase, is nitrosylated by nitric oxide and this nitrosyl-ation leads to inhibition of its phosphatase activity.18 Here, we present data demonstrating that MKP7 activity is also inhibited by NOX inhibitors or NOX5 siRNAs, suggesting that the intra-cellular redox status is critical for maintaining cellular func-tions for endothelial cells. Cysteine oxidation is one important mechanism by which ROS modulate protein function, and the oxidation of the catalytic cysteine on MKP7 has been reported to be required for TNFα-induced JNK activation.22,51 Therefore, we propose that MKP7 is oxidized by NOX5-generated ROS after SDF-1α administration, which in turn inhibits MKP7 phosphatase activity and sustains JNK3 activation. All MKPs share this cysteine-containing catalytic motif, suggesting they are similarly regulated by the intracellular redox status. In addition to cysteine oxidation, other amino acids such as lysine, proline, arginine, and threonine can also be oxidized, and these modifications are recognized as being important for the pathogenesis of many degenerative diseases.52 Future stud-ies aimed at further characterizing the role of protein oxida-tion may provide essential information that helps to explain the pathophysiologic mechanism of oxidative stress.

AcknowledgmentsWe thank Robert Bagnell, Victoria J. Madden, and Steven J. Ray in the UNC Microscopy Services Laboratories for help with immuno-fluorescence microscopy experiments.

Sources of FundingThis work was supported in part by American Heart Association NCRP (National Center Research Program) Science Development Grant 0930261N (to X. Pi), National Institutes of Health (NIH) R01 HL112890 (to X. Pi), and NIH R01 HL061656 (X. Pi).

DisclosuresNone.

References 1. Chen W, Gabel S, Steenbergen C, Murphy E. A redox-based mechanism

for cardioprotection induced by ischemic preconditioning in perfused rat heart. Circ Res. 1995;77:424–429.

2. Skyschally A, Schulz R, Gres P, Korth HG, Heusch G. Attenuation of ischemic preconditioning in pigs by scavenging of free oxyradicals with ascorbic acid. Am J Physiol Heart Circ Physiol. 2003;284:H698–H703.

3. Tanaka K, Weihrauch D, Kehl F, Ludwig LM, LaDisa JF, Jr, Kersten JR, Pagel PS, Warltier DC. Mechanism of preconditioning by isoflu-rane in rabbits: a direct role for reactive oxygen species. Anesthesiology. 2002;97:1485–1490.

4. Ushio-Fukai M, Tang Y, Fukai T, Dikalov SI, Ma Y, Fujimoto M, Quinn MT, Pagano PJ, Johnson C, Alexander RW. Novel role of gp91(phox)- containing NAD(P)H oxidase in vascular endothelial growth factor-induced signaling and angiogenesis. Circ Res. 2002;91:1160–1167.

5. Colavitti R, Pani G, Bedogni B, Anzevino R, Borrello S, Waltenberger J, Galeotti T. Reactive oxygen species as downstream mediators of angio-genic signaling by vascular endothelial growth factor receptor-2/KDR. J Biol Chem. 2002;277:3101–3108.

6. Lelkes PI, Hahn KL, Sukovich DA, Karmiol S, Schmidt DH. On the pos-sible role of reactive oxygen species in angiogenesis. Adv Exp Med Biol. 1998;454:295–310.

7. Maulik N. Redox regulation of vascular angiogenesis. Antioxid Redox Signal. 2002;4:783–784.

8. Cai T, Fassina G, Morini M, Aluigi MG, Masiello L, Fontanini G, D’Agostini F, De Flora S, Noonan DM, Albini A. N-acetylcysteine inhibits endothelial cell invasion and angiogenesis. Lab Invest. 1999;79:1151–1159.

9. Wheeler MD, Smutney OM, Samulski RJ. Secretion of extracellular superoxide dismutase from muscle transduced with recombinant adeno-virus inhibits the growth of B16 melanomas in mice. Mol Cancer Res. 2003;1:871–881.

10. Lassègue B, San Martín A, Griendling KK. Biochemistry, physiology, and pathophysiology of NADPH oxidases in the cardiovascular system. Circ Res. 2012;110:1364–1390.

11. Bánfi B, Clark RA, Steger K, Krause KH. Two novel proteins activate superoxide generation by the NADPH oxidase NOX1. J Biol Chem. 2003;278:3510–3513.

12. Geiszt M, Lekstrom K, Witta J, Leto TL. Proteins homologous to p47phox and p67phox support superoxide production by NAD(P)H oxidase 1 in colon epithelial cells. J Biol Chem. 2003;278:20006–20012.

13. Ueno N, Takeya R, Miyano K, Kikuchi H, Sumimoto H. The NADPH oxidase Nox3 constitutively produces superoxide in a p22phox-dependent manner: its regulation by oxidase organizers and activators. J Biol Chem. 2005;280:23328–23339.

14. Bedard K, Krause KH. The NOX family of ROS-generating NADPH oxi-dases: physiology and pathophysiology. Physiol Rev. 2007;87:245–313.

15. Sumimoto H. Structure, regulation and evolution of Nox-family NADPH oxidases that produce reactive oxygen species. FEBS J. 2008;275:3249–3277.

16. Serrander L, Cartier L, Bedard K, Banfi B, Lardy B, Plastre O, Sienkiewicz A, Fórró L, Schlegel W, Krause KH. NOX4 activity is determined by mRNA levels and reveals a unique pattern of ROS generation. Biochem J. 2007;406:105–114.

17. Bánfi B, Tirone F, Durussel I, Knisz J, Moskwa P, Molnár GZ, Krause KH, Cox JA. Mechanism of Ca2+ activation of the NADPH oxidase 5 (NOX5). J Biol Chem. 2004;279:18583–18591.

18. Pi X, Wu Y, Ferguson JE, 3rd, Portbury AL, Patterson C. SDF-1alpha stimulates JNK3 activity via eNOS-dependent nitrosylation of MKP7 to enhance endothelial migration. Proc Natl Acad Sci U S A. 2009;106:5675–5680.

19. Meng TC, Fukada T, Tonks NK. Reversible oxidation and inactivation of protein tyrosine phosphatases in vivo. Mol Cell. 2002;9:387–399.

20. Kwon J, Lee SR, Yang KS, Ahn Y, Kim YJ, Stadtman ER, Rhee SG. Reversible oxidation and inactivation of the tumor suppressor PTEN in cells stimulated with peptide growth factors. Proc Natl Acad Sci U S A. 2004;101:16419–16424.

21. Lee SR, Kwon KS, Kim SR, Rhee SG. Reversible inactivation of protein-tyrosine phosphatase 1B in A431 cells stimulated with epidermal growth factor. J Biol Chem. 1998;273:15366–15372.

by guest on July 18, 2016http://atvb.ahajournals.org/Downloaded from

2032 Arterioscler Thromb Vasc Biol September 2014

22. Kamata H, Honda S, Maeda S, Chang L, Hirata H, Karin M. Reactive oxygen species promote TNFalpha-induced death and sustained JNK acti-vation by inhibiting MAP kinase phosphatases. Cell. 2005;120:649–661.

23. Ushio-Fukai M, Alexander RW. Reactive oxygen species as mediators of angiogenesis signaling: role of NAD(P)H oxidase. Mol Cell Biochem. 2004;264:85–97.

24. Ushio-Fukai M, Nakamura Y. Reactive oxygen species and angio-genesis: NADPH oxidase as target for cancer therapy. Cancer Lett. 2008;266:37–52.

25. Deshane J, Chen S, Caballero S, Grochot-Przeczek A, Was H, Li Calzi S, Lach R, Hock TD, Chen B, Hill-Kapturczak N, Siegal GP, Dulak J, Jozkowicz A, Grant MB, Agarwal A. Stromal cell-derived factor 1 pro-motes angiogenesis via a heme oxygenase 1-dependent mechanism. J Exp Med. 2007;204:605–618.

26. Sedeek M, Montezano AC, Hebert RL, Gray SP, Di Marco E, Jha JC, Cooper ME, Jandeleit-Dahm K, Schiffrin EL, Wilkinson-Berka JL, Touyz RM. Oxidative stress, Nox isoforms and complications of diabetes–poten-tial targets for novel therapies. J Cardiovasc Transl Res. 2012;5:509–518.

27. Butler JM, Guthrie SM, Koc M, Afzal A, Caballero S, Brooks HL, Mames RN, Segal MS, Grant MB, Scott EW. SDF-1 is both necessary and suf-ficient to promote proliferative retinopathy. J Clin Invest. 2005;115:86–93.

28. Manea A, Fenyo IM, Florea IC, Raicu M. Distinct roles of NADPH oxi-dase isoforms in high glucose-induced reactive oxygen species formation in human aortic smooth muscle cells. Annals RSCB 2010;XV:26–31.

29. Sonta T, Inoguchi T, Tsubouchi H, Sekiguchi N, Kobayashi K, Matsumoto S, Utsumi H, Nawata H. Evidence for contribution of vascular NAD(P)H oxidase to increased oxidative stress in animal models of diabetes and obesity. Free Radic Biol Med. 2004;37:115–123.

30. Chen K, Craige SE, Keaney JF, Jr. Downstream targets and intracel-lular compartmentalization in Nox signaling. Antioxid Redox Signal. 2009;11:2467–2480.

31. Zhang Q, Malik P, Pandey D, Gupta S, Jagnandan D, Belin de Chantemele E, Banfi B, Marrero MB, Rudic RD, Stepp DW, Fulton DJ. Paradoxical activation of endothelial nitric oxide synthase by NADPH oxidase. Arterioscler Thromb Vasc Biol. 2008;28:1627–1633.

32. Brar SS, Corbin Z, Kennedy TP, Hemendinger R, Thornton L, Bommarius B, Arnold RS, Whorton AR, Sturrock AB, Huecksteadt TP, Quinn MT, Krenitsky K, Ardie KG, Lambeth JD, Hoidal JR. NOX5 NAD(P)H oxi-dase regulates growth and apoptosis in DU 145 prostate cancer cells. Am J Physiol Cell Physiol. 2003;285:C353–C369.

33. Fu X, Beer DG, Behar J, Wands J, Lambeth D, Cao W. cAMP-response element-binding protein mediates acid-induced NADPH oxidase NOX5-S expression in Barrett esophageal adenocarcinoma cells. J Biol Chem. 2006;281:20368–20382.

34. Kamiguti AS, Serrander L, Lin K, Harris RJ, Cawley JC, Allsup DJ, Slupsky JR, Krause KH, Zuzel M. Expression and activity of NOX5 in the circulating malignant B cells of hairy cell leukemia. J Immunol. 2005;175:8424–8430.

35. Kumar B, Koul S, Khandrika L, Meacham RB, Koul HK. Oxidative stress is inherent in prostate cancer cells and is required for aggressive pheno-type. Cancer Res. 2008;68:1777–1785.

36. Xu L, Duda DG, di Tomaso E, Ancukiewicz M, Chung DC, Lauwers GY, Samuel R, Shellito P, Czito BG, Lin PC, Poleski M, Bentley R, Clark JW, Willett CG, Jain RK. Direct evidence that bevacizumab, an anti-VEGF antibody, up-regulates SDF1alpha, CXCR4, CXCL6, and neuropilin 1 in tumors from patients with rectal cancer. Cancer Res. 2009;69:7905–7910.

37. Kim J, Takeuchi H, Lam ST, Turner RR, Wang HJ, Kuo C, Foshag L, Bilchik AJ, Hoon DS. Chemokine receptor CXCR4 expression in colorec-tal cancer patients increases the risk for recurrence and for poor survival. J Clin Oncol. 2005;23:2744–2753.

38. Guzik TJ, Chen W, Gongora MC, Guzik B, Lob HE, Mangalat D, Hoch N, Dikalov S, Rudzinski P, Kapelak B, Sadowski J, Harrison DG. Calcium-dependent NOX5 nicotinamide adenine dinucleotide phosphate oxidase contributes to vascular oxidative stress in human coronary artery disease. J Am Coll Cardiol. 2008;52:1803–1809.

39. Bedard K, Jaquet V, Krause KH. NOX5: from basic biology to signaling and disease. Free Radic Biol Med. 2012;52:725–734.

40. Basu S, Broxmeyer HE. Transforming growth factor-{beta}1 modulates responses of CD34+ cord blood cells to stromal cell-derived factor-1/CXCL12. Blood. 2005;106:485–493.

41. Ueyama T, Geiszt M, Leto TL. Involvement of Rac1 in activation of mul-ticomponent Nox1- and Nox3-based NADPH oxidases. Mol Cell Biol. 2006;26:2160–2174.

42. Cheng G, Diebold BA, Hughes Y, Lambeth JD. Nox1-dependent reactive oxygen generation is regulated by Rac1. J Biol Chem. 2006;281:17718–17726.

43. Zhao X, Carnevale KA, Cathcart MK. Human monocytes use Rac1, not Rac2, in the NADPH oxidase complex. J Biol Chem. 2003;278:40788–40792.

44. Roberts AW, Kim C, Zhen L, Lowe JB, Kapur R, Petryniak B, Spaetti A, Pollock JD, Borneo JB, Bradford GB, Atkinson SJ, Dinauer MC, Williams DA. Deficiency of the hematopoietic cell-specific Rho family GTPase Rac2 is characterized by abnormalities in neutrophil function and host defense. Immunity. 1999;10:183–196.

45. Niu XL, Madamanchi NR, Vendrov AE, Tchivilev I, Rojas M, Madamanchi C, Brandes RP, Krause KH, Humphries J, Smith A, Burnand KG, Runge MS. Nox activator 1: a potential target for modulation of vascular reactive oxygen species in atherosclerotic arteries. Circulation. 2010;121:549–559.

46. Anilkumar N, Weber R, Zhang M, Brewer A, Shah AM. Nox4 and nox2 NADPH oxidases mediate distinct cellular redox signaling responses to agonist stimulation. Arterioscler Thromb Vasc Biol. 2008;28:1347–1354.

47. Peshavariya H, Dusting GJ, Jiang F, Halmos LR, Sobey CG, Drummond GR, Selemidis S. NADPH oxidase isoform selective regulation of endo-thelial cell proliferation and survival. Naunyn Schmiedebergs Arch Pharmacol. 2009;380:193–204.

48. Garrido-Urbani S, Jemelin S, Deffert C, Carnesecchi S, Basset O, Szyndralewiez C, Heitz F, Page P, Montet X, Michalik L, Arbiser J, Rüegg C, Krause KH, Imhof BA, Imhof B. Targeting vascular NADPH oxidase 1 blocks tumor angio-genesis through a PPARα mediated mechanism. PLoS One. 2011;6:e14665.

49. Ikeda S, Yamaoka-Tojo M, Hilenski L, Patrushev NA, Anwar GM, Quinn MT, Ushio-Fukai M. IQGAP1 regulates reactive oxygen species-depen-dent endothelial cell migration through interacting with Nox2. Arterioscler Thromb Vasc Biol. 2005;25:2295–2300.

50. Craige SM, Chen K, Pei Y, Li C, Huang X, Chen C, Shibata R, Sato K, Walsh K, Keaney JF Jr. NADPH oxidase 4 promotes endothelial angio-genesis through endothelial nitric oxide synthase activation. Circulation. 2011;124:731–740.

51. Miki H, Funato Y. Regulation of intracellular signalling through cysteine oxidation by reactive oxygen species. J Biochem. 2012;151:255–261.

52. Berlett BS, Stadtman ER. Protein oxidation in aging, disease, and oxida-tive stress. J Biol Chem. 1997;272:20313–20316.

NADPH oxidases (NOXs) are well accepted as being novel players in the regulation of angiogenesis. However, the underlying mechanisms by which they exert their actions have not been completely elucidated. Using NOX inhibitors and NOX5-specific small interfering RNA, our data identify NOX5 as a novel component of stromal cell–derived factor-1α–dependent c-Jun N-terminal kinase 3 activation and support the hypothesis that NOX-generated reactive oxygen species are important mediators of stromal cell–derived factor-1α–dependent angiogenesis. More importantly, both nitrous and oxidative stresses, and therefore the intracellular redox status, are important for stromal cell–derived factor-1α–dependent signaling and endothelial activation in response to pathophysiologic conditions such as hyperglycemia.

Significance

by guest on July 18, 2016http://atvb.ahajournals.org/Downloaded from

PattersonXinchun Pi, Liang Xie, Andrea L. Portbury, Sarayu Kumar, Pamela Lockyer, Xi Li and Cam

Stimulated Angiogenesis−αDerived Factor-1−Generated Reactive Oxygen Species Are Required for Stromal Cell−NADPH Oxidase

Print ISSN: 1079-5642. Online ISSN: 1524-4636 Copyright © 2014 American Heart Association, Inc. All rights reserved.

Greenville Avenue, Dallas, TX 75231is published by the American Heart Association, 7272Arteriosclerosis, Thrombosis, and Vascular Biology

doi: 10.1161/ATVBAHA.114.3037332014;34:2023-2032; originally published online July 2, 2014;Arterioscler Thromb Vasc Biol. 

http://atvb.ahajournals.org/content/34/9/2023World Wide Web at:

The online version of this article, along with updated information and services, is located on the

http://atvb.ahajournals.org/content/suppl/2014/07/02/ATVBAHA.114.303733.DC1.htmlData Supplement (unedited) at:

  http://atvb.ahajournals.org//subscriptions/

at: is onlineArteriosclerosis, Thrombosis, and Vascular Biology Information about subscribing to Subscriptions:

  http://www.lww.com/reprints

Information about reprints can be found online at: Reprints: 

document. Question and AnswerPermissions and Rightspage under Services. Further information about this process is available in the

which permission is being requested is located, click Request Permissions in the middle column of the WebCopyright Clearance Center, not the Editorial Office. Once the online version of the published article for

can be obtained via RightsLink, a service of theArteriosclerosis, Thrombosis, and Vascular Biologyin Requests for permissions to reproduce figures, tables, or portions of articles originally publishedPermissions:

by guest on July 18, 2016http://atvb.ahajournals.org/Downloaded from

SUPPLEMENTAL DATA

Figure SI. ROS are generated following SDF-1α treatment. A, Flow cytometry image of control cells without

DCFDA-staining. B-D, BAECs were preloaded with CM-H2DCFDA, followed by incubation with 1000 U/ml PEG-catalase (B), 1 mM diclofenac (C) or 10 mM L-NAME (D) for 30 minutes and SDF-1α (100 ng/ml) for 5

minutes or H2O2 (1 mM) for 30 minutes. The intracellular H2O2 production was quantified as CM-H2DCFDA

fluorescence, detected by flow cytometry analysis. *, P<0.05, compared to control BAECs; #, P<0.05; compared to control cells with SDF-1α or H2O2; n=3. E, Representative images of an endothelial cell

monolayer following incubation with 1 mM diclofenac or 10 mM L-NAME for 30 minutes and SDF-1α (100

ng/ml) for 5 minutes. BAECs were fixed for CM-H2DCFDA staining. Scale bar, 10 µm.

!"#$%"& '()*

+,-./0

(12&"34#52

+,-./06'()*

(12&"34#526'()*

!*789

*::;

<<;=*

<<77>

<<<<<<:*::<<<<<<<<<<<<<<<<<<<*:*<<<<<<<<<<<<<<<<<<<<<*:9<<<<<<<<<<<<<<<<<<<<*:7<<<<<<<<<<<<<<<<<<<<*:8

!4&&<2"?#$

@%44#<3&?"%4A24#24

"

#

:

9:

8:

;:

B:

*::

!4&&<2"?#$<CD

E

!$%&

'()* F 9

G 9!$%&

'()* F 9

G 9

!"#$%"& !5$5&5A4,H0@

I

J

I

J

!"#$%"&(12&"34#52

:

9:

8:

;:

B:

*::

!4&&<2"?#$<CD

E

II

I

!$%&

'()* !$

%&

'()*

:

9:

8:

;:

B:

*::

!$%&

'()* !$

%&

'()*

!4&&<2"?#$<CD

E

!"#$%"&+,-./0

I

$

%

H1<4$<5&K<)1L?%4<'*

Figure SII. NOXs are required for SDF-1α-induced angiogenesis in BAECs. A, Total RNA of BAECs was

reverse transcribed to cDNA and analyzed for the RNA levels of NOX1-5 using their specific primers. GAPDH

was used as a housekeeping gene control. B-C, The knockdown of p22phox and NOX5 using their specific

siRNAs. BAECs were transfected with p22phox siRNA, NOX5 siRNA or control siRNA. Three days later, the

total RNA of BAECs were reverse transcribed to cDNA and analyzed for the RNA levels of NOX1, 2, 4, 5 and

p22phox by quantitative PCR (B). In addition, the cell lysates were used for Western blotting to detect

endogenous p22phox and NOX5 (C). D, BAECs were transfected with p22phox, NOX5, or control siRNAs and then treated with SDF-1α in 100 ng/ml for 5 minutes. O2

˙¯ production was quantified as ethidium bromide

fluorescence. *, P<0.05, compared to control BAECs; #, P<0.05; compared to control cells with SDF-1α; n=3.

E, Nox5 and p22phox were required for the sprouting angiogenesis of BAEC spheroids. The spheroid

angiogenesis assays were performed with BAECs that were transfected with p22phox, NOX5 or control siRNAs. 100 ng/ml SDF-1α was used to induce sprout formation. Images of collagen-embedded BAEC

spheroids demonstrate the formation of sprouts after 72 hours of treatment with SDF-1α (100 ng/ml). The

number of sprouts per spheroids were counted and the quantitative data were presented. *, P<0.05; compared to control cells without SDF-1α. #, P<0.05; compared to control cells with SDF-1α; n=3. Scale bar, 100 µm.

!""!#$%

&'()

*+,-.

!""!#$%&&&&&&&&'()*+,-'. /$012$3

/-01,023405-678109"

&'(:

&'("

&'(;

&'()

!""!#$%

<

:

"

=+$.,8$30>-?&* !""!#$%0>-?&*&'()0>-?&*

@ @

! "

0$%*&&0$%"&&0$%4&&0$%5&&0$%*&&67!8#

A)0B!

#

<

:<<<<

"<<<<

=<<<<

+$.,8$30>-?&* !""!#$%0>-?&*&'()0>-?&*

CD0E37$81>+1.+10F2474G

$

@

H

H

5$3I0+#2.610$E0J?&*

K$.,8$3

9L5:

!""!#$% '()*>-?&*M /$012$3

!""!#$% '()*+,-'. /$012$3

K$.,8$39L5:

9!8$7,0.7JB18N>!#18$-I

<

)

:<

:)

%K$.,8$3 9L5:

@

HH

)<<0B!

Figure SIII. NOX5 is required for SDF-1α-induced JNK3 and ERK activation in BAECs and HUVECs. A,

Antioxidant NAC inhibited SDF-1α-induced JNK3 activation in BAECs. BAECs were incubated with NAC in

different concentration for 30 minutes and then treated with SDF-1α (100 ng/ml) for 10 minutes to activate

JNK3. Cell lysates were used for Western blotting to detect the activation of JNK3. *, P<0.05; compared to control cells without SDF-1α. #, P<0.05; compared to control cells with SDF-1α; n=4. B-C, HUVECs were

incubated with NOX inhibitors DPI (10 µM) or apocynin (Apo, 100 µM) for 30 minutes and then treated with

SDF-1α (100 ng/ml) for 10 minutes to activate JNK3 and ERK. Cell lysates were used for Western blotting to

detect the activation of JNK3 (B) and ERK (C). *, P<0.05; compared to control cells without SDF-1α. #,

P<0.05; compared to control cells with SDF-1α; n=3. D, HUVECs were transfected with p22phox siRNA, NOX5

siRNA or control siRNA. Three days later, cells were treated with SDF-1α (100 ng/ml) for 10 minutes to

activate JNK3. Cell lysates were used for Western blotting to detect the activation of JNK3. *, P<0.05; compared to control cells without SDF-1α. #, P<0.05; compared to control cells with SDF-1α; n=3. E, BAECs

were transfected with p22phox siRNA, NOX5 siRNA or control siRNA. Three days later, cells were treated with SDF-1α (100 ng/ml) for 10 minutes to activate JNK3. Cell lysates were used for Western blotting to detect the

activation of ERK. *, P<0.05; compared to control cells without SDF-1α. #, P<0.05; compared to control cells

!

!""!#$%&&&&&&&&&&'()*+,-'. /$012$3!"#$ %%%%&%%%%%%%%%'%%%%%%%%&%%%%%%%%%'%%%%%%%%%&%%%%%%%%'

()*+

)*+,

(-.+

-.+

()*+

)*+,

(-.+

-.+

"/0%%%%%%%%%%%1(23245627!"#$ %%%%&%%%%%%%%'%%%%%%%%&%%%%%%%'%%%%%%%&%%%%%%'

"!"#$ %%%&%%%%&%%%%%%&%%%%&%%%%%%&%%%%%'%%%%'%%%%'%%%%'%%%%%'*18%%%%%%%9%%9:9;%%9:<%%%=%%%%%$9%%%%9%%%9:9;%9:<%%%%=%%%%$9%>?

#

()*+

)*+,

'()*&&&&&&&&&!""!#$%+,-'. /$012$3!"#$ %%%%&%%%%%%%%%'%%%%%%%%&%%%%%%%%%'%%%%%%%%%&%%%%%%%%'

(-.+

-.+

$

/@%A5%B7:%#@CD6A%!,

9

$

=

, 8245627!"#$E E

F

#27G%3HB4CA%2I%

-.+%B35@JB5@24

9:9

$:9

=:9

*18%%%%%%%9%%%%%%9:9;%%%%%9:<%%%%%%%%=%%%%%%%%$9%>?

E E

F

'()*&&&&&&&&&!""!#$%/$012$3+,-'.

%#27G%3HB4CA%2I%

)*++%B35@JB5@24

9

$

=

,

#27G%3HB4CA%2I%

-.+%B35@JB5@24

%#27G%3HB4CA%2I%

)*+,%B35@JB5@24

E

F

EF

8245627!"#$

!""!#$%&&&&&&&&'()*/$012$3+,-'.

8245627!"#$

!""!#$%&&&&&&&&'()*/$012$3+,-'.

8245627!"#$

!""!#$%&&&&&&&&&&'()*+,-'. /$012$3!"#$ %%%%&%%%%%%%%%'%%%%%%%%&%%%%%%%%%'%%%%%%%%%&%%%%%%%%'

%

3245627 "/0 1(2

#27G%8HB4CA !"#$

8245627E

"/0%%%%%%%%%%%1(23245627!"#$ %%%%&%%%%%%%%'%%%%%%%%&%%%%%%%'%%%%%%%&%%%%%%'

&

3245627 "/0 1(29:9

9:K

$:9

$:K

=:9

=:K

#27G%8HB4CA

!"#$8245627E

F F

9

$

=

,

FF

9

$

=

,

<E E

with SDF-1α; n=4. F, HUVECs were transfected with p22phox siRNA, NOX5 siRNA or control siRNA. Three

days later, cells were treated with SDF-1α (100 ng/ml) for 10 minutes to activate JNK3. Cell lysates were used

for Western blotting to detect the activation of ERK. *, P<0.05; compared to control cells without SDF-1α. #,

P<0.05; compared to control cells with SDF-1α; n=3.

Figure SIV. Multiple cytokines induce ROS generation, TNFα-induced JNK3 activation and VEGF-induced

tube formation are mediated by NOX5. A-C, BAECs were treated with 100 ng/ml b-FGF (A), 50 ng/ml TNFα

(B) and 50 ng/ml IL1β (C) for different time periods. O2˙¯ production was quantified as ethidium bromide

fluorescence. *, P<0.05, compared to control BAECs. n=4. D, BAECs were transfected with p22phox siRNA,

NOX5 siRNA or control siRNA. Three days later, cells were treated with 100 ng/ml b-FGF (F) or 50 ng/ml TNFα (T) for 10 minute. Cell lysates were used for Western blotting to detect the activation of JNK3 and ERK.

E-F, In vitro Matrigel angiogenesis assays were performed with HUVECs that were transfected with NOX5 or

control siRNAs. 50 ng/ml VEGF was used to induce tube formation. Images of formed tubes were taken (E)

and tube numbers were counted (F). *, P<0.001; compared to control cells without VEGF. NS, not significant. Scale bar, 50 µm.

!

! "! #! $! %!&'()*"!!!

*#!!!

*$!!!

*%!!!

+,- ./01

! "! #! $! %!&'()!

*!!!!

"!!!!

2!!!!

#!!!!

3!!!!

! "! #! $! %!&'()*#!!!

*$!!!

*%!!!

"!!!!

""!!!

45* ./01

6178

69,8

9,8

178

"

# $!"#$% &'()******+,,***$-./***&'()****+,,***$-./&&&&&&&&&&&&&.&&&&&-&&&.&&&&-&&&.&&&&-&&&.&&&&+&&&.&&&&+&&&&.&&&+

:

:

1;&<=>?@ABCA)CA&DEF>FG

1;&<=>?@ABCA)CA&DEF>FG

1;&<=>?@ABCA)CA&DEF>FG

H-I-./01

J(&AK&E=F&-(L>@A&M#

C?)K@?=& & & N1I-

&01'(0)*!"#$%

$02/*!"#$%

%

!

"!

#!

$!

%!

*!!

&01'(0)*!"#$% $-./*!"#$%

C?)K@?=N1I-

K>HA&)>'HA@&6A@&<(A=O

:

,M

&

Figure SV. HO-1 induction in response to SDF-1α stimulation is mediated by NOX5. BAECs were transfected

with p22phox siRNA, NOX5 siRNA or control siRNA. Three days later, cells were treated with SDF-1α (100

ng/ml) for 8 hours. Cell lysates were used for Western blotting to detect the HO-1 protein level. *, P<0.05; compared to control cells without SDF-1α. #, P<0.05; compared to control cells with SDF-1α; n=5.

!""!#$%&&&&&&&&&&'()*+,-'. /$012$3!"#$ %%%%&%%%%%%%%%'%%%%%%%%&%%%%%%%%%'%%%%%%%%%&%%%%%%%%'

()&$

*+,-.

/0/

/01

$0/

$01

20/

!""!#$%&&&&&&&&&&'()*+,-'. /$012$3

3

4

()&$%56765%89:5;%+<=.>6? @:.,A:5

!"#$

B-%6,%=50%#->CA6%!1

3201

Figure SVI. NOX5 is associated with MKP7. A, NOX5 is associated with MKP7. HEK293 cells were

transfected with Flag-MKP7 and Myc-NOX5 constructs. One day later, total cell lysates were used for

immunoprecipitation for Flag-MKP7 and immunoblotting for Myc-NOX5. B, HEK293 cells were transfected with

different Flag-MKP7 mutants and Myc-NOX5 constructs. One day later, total cell lysates were used for

immunoprecipitation for Myc-NOX5 and immunoblotting for Flag-MKP7.

!"#$%&'()

&*+%,-./

!"#$%&'()000%0000001000000%00000001000000&*+%,-.0/00%0000000%0000001000000012(30!"#$2430&*+

&*+%,-./2430&*+

2430!"#$

(50670#"80!5$9:60;<

! "

!"#$%&'()=>%<</?

!"#$%&'()=@>A%<</?!"#$%&'()=>%@>)?

!"#$%&'()=>%<</?

!"#$%&'()=@>A%<</?

&*+%,BC0/

!"#$%&'()00%000000000000000000000000000%000000&*+%,-.0/00%000000%000000%00000%0000001000001000001000001

>%<</>%@>)@>A%<</

>%<</>%@>)@>A%<</

2(30&*+2430!"#$

2430&*+

2430!"#$

Figure SVII. JNK3 activation in response to SDF-1α stimulation is mediated by both NOX5 and eNOS. B,

BAECs were transfected with NOX5 siRNA, eNOS siRNA or control siRNA. Three days later, cells were treated with SDF-1α (100 ng/ml) for 10 minutes. Cell lysates were used for Western blotting to detect the

activation of JNK3. *, P<0.05; compared to control cells without SDF-1α. #, P<0.05; compared to control cells

with SDF-1α; n=3.

!"#$%#&'(#)"*+,$#-.

!"#$%"&

/012

0123

4%,'#5"

1678#5"

$16-#5"

16789

####$16-#5"

4%,'#5"1678#5"

$16-#5"

16789

####$16-#5"

:(:

:(8

;(:

;(8

<(:

<(8=>?%,>' -@);

A

BB B

#)>'C#4D&?*$#>E#

0123#&4%"F&%">?

4%,'#5"

1678#5"

$16-#5"

16789

####$16-#5"

4%,'#5"

1678#5"

$16-#5"

16789

####$16-#5"

MATERIAL AND METHODS Reagents Recombinant SDF-1α, FGF, TNFα, IL-1β and VEGF proteins were obtained from R&D Systems (Minneapolis, MN). For Western blotting, NOX5 and p22phox antibodies were obtained from Santa Cruz Biotechnology (Santa Cruz, CA) and the JNK3 antibody from Millipore (Billerica, MA). The MKP7 antibody was purchased from Novus Biologicals (Littleton, CO). Anti-pJNK, JNK, pERK1/2, ERK1/2, HO-1 antibodies were purchased from Cell Signaling (Danvers, MA) and used for Western blotting experiments. The oxidative stress indicators Dihydroethidium (DHE) and CM-H2DCFDA were obtained from Life Technologies (Carlsbad, CA). Antioxidants (butylated hydroxyanisole (BHA), N-acetyl-L-cysteine (NAC), ebselen (Ebs)) and NOX inhibitors (diphenylene iodoniu (DPI), apocynin (Apo)) were purchased from Sigma (St. Louis, MO). The growth factor bFGF was obtained from R&D systems (Minneapolis, MN), and cytokines TNFα and IL1β were obtained from Sigma. Cell culture and transient transfection BAECs (bovine aortic endothelial cells) were purchased from Genlantis (San Diego, CA) and cultured in bovine aortic endothelial cell medium (Genlantis). HUVEC (Human umbilical vein endothelial cells) were purchased from Lonza (Allendale, NJ) and cultured in EGM-2 medium (Lonza). The cells from passages 4-8 were used for experiments. ∼70% confluent cells in 6-cm dishes were transfected for 3 hours with 2 µg of plasmids using 8 µl of Lipofectamine LTX and 8 µl of Plus reagent (Life Technologies). Two days later, cells were treated with SDF-1α in the spent media. For transient expression experiments with 293 cells, 90-100% confluent 293 cells in 10-cm dishes were transfected 24 hours after plating with 2 µg of plasmid using 20 µl of Lipofectamine 2000 (Life Technologies). siRNA design and transient transfection The stealth siRNA duplexes were obtained from Life Technologies. The siRNA against bovine p22phox is a duplex of 5'- GCGUAUUGGUCUGCCUGCUGGAAUA-3'. The siRNAs against bovine NOX5 are a mixture of duplexes of 5'- CACCAGUUCUGUAACAUCAAGUGCU-3' and 5'- UGCCUCAACUUUGACUGCAGCUUCA-3'. eNOS siRNAs are a mixture of duplexes of 5’-CGGUGAAGAUCUCUGCCU CACUCAU-3’, 5’-UGUUGCUGGACUCCUUUCUCUUCCG-3’ and 5’- UACGUAUACGGCUUGUCACCUCCUG-3’. The control siRNA is the StealthTM RNAi negative control duplex (Cat. No. 12935-300) and was purchased from Life Technologies. The siRNAs were transfected into BAECs according to our previous published protocol1. Briefly, for each sample, 2x105 BAECs were transfected with 200 pmol siRNA. The experiments with p22phox, Nox5 or eNOS siRNA-transfected BAECs were performed three days later. RT-PCR and PCR The RNA was reverse transcribed into cDNAs with iScriptTM cDNA synthesis kit (Bio-Rad, Hercules, CA, USA). The specific pairs of primers used for the end-point PCR and real-time PCR are the following: nox1 (forward primer: 5'- tccctttaccctgacctctg-'3 and reverse primer: 5'-cccactgctcggatatgaat-'3), nox2 (forward primer: 5'- ctgccagtgaggacgtgtt-'3 and reverse primer: 5'-gtaaccccgatccctgct-'3), nox3 (forward primer: 5'- aggacagcccctgaagga-'3 and reverse primer: 5'-agtcagtgcagctccgaac-'3), nox4

(forward primer: 5'- ccgtttgcgtcaatcctc-'3 and reverse primer: 5'-tcttctaagcttgtatggtttcca-'3), nox5 (forward primer: 5'- ggtctccgagatgtcctctg-'3 and reverse primer: 5'-gccatcgatgtagcacttga-'3), p22phox (forward primer: 5'- cgcaagaagccgagtgag-'3 and reverse primer: 5'-cactggcatggggttttc-'3; designed by Universal ProbeLibrary Assay Design Center tool from Roche, Indianapolis, IN). The end-point PCR was performed with PlatinumTM PCR supermix (Life technologies) in a Mastercycler EP PCR machine (Eppendorf, Hamburg, Germany) with the following program: 3 min 94°C to activate the polymerase followed by 40 cycles of 45 s 94°C, 45 s 55°C, and 45 s 72°C. GAPDH was the housekeeping gene used for internal control. The real-time PCR was performed with FastStart Universal Probe Master mix and specific primers and probes for each gene (Universal ProbeLibrary Single Probes #6 for nox1, #9 for nox2, #1 for nox3, #42 for nox4, #131 for nox5 and #70 for p22phox) in Roche Lightcycler 480 PCR machine. The reaction mixtures were incubated at 95°C for 10 min followed by 55 cycles at 95°C for 10 sec and 60°C for 30 sec. 18S was used as the housekeeping gene. Immunoprecipitation and immunoblotting analysis Cells were harvested in lysis buffer (1% Triton X-100, 50 mmol/L Tris (pH 7.4), 150 mmol/L NaCl, 1 mmol/L Na3VO4 and 0.1% protease inhibitor mixture; Sigma) and clarified by centrifugation at 15,000 g. Equal amounts of proteins were incubated with a specific antibody overnight at 4 °C with gentle rotation. Protein A/G Plus-agarose beads (Santa Cruz Biotechnology) were used to pull down the antibody complexes. Immune complexes were then separated by SDS-PAGE and analyzed by Western blot. Immunofluorescence Immunofluorescence histochemistry was performed following our previously published protocol2. Cells were fixed in 3.7% paraformaldehyde for 10 min at room temperature. After 3 washes with PBS, the cells were sequentially treated with 0.2% Triton X-100 for 5 min (for permeabilization), with 5% boiled serum for 1 hour (for blocking), then with the primary antibody overnight in the blocking solution. After 3 washes, cells were incubated in the dark with secondary antibody conjugated with Alexa Fluor® 568 (Life Technologies) in blocking solution for 90 min at 37°C. After 3 washes in PBS, the slides were mounted and the fluorescent signal was visualized by confocal laser scanning microscopy (Zeiss Pascal, Zeiss, Germany). Detection of ROS generation ROS were detected by staining BAECs with CM-H2DCFDA or DHE. CM-H2DCFDA is oxidized to green fluorescent DCF (dichlorofluorescein) by H2O2, and DHE is oxidized to red fluorescent ethidium by O2·-. Cells were loaded with 5 µM CM-H2DCFDA or 10 µM DHE for 0.5 hour followed by SDF-1α treatment for the indicated time periods. Cells were then washed with PBS and the mean fluorescence intensity was determined as ROS generation by flow cytometry CyAnTM ADP analyzer (for CM-H2DCFDA, excitation: 485 nm and emission: 530 nm; Beckman-Coulter, Indianapolis, IN) or a Tecan microplate reader (for DHE, excitation: 520 nm and emission: 620 nm). Boyden chamber assay Boyden chamber assays were performed as previously described3. Wound healing assay For detection of cell migration, a wound healing assay was performed as previously described3.

In vitro matrigel angiogenesis assay Endothelial cell tube formation was analyzed with the Matrigel-based tube formation assay as previously described3. Spheroid angiogenesis assay BAECs were transfected with p22phox, NOX5 or control siRNAs. Two days later, BAEC spheroids were generated as previously described4, 5. Spheroids were cultured in the polymerized gel of neutralized collagen and carboxymethylcellulose at a ratio of 1:1, and containing 100 ng/ml SDF-1α. Images of the sprouts were taken at 4x magnification. The number of sprouts per spheroid was counted. For each experimental condition, at least 10 spheroids were analyzed. In vitro phosphatase assay In vitro phosphatase reactions were carried out on Flag-MKP7 protein immunoprecipitated from transfected cells and results visualized using Western blot analysis with phospho-specific JNK and JNK3 antibodies following our previous protocol1. Statistical analysis Data are shown as the mean ± SEM for 3 to 4 separate experiments. Differences were analyzed with Student’s t-test or ANOVA and followed by a post hoc test with Bonferroni correction when needed. Values of P≤0.05 were considered statistically significant. REFERENCES 1. Pi X, Wu Y, Ferguson JE, 3rd, Portbury AL, Patterson C. SDF-1alpha stimulates JNK3 activity via eNOS-dependent nitrosylation of MKP7 to enhance endothelial migration. Proc Natl Acad Sci U S A. 2009;106:5675-5680. 2. Pi X, Ren R, Kelley R, Zhang C, Moser M, Bohil AB, Divito M, Cheney RE, Patterson C. Sequential roles for myosin-X in BMP6-dependent filopodial extension, migration, and activation of BMP receptors. J Cell Biol. 2007;179:1569-1582. 3. Pi X, Garin G, Xie L, Zheng Q, Wei H, Abe J, Yan C, Berk BC. BMK1/ERK5 is a novel regulator of angiogenesis by destabilizing hypoxia inducible factor 1alpha. Circ Res. 2005;96:1145-1151. 4. Aitsebaomo J, Srivastava S, Zhang H, Jha S, Wang Z, Winnik S, Veleva AN, Pi X, Lockyer P, Faber JE, Patterson C. Recombinant human interleukin-11 treatment enhances collateral vessel growth after femoral artery ligation. Arterioscler Thromb Vasc Biol. 2011;31:306-312. 5. Laib AM, Bartol A, Alajati A, Korff T, Weber H, Augustin HG. Spheroid-based human endothelial cell microvessel formation in vivo. Nat Protoc. 2009;4:1202-1215.