Role of Redox Signaling in the Autonomous Proliferative Response of Endothelial Cells to Hypoxia
Endothelial cells exhibit an autonomous proliferative response to hypoxia, independent of paracrine effectors. In cultured endothelial cells of porcine aorta, we analyzed the signaling of this response, with a focus on the roles of redox signaling and the MEK/ERK pathway. Transient hypoxia (1 hour) stimulated proliferation by 61±4% (n=16; P<0.05 versus control), quantified after 24 hours normoxic postincubation. Hypoxia induced an activation of ERK2 and of NAD(P)H oxidase and a burst of reactive oxygen species (ROS), determined by DCF fluorescence. To inhibit the MEK/ERK pathway, we used PD 98059 (PD, 20 μmol/L); to downregulate NAD(P)H oxidase, we applied p22phox antisense oligonucleotides; and to inhibit mitochondrial ROS generation, we used the ubiquinone derivate mitoQ (MQ, 10 μmol/L). All three inhibitions suppressed the proliferative response: PD inhibited NAD(P)H oxidase activation; p22phox antisense transfection did not inhibit ERK2 activation, but suppressed ROS production; and MQ inhibited ERK2 activation and ROS production. The autonomous proliferative response depends on the MEK/ERK pathway and redox signaling steps upstream and downstream of ERK. Located upstream is ROS generation by mitochondria, downstream is NAD(P)H oxidase.
Deprivation of tissues from oxygen represents one of the strongest stimuli for endothelial cell proliferation and the formation of new blood vessels. A proliferative response of endothelial cells to hypoxia can be the result of an upregulation of growth factors in adjacent nonendothelial cells, acting in a paracrine way on endothelial cells. VEGF, a growth factor controlled by the hypoxia-inducible transcription factor HIF-1, is an example. Many studies have indicated that the HIF/VEGF system is important for blood vessel growth during embryogenesis; a recent study showed, however, that it may not be essential.1
Apart from such a secondary response of endothelial cells to the effect of growth factors produced in adjacent cells in response to hypoxia, endothelial cells can also exhibit an autonomous proliferative response to hypoxia that is not dependent on the presence of other cells. Existence of the autonomous response to hypoxia has previously been identified in endothelial cell cultures that are devoid of other cell types,2 but its signaling mechanism has remained unclear.
In the present study, we analyzed the signaling of the autonomous proliferative response of endothelial cells to hypoxia. We focused on redox signaling and the role of reactive oxygen species (ROS) because (1) it was shown before that under hypoxic or ischemic conditions generation of ROS is increased, and (2) ROS were shown to be essential signaling intermediates in growth factor–induced proliferation of endothelial cells.3,4 ROS may originate from different sources within endothelial cells, among these are the mitochondrial respiratory chain and NAD(P)H oxidase. We focused our analysis on these two sources. To interfere with mitochondrial ROS production, we used the radical scavenger mitoQ (MQ), a ubiquinone derivate that is selectively targeted to mitochondria.5 To investigate the role of NAD(P)H oxidase, we used an antisense approach. Hypoxia promotes early activation of the MEK/ERK pathway,6 a well-known pathway mediating the growth factor–induced proliferative response of endothelial cells. We, therefore, investigated the relationship between ROS signaling and the MEK/ERK pathway within the signaling of the autonomous proliferative response to hypoxia.
Materials and Methods
Endothelial Cell Culture
Endothelial cells from porcine aorta were isolated and cultured as described.7 Primary endothelial cells were trypsinized and seeded at density of 5×104 cells/cm2. Experiments were performed with monolayers of about 90% confluence, 3 days after seeding.
Before start of experiments, subconfluent monolayers were incubated for 24 hours in serum-free medium 199 to arrest proliferation. Monolayers were then incubated at 35°C in glucose-free HEPES-buffered solution HBS (composed of [in mmol/L] 25 HEPES, 145 NaCl, 1.25 CaCl2, 2.6 KCl, 1.2 MgCl2, 1.2 KH2PO4; pH 7.4). After 30 minutes in normoxic HBS, hypoxia (1 hour) was performed by superfusion of cells in a gas-tight chamber in HBS, equilibrated with 100% N2. Inhibitors were applied 30 minutes before and during hypoxia.
After hypoxia, cultures were postincubated in serum free normoxic media for another 24 hours, trypsinized, and counted in a Neubauer chamber.
Determination of Reactive Oxygen Species (ROS)
ROS were determined by dichlorofluorescein (DCF) fluorescence by use of 2′,7′-dichlorodihydro-fluoresceine-diacetate (H2DCF-DA). Endothelial cells, cultured on round glass coverslips, were incubated (30 minutes) in HBS with 5 μmol/L H2DCF-DA. Monolayers were then superfused (0.5 mL/min) with HBS equilibrated with N2 containing 5 μmol/L H2DCF-DA. Fluorescence was analyzed using a fluorescence microscope combined with a video imaging system (T.I.L.L. Photonics).
ERK2 Kinase Activation
Activation of ERK2 (p42 kDa MAPK) was analyzed by phosphorylation of the kinase detected by SDS-Page, Western blot analysis, and densitometrical quantification, as previously described.8
Antisense Transfection Experiments
We used an antisense approach to downregulate the expression of p22phox, an essential subunit of the NAD(P)H oxidase. Endothelial monolayers were incubated for 24 hours (500 nmol/L) with two different overlapping sequences of 16-mer length, which were phosphorothioated. Sequence of p22phox antisense: 5′-TCTGTCCCATGGCGAT-3′ and 5′-TGTCCCATGGCGATGC-3′; of nonsense, ie, reversed: 5′-TAGCGGTACCCTGTCT-3′ and 5′-CGTAGCGGTACCCTGT-3′.
Consumption of NADH
Consumption of NADH was determined as described before by Patterson et al9 with minor modifications. DPI-sensitive consumption of NADH was estimated as the difference between samples with or without DPI (10 μmol/L) in each group and corrected for protein content of the culture dishes quantified according to Bradford.10 NADH consumption is given as percentage of consumption of normoxic control cultures.
Culture dishes were from Becton Dickinson; H2DCF-DA was from Molecular Probes; PD 98059 and DPI were obtained from Calbiochem; oligonucleotides, newborn calf serum (NCS), medium 199, penicillin-streptomycin and trypsin-EDTA were from Gibco Life Technologies. The polyclonal ERK2 antibody was from Santa Cruz Biotechnologies; the secondary antibody was from Sigma-Aldrich; and mitoQ was a generous gift from Michael P. Murphy (University of Otago, New Zealand). All other chemicals were of the best available quality, usually analytical grade.
Values are expressed as mean±SEM of cells/dishes taken from at least 3 experiments using independent monolayer preparations. Statistical analysis was performed by one-way ANOVA in conjunction with the Student-Newman-Keuls test.
Transient hypoxia for one hour resulted in increased proliferation of endothelial cells, quantified after 24 hours normoxic postincubation as compared with control cultures (Figure 1). The role of the MEK/ERK pathway, redox signaling by flavoproteins, and mitochondrial ROS formation was evaluated by application of pharmacological inhibitors during the hypoxic period. PD (20 μmol/L) was applied to inhibit MEK, DPI (10 μmol/L) to inhibit flavoproteins, and MQ (10 μmol/L) to scavenge mitochondrial ROS. All three inhibitors abolished the stimulation of proliferation by hypoxia. Control experiments in which these agents were applied for a 1 hour normoxic period did not alter the low basal rate of proliferation (data not shown), with exception of DPI, which slightly increased proliferation by 13±2% (n=16).
It was analyzed if ROS are generated by the endothelial cells during hypoxic exposure and if this was influenced by DPI or MQ. Hypoxia provoked a burst of ROS indicated by a rise of the DCF signal (Figure 2). Both DPI and MQ suppressed the hypoxic rise of the DCF signal, but had no effect during the preincubation period on the basal signal in normoxia.
It was shown before that hypoxia leads to an early upregulation of the pro-proliferative MEK/ERK pathway in endothelial cells.6 In order to analyze the interplay between ROS signaling and the MEK/ERK pathway, we investigated whether the use of DPI or MQ affects a possible activation of ERK2 during hypoxia. Hypoxia was found to activate ERK2 transiently, as indicated by an increase in the degree of its phosphorylation (Figure 3). Administration of DPI or MQ suppressed hypoxic activation of ERK2, but these agents did not change the basal phosphorylation state of ERK2 during normoxic preincubation. The data indicate that some DPI- and MQ-sensitive steps are located upstream of ERK2 in the signaling of hypoxia-induced proliferation.
The last described experiments did not exclude the possibility that ROS generation may also occur downstream of the MEK/ERK signaling. We, therefore, analyzed if ROS formation is affected by addition of the MEK inhibitor PD, which was shown to abolish hypoxic MAPK phosphorylation. As shown in Figure 4, PD abolished the hypoxia-induced rise in ROS but did not alter the low rate of normoxic ROS formation during preincubation. These results indicate that the major source for hypoxic ROS lies downstream of MEK/ERK.
We hypothesized that the endothelial NAD(P)H oxidase represents the source of ROS originating downstream of MEK/ERK. To test this hypothesis, we pretreated the cultures with p22phox antisense oligonucleotides or, as controls, with nonsense oligonucleotides. In p22phox antisense–transfected cultures, the hypoxia-induced rise in ROS formation was diminished, whereas this was unaffected by nonsense oligonucleotide treatment (Figure 5). As a positive control for the antisense transfections, we used angiotensin II (100 nmol/L) stimulation of endothelial cells, which is known to provoke ROS formation from NADP(H) oxidase. Under normoxic conditions, treatment with p22phox antisense oligonucleotides reduced the ROS response to angiotensin II stimulation by 85±5% (n=4); treatment with nonsense oligonucleotides had no effect.
In contrast to its effect on ROS formation, antisense oligonucleotide treatment did not affect the hypoxic activation of ERK2 (Figure 6), confirming that the antisense target, NAD(P)H oxidase, is located downstream of ERK2. To corroborate this finding, we also tested if PD reduces NAD(P)H oxidase activity determined by a NADH consumption assay (Figure 7). This was indeed the case: during hypoxia, NADH consumption of hypoxic endothelial monolayers treated with nonsense oligonucleotides was increased as compared with normoxic controls. PD diminished NADH consumption in those monolayers. In p22phox antisense-treated monolayers, consumption of NADH was only slightly increased and was not affected by inhibition of the MEK/ERK pathway.
The next experiment of this study was to test whether NAD(P)H oxidase is involved in the proliferative response to hypoxia. We found that in cultures treated with p22phox antisense oligonucleotides, the proliferative response to hypoxia was diminished, in contrast to the nonsense-treated controls (Figure 8). Treatment of endothelial monolayers with either nonsense or antisense p22phox oligonucleotides had no effect on the small proliferation rate found in normoxic controls (Figure 8).
In order to test for respiratory chain inhibition other than lack of oxygen, we incubated cells for 1 hour in presence of rotenone (10 μmol/L) instead of hypoxia. We found that this treatment also increased ROS formation and ERK2 activation (not shown) and induced endothelial cell proliferation to an extent comparable to that of hypoxia: cell number was increased to 183±10% of normoxic control (n=16; P<0.05 versus control). Treatment with p22phox antisense reduced cell proliferation to 112±6% of control (NS versus control), and incubation with the nonsense oligonucleotides had no effect of normoxic control (182±9%; P<0.05 versus control; NS versus hypoxia alone).
The aim of the present study was to analyze the role of redox signaling for the autonomous proliferative response of endothelial cells to hypoxia. We found that the proliferative response depends on the MEK/ERK pathway and that redox signaling steps are located both upstream and downstream of this pathway. Located in upstream position is ROS generation by mitochondria; located downstream is NAD(P)H oxidase.
The proliferative response to hypoxia was found sensitive to DPI and MQ. DPI can inhibit redox signaling via flavoproteins, as are present in the NAD(P)H oxidase complex and complex I of the mitochondrial respiratory chain. MQ is a highly selective scavenger of intramitochondrial generated ROS.5,11 In hypoxic cells, electrons can leak from the electron transport chain.12 A common denominator of DPI and MQ is the ability to reduce mitochondrial generation of ROS. The data of this study further show that the proliferative response is dependent on the MEK/ERK pathway, because it is sensitive to the MEK inhibitor, PD, and that ERK2 activation occurs downstream of a DPI and MQ sensitive step. Because ERK2 activation was not sensitive to p22phox antisense treatment, it does not depend on NAD(P)H oxidase, another target of DPI. Together, the results indicate that hypoxic generation of ROS in mitochondria causes activation of the MEK/ERK pathway and that both these events are required for the largest part of the proliferative response.
In the hypoxic cells, the largest part of ROS is generated downstream of the MEK/ERK pathway. This is shown by the finding that ROS generation can be effectively suppressed by the MEK inhibitor, PD. Presence of PD was also found to block the hypoxic activation of NAD(P)H oxidase, determined by the NADH consumption assay. NAD(P)H oxidase was confirmed as source of the largest part of hypoxically generated ROS by use of p22phox antisense oligonucleotides. Treatment with the latter had no effect on ERK2 activation, consistent with the aforementioned evidence that NAD(P)H oxidase is activated downstream of the MEK/ERK pathway. Downregulation of NAD(P)H oxidase by antisense oligonucleotides was sufficient to block the largest part (>80%) of cell proliferation in response to hypoxia. This shows that NAD(P)H oxidase represents an essential element in signaling of the autonomous proliferative response. In the described sequence of signaling events, NAD(P)H oxidase activation seems to play a dual role, as effector and amplifier of the original hypoxia signal at the mitochondrial level. Recent studies have revealed an essential role of NAD(P)H oxidase-dependent redox signaling also for other stimuli of endothelial cell proliferation, eg, growth factors3,4 or oxidized LDL.13
The finding that the mitochondrial ROS scavenger, MQ, suppresses ROS generation almost entirely is not in objection to the proposition that mitochondria are not the major source of ROS in the hypoxic cells. This can be understood when considering the fact that mitochondrial ROS production represents the initial step of a signaling cascade which leads to activation of a larger producer of ROS in downstream position, ie, NAD(P)H oxidase.
The identified signaling cascade starts at the mitochondrial level. Mitochondrial ROS seem to represent the initial event in “hypoxia sensing” of this proliferative response. The data suggest that these mitochondrial ROS are directly responsible for a small (<20%) part of the proliferative response because this part is insensitive to treatment with PD or antisense nucleotides. Others have also argued that hypoxic generation of ROS in mitochondria is an initial step in hypoxia-sensing mechanisms,14,15 eg, for NF-κB activation and subsequent gene expression in endothelial cells.15 It was also shown for other cells that mitochondria are essential for hypoxia-induced stimulation of proliferation.16 A general role of NAD(P)H oxidase as the primary trigger of oxygen sensing has become unlikely, because hypoxia sensing is preserved in transgenic mice lacking the essential gp91phox subunit of the NAD(P)H oxidase complex.17
ROS may be generated in hypoxic mitochondria at the flavoprotein containing complex I as well as at the ubisemiquinone site of complex III.14,16,18 We found that rotenone, an inhibitor acting behind complex I, activates ROS production and ERK2 and stimulates proliferation equally to hypoxia. These experiments indicate, independent of the experiments described, that ROS generated by respiratory chain inhibition trigger the proliferative response; complex I inhibition is a sufficient cause but ROS generated at other mitochondrial sites may also contribute.
The mechanism by which ROS lead to activation of the MEK/ERK pathway is not yet known. Our study is in line with reports that exogenously applied ROS lead to activation of the MEK/ERK pathway in vascular smooth muscle cells.19 PKC isoforms, well known as signaling elements upstream of the MEK/ERK pathway, are potential targets of ROS. Knapp and Klann20 recently showed that PKC isoforms, which possess cysteine-rich, zinc finger motifs, are activated in a redox-dependent way. The mechanism by which activation of the MEK/ERK pathway causes the activation of NAD(P)H oxidase is not identified either. In neutrophils, it was demonstrated that p47phox, a regulatory subunit of the NAD(P)H oxidase complex, is phosphorylated in a ERK dependent way.21 This may apply to endothelial cells and may also be responsible for the activation of NAD(P)H oxidase here reported.
In summary, the autonomous proliferative response of endothelial cells is initiated at the mitochondrial level with production of ROS or at a step before that. Subsequent to that mitochondrial event are activation of MEK/ERK and NAD(P)H oxidase. Both are also signaling elements for other stimuli of endothelial proliferation.
This work was supported by the Deutsche Forschungsgemeinschaft, grants A3 and A4 of Sonderforschungsbereich 547. We gratefully acknowledge the excellent technical assistance of Daniela Schreiber and Anna Stapler.
- Received August 1, 2002.
- Revision received March 28, 2003.
- Accepted March 28, 2003.
Oosthuyse B, Moons L, Storkebaum E, Beck H, Nuyens D, Brusselmans K, Van Dorpe J, Hellings P, Gorselink M, Heymans S, Theilmeier G, Dewerchin M, Laudenbach V, Vermylen P, Raat H, Acker T, Vleminckx V, Van Den Bosch L, Cashman N, Fujisawa H, Drost MR, Sciot R, Bruyninckx F, Hicklin DJ, Ince C, Gressens P, Lupu F, Plate KH, Robberecht W, Herbert JM, Collen D, Carmeliet P. Deletion of the hypoxia-response element in the vascular endothelial growth factor promoter causes motor neuron degeneration. Nat Genet. 2001; 28: 131–138.
Colavitti R, Pani G, Bedogni B, Anzevino R, Borrello S, Waltenberger J, Galeotti T. Reactive oxygen species as downstream mediators of angiogenic signaling by vascular endothelial growth factor receptor-2/KDR. J Biol Chem. 2002; 277: 3101–3108.
Kelso GF, Porteous CM, Coulter CV, Hughes G, Porteous WK, Ledgerwood EC, Smith RA, Murphy MP. Selective targeting of a redox-active ubiquinone to mitochondria within cells: antioxidant and antiapoptotic properties. J Biol Chem. 2001; 276: 4588–4596.
Spahr R, Piper HM. Microcarrier cultures of endothelial cells. In: Piper HM, ed. Cell Culture Techniques in Heart and Vessel Research. Heidelberg, Germany: Springer-Verlag; 1990: 220–229.
Patterson C, Ruef J, Madamanchi NR, Barry-Lane P, Hu Z, Horaist C, Ballinger CA, Brasier AR, Bode C, Runge MS. Stimulation of a vascular smooth muscle cell NAD(P)H oxidase by thrombin: evidence that p47phox may participate in forming this oxidase in vitro and in vivo. J Biol Chem. 1999; 274: 19814–19822.
Chandel NS, Schumacker PT. Cellular oxygen sensing by mitochondria: old questions, new insight. J Appl Physiol. 2000; 88: 1880–1889.
Heinloth A, Heermeier K, Raff U, Wanner C, Galle J. Stimulation of NADPH oxidase by oxidized low-density lipoprotein induces proliferation of human vascular endothelial cells. J Am Soc Nephrol. 2000; 11: 1819–1825.
Chandel NS, McClintock DS, Feliciano CE, Wood TM, Melendez JA, Rodriguez AM, Schumacker PT. Reactive oxygen species generated at mitochondrial complex III stabilize hypoxia-inducible factor-1α during hypoxia: a mechanism of O2 sensing. J Biol Chem. 2000; 275: 25130–25138.
Pearlstein DP, Ali MH, Mungai PT, Hynes KL, Gewertz BL, Schumacker PT. Role of mitochondrial oxidant generation in endothelial cell responses to hypoxia. Arterioscler Thromb Vasc Biol. 2002; 22: 566–573.
Chandel NS, Maltepe E, Goldwasser E, Mathieu CE, Simon MC, Schumacker PT. Mitochondrial reactive oxygen species trigger hypoxia-induced transcription. Proc Natl Acad Sci U S A. 1998; 95: 11715–11720.
Archer SL, Reeve HL, Michelakis E, Puttagunta L, Waite R, Nelson DP, Dinauer MC, Weir EK. O2 sensing is preserved in mice lacking the gp91phox subunit of NADPH oxidase. Proc Natl Acad Sci U S A. 1999; 96: 7944–7949.
Baas AS, Berk BC. Differential activation of mitogen-activated protein kinases by H2O2 and O2− in vascular smooth muscle cells. Circ Res. 1995; 77: 29–36.
Knapp LT, Klann E. Superoxide-induced stimulation of protein kinase C via thiol modification and modulation of zinc content. J Biol Chem. 2000; 275: 24136–24145.
Dewas C, Fay M, Gougerot-Pocidalo MA, El-Benna J. The mitogen-activated protein kinase extracellular signal-regulated kinase 1/2 pathway is involved in formyl-methionyl-leucyl-phenylalanine-induced p47phox phosphorylation in human neutrophils. J Immunol. 2000; 165: 5238–5244.