NADPH Oxidase Activation Increases the Sensitivity of Intracellular Ca2+ Stores to Inositol 1,4,5-Trisphosphate in Human Endothelial Cells*

Many stimuli that activate the vascular NADPH oxidase generate reactive oxygen species and increase intracellular Ca2+, but whether NADPH oxidase activation directly affects Ca2+ signaling is unknown. NADPH stimulated the production of superoxide anion and H2O2 in human aortic endothelial cells that was inhibited by the NADPH oxidase inhibitor diphenyleneiodonium and was significantly attenuated in cells transiently expressing a dominant negative allele of the small GTP-binding protein Rac1, which is required for oxidase activity. In permeabilized Mag-indo 1-loaded cells, NADPH and H2O2 each decreased the threshold concentration of inositol 1,4,5-trisphosphate (InsP3) required to release intracellularly stored Ca2+ and shifted the InsP3-Ca2+ release dose-response curve to the left. Concentrations of H2O2 as low as 3 μm increased the sensitivity of intracellular Ca2+ stores to InsP3 and decreased the InsP3 EC50 from 423.2 ± 54.9 to 276.9 ± 14.4 nm. The effect of NADPH on InsP3-stimulated Ca2+ release was blocked by catalase and by diphenyleneiodonium and was not observed in cells lacking functional Rac1 protein. Thus, NADPH oxidase-derived H2O2 increases the sensitivity of intracellular Ca2+ stores to InsP3 in human endothelial cells. Since Ca2+-dependent signaling pathways are critical to normal endothelial function, this effect may be of great importance in endothelial signal transduction.

The endothelial cell membrane contains an NADPH oxidaselike H 2 O 2 -generating enzyme (1,2) that is stimulated by posthypoxic reoxygenation (3), cyclic stretch (4,5), and low density lipoprotein (6). The signal transduction pathways stimulated following activation of the NADPH oxidase in the vascular endothelium have not been completely characterized. Many stimuli that activate the oxidase also increase endothelial cytosolic calcium concentration ([Ca 2ϩ ] i ) (7-10), but whether activation of the NADPH oxidase affects endothelial Ca 2ϩ signaling is unknown.
We recently showed that H 2 O 2 stimulates [Ca 2ϩ ] i oscillations in human aortic endothelial cells (HAEC) 1 (11). In con-trast to other agonists that stimulate [Ca 2ϩ ] i oscillations in endothelial cells like bradykinin (12), histamine (13), and adenosine trisphosphate (12), H 2 O 2 does not increase levels of inositol 1,4,5-trisphosphate (InsP 3 ) at the concentrations that produce [Ca 2ϩ ] i oscillations (14). Since the H 2 O 2 -generating enzyme xanthine oxidase has been shown to decrease luminal Ca 2ϩ content in vascular endothelial cells (15) and since redox sensitivity of the InsP 3 receptor has been demonstrated in other cell types (16,17), we hypothesized that activation of the NADPH oxidase affects endothelial Ca 2ϩ signaling by increasing the sensitivity of intracellular Ca 2ϩ stores to InsP 3 -stimulated Ca 2ϩ release. To test this hypothesis, the effect of NADPH oxidase stimulation on intracellular Ca 2ϩ stores was examined in permeabilized HAEC using the low affinity (micromolar range) Ca 2ϩ -sensitive fluorescent indicator Mag-indo 1. To characterize the specific role of NADPH oxidase stimulation, diphenyleneiodonium was used to pharmacologically inhibit the oxidase, and studies were performed using HAEC transiently expressing Rac1 N17 (18), a dominant negative allele of the small GTP-binding protein Rac1 that is required for oxidase activity.

MATERIALS AND METHODS
Culture of HAEC-HAEC were obtained as proliferating quaternary cultures (Clonetics, San Diego, CA) and were grown to confluence to passage 5-9 in endothelial cell growth medium supplemented with 2% fetal bovine serum, 10 g/liter human-recombinant epidermal growth factor, 1 mg/liter hydrocortisone, 50 g/ml gentamicin, 50 ng/ml amphotericin-B, and 12 g/ml bovine brain extract (Clonetics) in a 37°C humidified atmosphere of 95% air, 5% CO 2 . For Ca 2ϩ measurements, HAEC were plated at an approximate concentration of 1 ϫ 10 5 /ml on 25-mm diameter circular glass coverslips (VWR Scientific, Media, PA), which were precoated with 2% gelatin solution (Sigma) and washed three times with phosphate-buffered saline (Quality Biological, Inc., Gaithersburg, MD) before cell seeding. Cells were used for experiments after reaching 70% confluence after incubation for 1-2 days at 37°C in a humidified atmosphere of 95% air, 5% CO 2 . production by HAEC was measured fluorometrically using the Amplex TM Red Hydrogen Peroxide Assay Kit (Molecular Probes, Inc., Eugene, OR). Amplex TM Red is a fluorogenic substrate with very low background fluorescence, which reacts with H 2 O 2 with a 1:1 stoichiometry to produce highly fluorescent resorufin (19). Measurements of H 2 O 2 production after the addition of NADPH were performed using suspensions (ϳ10 5 / ml) of permeabilized HAEC in Corning 96-well microplates. Fluorescence intensity was measured in a Cytofluor 2300 System (Millipore Corp., Bedford, MA) at an excitation wavelength of 530 Ϯ 25 nm and an emission wavelength of 590 Ϯ 35 nm at room temperature. After subtracting background fluorescence, cumulative H 2 O 2 concentrations (M/ 10 5 cells) were calculated using a resorufin-H 2 O 2 standard calibration curve generated from cell-free experiments using H 2 O 2 and Amplex TM Red. In some experiments, HAEC were preincubated for 30 min prior to NADPH addition with either DPI (10 M), catalase (1000 units/ml), or SOD (200 units/ml). These inhibitors were also present during fluorescence measurements.

Determination of Superoxide Generation by HAEC-Superoxide
Measurement of Ca 2ϩ Release from Intracellular Ca 2ϩ Stores-HAEC monolayers on glass coverslips were incubated in a HEPES (Sigma)buffered saline containing 10 M of the ester derivative (acetoxymethyl ester form) of Mag-indo 1 (Molecular Probes) at room temperature for 45 min. The HEPES-buffered saline contained 137 mM NaCl, 4.9 mM KCl, 1.5 mM CaCl 2 , 1.2 mM MgSO 4 , 1.2 mM NaH 2 PO 4 , 15 mM D-glucose, 20 mM HEPES (pH adjusted to 7.40 at room temperature with NaOH). All salt solutions were made with pure water (free Ca 2ϩ Ͻ 0.05 ppb; PICOpure Water Systems, Research Triangle Park, NC). After Magindo 1 loading, coverslips were gently washed with indicator-free HEPES-buffered saline and maintained for an additional 10 min to allow for de-esterification of the indicator. Mag-indo 1-loaded HAEC were exposed for 2-3 min to Mg 2ϩ /ATP-free ICM. HAEC were then permeabilized by exposure to Mg 2ϩ /ATP-free ICM containing 30 g/ml (w/v) saponin (Sigma) for 1.5-2 min at room temperature. Permeabilization was monitored by observing the release of Mag-indo 1 cytosolic fluorescence. After permeabilization, HAEC were washed with Mg 2ϩ / ATP-free ICM and then exposed to complete ICM (containing 1 mM ATP and 1.4 mM MgCl 2 ; free Mg 2ϩ concentration ϭ 0.1 mM) for 10 min to allow for filling of intracellular Ca 2ϩ stores. Permeabilized HAEC were then superfused for at least 10 min with Ca 2ϩ -releasing medium containing 125 mM KCl, 19 mM NaCl, 10 mM HEPES, 1.4 mM MgCl 2 , 150 nM CaCl 2 (calculated concentration confirmed by fluorescence of indo-1 free acid). The pH of the Ca 2ϩ -releasing medium was adjusted to 7.20 with KOH.
Mag-indo 1 fluorescence was recorded at room temperature from HAEC on coverslips in a perfusion chamber mounted on the stage of a modified Nikon Diaphot inverted epifluorescence microscope. Mag-indo 1 fluorescence was excited at 350 Ϯ 50 nm using a xenon short arc lamp (UXL-75 XE, Ushio Inc., Tokyo, Japan), and bandpass interference filters (Omega Optical, Brattleboro, VT) selected wavelength bands of emitted fluorescence at 405 Ϯ 10 and 485 Ϯ 10 nm, corresponding to the Ca 2ϩ -bound and Ca 2ϩ -free forms of the indicator, respectively. Emitted fluorescence from single permeabilized HAEC was collected and measured using a spectrofluorimeter (PTI, Deltascan). The photometer had a series of fixed pinhole diaphragms to regulate the area of the recording field. After permeabilization, autofluorescence from unloaded HAEC was generally Ͻ5% of Mag-indo 1-loaded HAEC and was subtracted automatically from Mag-indo 1 fluorescence recordings. Since low affinity Ca 2ϩ -sensitive indicators like Mag-indo 1 (20) report free Ca 2ϩ concentration in the micromolar range and accumulate in intracellular organelles, they may be effectively used to monitor the free Ca 2ϩ content of intracellular Ca 2ϩ stores but are relatively insensitive to changes in cytosolic [Ca 2ϩ ], which typically occur in the nanomolar range.
When intact HAEC monolayers were loaded with Mag-indo 1, the fluorescence of the indicator was distributed uniformly throughout the cytoplasm. After incubation in Mg 2ϩ /ATP-free ICM and saponin permeabilization, much of the intracellular fluorescence was lost, presumably due to permeabilization of the cell membrane and release of intracellular dye. The remaining fluorescence appeared to reside in the perinuclear region, presumably in membrane-bound compartments like the endoplasmic reticulum.
The fluorescence ratios at zero and saturating free calcium were determined from 10 M Mag-indo 1 free acid solutions in Mg 2ϩ /ATP-free ICM containing no added CaCl 2 and 1 mM EGTA or containing 10 mM CaCl 2 , respectively, and were used as the minimum and maximum ratios (R min and R max , respectively). Under these experimental conditions, R min ϭ 0.19 Ϯ 0.04, and R max ϭ 4.56 Ϯ 0.75 for Mag-indo 1 (n ϭ 6 for each). The approximate intracellular store [Ca 2ϩ ] was calculated as described previously (20,21), according to the following formula: where K d is the dissociation constant of Mag-indo 1, and S f2 and S b2 are the fluorescence intensities at 490 nm of the Ca 2ϩ -free and Ca 2ϩ -saturated indicator, respectively.
Sensitivity of Mag-indo 1 to Intracellular Store [Ca 2ϩ ]-The distribution and intensity of the Mag-indo 1 fluorescence signal was relatively stable in the absence of agonist stimulation. Over a 25-min period, fluorescence intensity decreased by approximately 15% (F 405 , 14.5 Ϯ 2.4%; F 485 , 15.9 Ϯ 3.2%), although the F 405 :F 485 fluorescence ratio changed by Ͻ1%. When permeabilized HAEC were exposed to Mg 2ϩ -and ATP-containing medium to allow for filling of intracellular Ca 2ϩ stores, the increase of the Mag-indo 1 ratio was dependent on the Ca 2ϩ concentration in the presence of Mg 2ϩ and ATP but was not affected by either Mg 2ϩ alone, Mg 2ϩ and ATP, or Ca 2ϩ alone (data not shown). These findings are consistent with the known characteristics of Mag-indo 1, whose sensitivity to Mg 2ϩ is 100-fold less than to Ca 2ϩ (K d Mag-indo 1 ratio was sensitive to application of InsP 3 , and the rate of the increase in the Mag-indo 1 ratio during filling of intracellular Ca 2ϩ stores was inhibited by the endoplasmic reticulum Ca 2ϩ -ATPase inhibitor thapsigargin (data not shown). HAEC Transiently Expressing the Dominant Negative Allele of Rac1-An adenovirus encoding the Myc epitope-tagged-dominant negative Rac1 cDNA containing a substitution at position 17 (Rac1 N17 ) was used as described previously (18).
Expression of the Rac1 N17 mutant was confirmed by protein immunoblotting with an antibody to the Myc epitope (9E10; Santa Cruz Biotechnology, Inc., Santa Cruz, CA).
Statistical Analysis-Data are reported as mean Ϯ S.E. Statistical comparisons were made using Student's t test for the paired and the unpaired groups. An analysis of variance was used when multiple comparisons were performed. A difference was considered significant at p Ͻ 0.05.

NADPH-stimulated O 2 . Generation in HAEC-Neither
HAEC suspensions alone nor NADPH alone generated detectable chemiluminescence in the presence of 25 M lucigenin.

Effect of NADPH on H 2 O 2 Production by HAEC-
The effect of NADPH on H 2 O 2 production by HAEC was examined by measuring resorufin fluorescence intensity from HAEC suspensions after NADPH addition (1-300 M). NADPH stimulated production of H 2 O 2 in a time-and concentration-dependent manner ( Fig. 2A), with a threshold of ϳ30 M. H 2 O 2 production reached a plateau after ϳ12.5 min. The apparent maximum rate of H 2 O 2 production was 0.46, 0.86, and 1.11 M/min/10 5 cells after the addition of 30, 100, and 300 M, respectively. No measurable spontaneous H 2 O 2 production was observed in the absence of NADPH. H 2 O 2 production was significantly inhibited by pretreatment with either catalase or with the NADPH oxidase inhibitor DPI at each concentration of NADPH examined (Fig. 2B). In contrast, SOD did not significantly affect H 2 O 2 production at any NADPH concentration (p ϭ NS, n ϭ 3).
Effect of H 2 O 2 on InsP 3 -sensitive Intracellular Ca 2ϩ Stores-In permeabilized HAEC in which Ca 2ϩ stores are filled by perfusion with Ca 2ϩ -, Mg 2ϩ -, and ATP-containing ICM, 100 M H 2 O 2 did not affect intracellular Ca 2ϩ stores during a 15-min exposure (⌬ ratio ϭ 0.01 Ϯ 0.00, n ϭ 3, p ϭ NS versus 0.01 Ϯ 0.01, n ϭ 8 for time control). In the same HAEC, however (Fig. 3A), the addition of InsP 3 stimulated a rapid decrease in the content of intracellular Ca 2ϩ stores. A higher concentration of H 2 O 2 (1 mM for 10 min) in the absence of InsP 3 also did not affect intracellular Ca 2ϩ stores in Ca 2ϩ -replete, permeabilized HAEC (⌬ ratio ϭ 0.01 Ϯ 0.01, n ϭ 3, p ϭ NS versus time control).
As shown in Fig. 3B, when HAEC were exposed to a submaximal concentration of InsP 3 (300 nM), the Mag-indo 1 ratio decreased from 1.06 Ϯ 0.05 to 0.92 Ϯ 0.05 (n ϭ 4, p Ͻ 0.001) and was then maintained at this level for more than 10 min of observation. The decrease in the content of intracellular Ca 2ϩ stores stimulated by InsP 3 was reversibly inhibited by heparin (⌬ ratio ϭ 0.02 Ϯ 0.01 with heparin versus 0.01 Ϯ 0.01 for time control, p ϭ NS, n ϭ 8 for each). While 100 M H 2 O 2 did not affect intracellular Ca 2ϩ stores in permeabilized HAEC in the absence of InsP 3 , during an established response to a submaximal InsP 3 concentration of 300 nM (Fig. 3C) tional InsP 3 -insensitive intracellular Ca 2ϩ stores in HAEC (Fig. 4A).
H 2 O 2 (3-300 M) shifted the InsP 3 -stimulated intracellular Ca 2ϩ release dose-response curve to the left, decreased the threshold concentration to Յ10 nM InsP 3 , and decreased the EC 50 in response to InsP 3 (Fig. 4, B and C). In the presence of 100 M H 2 O 2 , a significant decrease in the Mag-indo 1 ratio was observed during exposure to 10 nM InsP 3 (⌬ ratio ϭ 0.10 Ϯ 0.01, n ϭ 4, p Ͻ 0.001 versus 10 nM InsP 3 alone). When permeabilized HAEC were exposed to 100 and 300 nM InsP 3 in the presence of 100 M H 2 O 2 , the decrease in the content of intracellular Ca 2ϩ stores was greater than that observed in the absence of Effect of NADPH on InsP 3 -sensitive Intracellular Ca 2ϩ Stores-To determine the effect of NADPH oxidase stimulation on InsP 3 -sensitive intracellular Ca 2ϩ stores, Mag-indo 1-loaded HAEC were stimulated by the addition of 100 M NADPH. This concentration of NADPH has previously been used to stimulate the oxidase in endothelial cells (2,3,23). The addition of 100 M NADPH to permeabilized HAEC rapidly and irreversibly decreased the Mag-indo 1 ratio by 0.08 Ϯ 0.01 (n ϭ 4) over 10 min, due primarily to an increase in the fluorescence intensity of Mag-indo 1 at the 485-nm wavelength. This appeared to be an effect on the indicator itself, rather than an effect on intracellular Ca 2ϩ stores, since in a cell-free system, the ratio of Magindo 1 free acid (10 M in 100 M free Ca 2ϩ ) was similarly decreased (0.08 Ϯ 0.02, n ϭ 2). NADPH shifted the InsP 3 dose-response relationship to the left, reduced the threshold concentration of InsP 3 , and decreased the EC 50 in response to  on intracellular Ca 2ϩ release were similarly enhanced by 100 M NADPH (⌬ ratio ϭ 0.22 Ϯ 0.04 for 100 nM InsP 3 , p Ͻ 0.01 versus 100 nM InsP 3 alone and ⌬ ratio ϭ 0.31 Ϯ 0.05 for 300 nM InsP 3 , p Ͻ 0.05 versus 300 nM InsP 3 alone, n ϭ 4 for each concentration). The decrease in the content of intracellular Ca 2ϩ stores stimulated by 1 and 3 M InsP 3 was not affected by the presence of NADPH.
As shown in Fig. 6, NADPH produced a dose-dependent decrease in the EC 50 of InsP 3 , which became significant at a concentration of 100 M. At this concentration, NADPH decreased the EC 50 in response to InsP 3 by more than 50% from 423.2 Ϯ 54.9 to 170.6 Ϯ 23. Importance of Rac1 in NADPH-induced H 2 O 2 Production and Increased Sensitivity to InsP 3 -To further characterize the role of NADPH oxidase activation in stimulating the production of H 2 O 2 and increasing the sensitivity to InsP 3 -stimulated intracellular Ca 2ϩ release, studies were performed in HAEC transiently expressing Rac1 N17 , a dominant negative allele of Rac1 (18). When transfected cells were stimulated with NADPH, H 2 O 2 production was significantly reduced in comparison with vector controls (Fig. 7A)

DISCUSSION
This study shows that NADPH oxidase activation stimulates the production of H 2 O 2 and increases the sensitivity of intracellular Ca 2ϩ stores to InsP 3 in human endothelial cells. A vascular NADH/NADPH oxidase is a major source of reactive oxygen species in the vasculature (24), and previous studies have demonstrated the expression of one or more components of the NADPH oxidase in endothelial cells (1,2). Although it has been suggested that reactive oxygen species produced upon activation of the vascular oxidase may play a role in endothelial dysfunction (25)(26)(27), the signal transduction pathways stimulated following activation of the endothelial NADPH oxidase are largely unknown. Many stimuli that activate the oxidase increase endothelial [Ca 2ϩ ] i , but this is the first study to show a direct effect of NADPH oxidase activation on endothelial Ca 2ϩ signaling.
In the present study, stimulation of the endothelial NADPH oxidase by NADPH (2,3,23)  Redox sensitivity of the InsP 3 receptor has been reported previously in other cell types (16,17). The present study suggests that small changes in the intracellular concentration of H 2 O 2 are likely to have significant effects on InsP 3 -sensitive intracellular Ca 2ϩ stores, given the effects on the EC 50 of InsP 3 of relatively small changes in the concentration of either H 2 O 2 or NADPH. Such a narrow range between subthreshold and maximal responses suggests that the intracellular concentration of H 2 O 2 is tightly regulated in vivo. Since the concentrations of H 2 O 2 that are generated by vascular cells as a result of NADPH oxidase activation are not accurately known, it is difficult to determine how the concentrations of H 2 O 2 observed following NADPH addition to permeabilized cell suspensions in the present study compare with that which occurs in vivo. Importantly, however, even concentrations of H 2 O 2 in the low micromolar range increased the sensitivity of intracellular Ca 2ϩ stores to InsP 3 in HAEC.
Our data indicate that NADPH oxidase-derived H 2 O 2 decreases the threshold concentration for intracellular Ca 2ϩ release in endothelial cells to Ͻ10 nM InsP 3 . Since basal intracellular levels of InsP 3 are in the 50 nM range (29,30), stimuli that activate the oxidase may affect endothelial Ca 2ϩ signaling, even without altering levels of InsP 3 . It may be of interest in this regard that we recently showed that H 2 O 2 , at concentrations that do not increase levels of InsP 3 (14), stimulate Ca 2ϩ oscillations in human endothelial cells (11). Since intracellular Ca 2ϩ oscillations can occur even at constant levels of InsP 3 (31), the increased sensitivity of intracellular Ca 2ϩ stores to InsP 3 resulting from NADPH oxidase activation may facilitate the generation of intracellular Ca 2ϩ oscillations or the regulation of their frequency. Such an effect may underlie recent observations in platelet-derived growth factor-stimulated vascular smooth muscle cells. In this cell type, platelet-derived growth factor induces intracellular Ca 2ϩ oscillations (32), generates O 2 . , and stimulates the activity of the proinflammatory transcription factor NF-B (33). Activation of NF-B by plateletderived growth factor is inhibited by both SOD and DPI, suggesting a link between NADPH oxidase activation and proinflammatory signaling pathways in the vasculature. Since the frequency of intracellular Ca 2ϩ oscillations regulates the activity of proinflammatory transcription factors like NF-B (34,35), the effect of NADPH oxidase activation on InsP 3mediated Ca 2ϩ release reported in this study may be of great importance in vascular signal transduction.