www.elsevier.com/locate/vph

Vascular Pharmacology 4

Rac1 inhibition protects against hypoxia/reoxygenation-induced lipid

peroxidation in human vascular endothelial cells

Sergio F. Martin a, Subroto Chatterjee b, Narasimham Parinandi c, B. Rita Alevriadou d,*

a Department of Biomedical Engineering, Johns Hopkins University School of Medicine, Baltimore, MD 21205, USAb Lipid Research Atherosclerosis Unit, Department of Pediatrics, Johns Hopkins University School of Medicine, Baltimore, MD 21205, USA

c Davis Heart and Lung Research Institute, Ohio State University College of Medicine, Columbus, OH 43210, USAd Davis Heart and Lung Research Institute and Biomedical Engineering Center, Ohio State University, 473 West 12th Avenue, Columbus, OH 43210, USA

Received 12 November 2004; received in revised form 10 February 2005; accepted 17 May 2005

Abstract

Both in vivo models of ischemia/reperfusion and in vitro models of hypoxia (H)/reoxygenation (R) have demonstrated the crucial role

of the Rac1-regulated NADPH oxidase in the production of injurious reactive oxygen species (ROS) by vascular endothelial cells (ECs).

Since membrane lipid peroxidation has been established as one of the mechanisms leading to cell death, we examined lipid peroxidation

in H/R-exposed cultured human umbilical vein ECs (HUVECs) and the role of Rac1 in this process. H (24 h at 1% O2)/R (5 min)

caused an increase in intracellular ROS production compared to a normoxic control, as measured by dichlorofluorescin fluorescence.

Nutrient deprivation (ND; 24 h), a component of H, was sufficient to induce a similar increase in ROS under normoxia. Either H(24 h)/

R (2 h) or ND (24 h) induced increases in lipid peroxidation of similar magnitude as measured by flow cytometry of diphenyl-1-

pyrenylphosphine-loaded HUVECs and Western blotting analysis of 4-hydroxy-2-nonenal-modified proteins in cell lysates. In cells

infected with a control adenovirus, H (24 h)/R (2 h) and ND (24 h) resulted in increases in NADPH-dependent superoxide production by

5- and 9-fold, respectively, as measured by lucigenin chemiluminescence. Infection of HUVECs with an adenovirus that encodes the

dominant-negative allele of Rac1 (Rac1N17) abolished these increases. Rac1N17 expression also suppressed the H/R- and ND-induced

increases in lipid peroxidation. In conclusion, ROS generated via the Rac1-dependent pathway are major contributors to the H/R-induced

lipid peroxidation in HUVECs, and ND is able to induce Rac1-dependent ROS production and lipid peroxidation of at least the same

magnitude as H/R.

D 2005 Elsevier Inc. All rights reserved.

Keywords: Endothelial cells; Hypoxia/reoxygenation; Nutrient deprivation; Reactive oxygen species; Rac1-regulated NADPH oxidase; Lipid peroxidation

1. Introduction

Reperfusion of ischemic tissue results in the generation

of reactive oxygen species (ROS) that contribute to tissue

injury (Granger and Korthuis, 1995; Grisham et al., 1998;

Zweier et al., 1988). Earlier studies suggested that xanthine

oxidase and the mitochondrial electron transport chain are

important sources of ROS produced during reperfusion

(Acosta and Li, 1979; Ratych et al., 1987; Russell et al.,

1994). However, ROS production by the isolated and

1537-1891/$ - see front matter D 2005 Elsevier Inc. All rights reserved.

doi:10.1016/j.vph.2005.05.002

* Corresponding author. Tel.: +1 614 292 5160; fax: +1 614 292 8778.

E-mail address: alevriadou-1@medctr.osu.edu (B.R. Alevriadou).

continuously ventilated ischemic mouse lungs was shown

to originate from the membrane-associated NADPH oxidase

complex of vascular endothelial cells (ECs) (Al-Mehdi et

al., 1998). ROS production by the membrane-bound

NADPH oxidase in both phagocytic and nonphagocytic

cells is regulated by Rac1, a member of the Rho family of

small GTPases (Abo et al., 1991). The crucial role of the

Rac1-regulated NADPH oxidase in generating the injurious

ROS was demonstrated in an in vivo model of mouse

hepatic ischemia/reperfusion injury, in which the recombi-

nant adenoviral expression of a dominant negative Rac1

(Rac1N17) completely suppressed the ischemia/reperfusion-

induced ROS production and lipid peroxidation and

3 (2005) 148 – 156

S.F. Martin et al. / Vascular Pharmacology 43 (2005) 148–156 149

significantly reduced apoptosis and necrosis in the liver

(Ozaki et al., 2000).

In in vitro models of ischemia/reperfusion, injury due to

exposure of cultured ECs to hypoxia (H)/reoxygenation (R)

was mediated by endogenous ROS production and was a

function of severity of H and duration of both H and R

(McLeod and Sevanian, 1997; Michiels et al., 1992; Terada,

1996). As in the in vivo models, Rac1N17 expression in

cultured human umbilical vein ECs (HUVECs) inhibited the

H/R-induced burst in ROS generation and significantly

reduced cell death, as compared to control cells (Kim et al.,

1998). Peroxidation of the phospholipids in the plasma

membrane mediated by ROS has been established as one of

the mechanisms of cell death (Girotti, 1998). Phospholipid

peroxidation perturbs the integrity of the lipid bilayer of cell

membranes leading to increased plasma membrane perme-

ability and apoptotic cell death (Kriska et al., 2002; Pak et

al., 2002). Exposure of cultured ECs to H/R was shown to

increase both the extent of lipid peroxidation and the number

of apoptotic cells (Li et al., 1998; McLeod and Alayash,

1999; Yang et al., 2001). However, the effect of Rac1 that

regulates the NADPH oxidase-mediated generation of ROS

on the membrane lipid peroxidation in ECs exposed to H/R

has not been reported.

Therefore, in the present study, we investigated the

effect of H/R exposure on lipid peroxidation in cultured

HUVECs and the role of Rac1 in this process. Cells were

exposed to severe prolonged H (24 h at 1% O2) followed

by R up to 2 h. During exposure to H, HUVECs were

maintained in serum-free media, so that they were also

subjected to nutrient deprivation (ND), to create H more

physiologically relevant to ischemia. Since a recent report

has revealed that ND alone is sufficient to induce

production of monocyte chemoattractant protein (MCP)-

1 in ECs that requires ROS generated by the Rac1-

dependent pathway (Lopes et al., 2002), we also

investigated the effect of ND (24 h) on lipid peroxidation

in cultured HUVECs infected either with control adeno-

virus or with adenovirus encoding Rac1N17. The extent

of lipid peroxidation in HUVECs was determined by

flow-cytometric analysis of the intracellular formation of

lipid hydroperoxides using diphenyl-1-pyrenylphosphine

(DPPP), and by Western blotting analysis of the for-

mation of 4-hydroxy-2-nonenal (HNE)-modified proteins

in cell lysates using anti-HNE polyclonal antibodies.

Intracellular generation of ROS was determined by 2V,7V-dichlorodihydrofluorescein diacetate (DCFH-DA)-coupled

flow cytometry. Cellular NADPH oxidase activity was

determined by lucigenin chemiluminescence assay. Our

results showed that H (24 h)/R (5 min–2 h) induces ROS

production and lipid peroxidation mediated by the Rac1-

regulated NADPH oxidase in HUVECs. Our results

further revealed that ND (24 h) is sufficient to induce

lipid peroxidation of similar magnitude as compared to H

(24 h)/R (2 h), which is partially due to Rac1-mediated

ROS production.

2. Materials and methods

2.1. Reagents

tert-Butyl hydroperoxide (tBOOH), cumene hydroper-

oxide (CumOOH), hydrogen peroxide (H2O2), lucigenin,

NADPH, and recombinant (Escherichia coli) human tumor

necrosis factor (TNF)-a were purchased from Sigma

Chemical Company (St. Louis, MO). DCFH-DA and DPPP

were purchased from Molecular Probes (Eugene, OR).

Stock reagents were prepared in medium M199, except

for DPPP that was dissolved in dimethylsulfoxide (DMSO)

at a concentration of 5 mM. Rabbit anti-HNE antiserum was

from Alpha Diagnostic International, Inc. (San Antonio,

TX). Its ability to react with HNE-protein adducts was

demonstrated in immunoblots of homogenates of rat liver

hepatocytes that had been exposed to oxidative stress with

tBOOH (Uchida et al., 1993). Specifically, proteins from

untreated and treated with tBOOH hepatocytes were

subjected to immunoblot analysis in both the absence and

presence of an HNE-peptide competitor. In the presence of

the competitor, the patterns of the unoxidized and oxidized

samples were almost identical and indistinguishable from

the pattern obtained for the control sample in the absence of

competitor, thus proving the specificity of the antibody

reaction.

2.2. EC culture and infection with adenovirus

Cryopreserved primary HUVECs were purchased from

Clonetics (San Diego, CA) and grown in EGM-2 complete

growth medium containing 2% fetal bovine serum (FBS)

and growth supplements (Clonetics). On confluence, cells

were passaged following trypsinization with a 0.05% trypsin

solution (Clonetics). ECs of passages 2 to 8 were used in all

experiments. Both the replication-deficient adenovirus

AdRac1N17, which encodes the myc epitope-tagged dom-

inant-negative allele of Rac1, and the control virus AdhGal,which encodes the inert bacterial LacZ gene, were described

previously (Sulciner et al., 1996; Sundaresan et al., 1995).

Viruses were amplified in HEK-293 cells, purified on

double cesium gradients and plaque-titered (Guzman et

al., 1994). Purified AdRac1N17 and AdhGal had titers of

2�1010 and 5�1010 plaque-forming units (pfu)/ml,

respectively. Some ECs were infected for 48 h before

exposure to H/R or ND (¨80% confluence) at a multiplicity

of infection (MOI) of 150 in EGM-2 complete growth

medium, as described previously (Ng et al., 2002).

2.3. EC treatment with oxidants, H/R or ND

As a positive control for lipid peroxidation, some ECs

were incubated with exogenous hydroperoxides with differ-

ent lipid solubilities, such as tBOOH (0.75 or 1.5 AM),

CumOOH (1 or 4 AM) or H2O2 (0.5 or 5 AM) in M199 for 1

h prior to measurement of ROS production and lipid

S.F. Martin et al. / Vascular Pharmacology 43 (2005) 148–156150

peroxidation. Since the TNF-a-induced signaling pathway

is known to involve intracellular ROS production (Desh-

pande et al., 2000), some ECs were preincubated with TNF-

a at 30 ng/ml in M199 for 1 h. For H/R treatment, ECs in

M199 were exposed to an anoxic gas mixture (95% N2/5%

CO2; Puritan Bennett, Linthicum Heights, MD) in a

humidified incubator (Billups-Rothenburg, Del Mar, CA)

for 2 or 24 h, as described previously (Ng et al., 2002).

Using a Clark-style O2 probe placed in the fluid in the tissue

culture container, the starting partial pressure of oxygen

(PO2) was measured at 150 mm Hg (20% O2) and it dropped

to 10 mm Hg (1% O2) within 15 min of H. Temperature

inside the incubator was maintained at 37 -C with a heating

pad. After H, R was initiated by returning the cells to 95%

air/5% CO2 for 5, 30 min or 2 h. EGM-2 complete medium

was added to the cells right at the onset of R. For ND

treatment, ECs were incubated in M199 in the regular tissue

culture incubator for 24 h. Cell viability was determined by

trypan blue exclusion following normoxia, H/R or ND, and

in each case it was �90%.

2.4. Measurement of ROS and hydroperoxides

To detect the intracellular ROS formation, ECs with and

without adenoviral infection were preloaded with DCFH-

DA (5 AM) in EGM-2 complete medium for 15 min at 37

-C prior to treatment with oxidants, H/R or ND. DCFH-

DA is trapped within cells in the form of DCFH. DCFH is

not fluorescent per se but it is oxidized by potent ROS to

yield measurable fluorescent dichlorofluorescin (DCF)

(Ischiropoulos et al., 1999). To determine the levels of

lipid hydroperoxides, ECs were instead preloaded with

DPPP (200 AM) in EGM-2 complete medium for 15 min

at 37 -C. DPPP reacts stoichiometrically with hydro-

peroxides to give DPPP oxide (DPPP_O), which, due to

its high fluorescence intensity, can be used for determi-

nation of lipid peroxidation in cell membranes (Okimoto et

al., 2000; Takahashi et al., 2001). In either case, after

washing with M199 to remove any extracellular probe,

HUVECs were exposed to stimuli (tBOOH, CumOOH,

H2O2 or TNF-a for 1 h), H (2 or 24 h)/R (5 or 30 min) or

ND (24 h), as described above, following which the cells

were trypsinized, neutralized with trypsin neutralizing

solution (Clonetics) and DCF fluorescence was measured

on a FACScan flow cytometer (Becton Dickinson, San

Jose, CA) with excitation and emission wavelengths set at

475 and 525 nm, respectively. DPPP fluorescence was

measured on a LSRScan flow cytometer (Becton Dick-

inson) with excitation and emission wavelengths set at 350

nm and 400 nm, respectively. For each analysis, 5000

events were recorded.

2.5. Detection of HNE-modified proteins

Following treatment with oxidants, H/R or ND, HUVECs

were scraped into ice-cold lysis buffer consisting of 145 mM

NaCl, 0.1 mM MgCl2, 15 mM HEPES, pH 7.4, 10 mM

EGTA, 1% Triton X-100, 1 mM Na3VO4, and protease

inhibitors such as chymostatin, aprotinin, leupeptin and

pepstatin, each at 10 Ag/ml. Cell extracts were homogen-

ized by sonication and centrifuged at 10,000 rpm for 10

min at 4 -C. Protein concentration in cell lysates was

determined using the bicinchoninic acid assay (Pierce,

Rockford, IL). Supernatant containing 30 Ag of protein was

mixed with 2� sample buffer, boiled for 3 min and applied

in duplicate on 4–20% Tris–glycine gels (Novex, San

Diego, CA) for protein separation by sodium dodecyl

sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) at

100 V constant voltage. One gel was stained with

Coomassie brilliant blue to confirm loading of equal

amount of proteins and the other was used for Western

blotting analysis. For the latter, proteins were electro-

phoretically transferred at 30 V of constant voltage at 4 -Cfor 12 h onto nitrocellulose membranes (Amersham Life

Science, Arlington Heights, IL). Blots were probed with

1:5000 diluted rabbit anti-HNE polyclonal antibody (Alpha

Diagnostic International) and 1:1000 diluted alkaline

phosphatase-conjugated goat anti-rabbit IgG (Calbiochem,

San Diego, CA) and were visualized using the Western-

Breeze\ chemiluminescent Western blot immunodetection

kit (Invitrogen, Carlsbad, CA).

2.6. Measurement of NADPH-dependent O2� production

Following treatment, NADPH-dependent O2� produc-

tion by intact HUVECs was measured using lucigenin-

enhanced chemiluminescence, as described previously (Li

and Shah, 2001). Lucigenin is a compound that emits light

upon interaction with O2� (Gyllenhammar, 1987). Briefly,

ECs were harvested in phosphate-buffered saline (PBS)

and cell pellets were suspended in a balanced salt solution

(130 mM NaCl, 5 mM KC1, 1 mM MgCl2, 1 mM

CaCl2I2H2O, 35 mM H3PO4, and 20 mM HEPES, pH 7.4).

Cell viability determined by trypan blue exclusion was

�90%. HUVEC suspensions were added to duplicate

wells of 96-well plates, followed by NADPH (100 AMfinal concentration) and dark-adapted lucigenin (250 AMfinal concentration). Increases in photon emission were

measured every 60 s for 16 min in a scintillation counter

(Top counter; Packard Instrument, Meriden, CT). A plateau

was achieved after ¨11 min. Light emission (arbitrary

units/min) was first averaged over duplicate wells and then

over the last 5 min in the counter and was normalized by

protein content obtained by a protein assay (bicinchoninic

acid; Pierce). The NADPH-dependent O2� production for

each treatment, a measure of the NADPH oxidase activity,

was estimated by averaging over three independent experi-

ments and expressed as relative lucigenin chemilumines-

cence in mean arbitrary units/min/mg proteinTS.D.

Significant differences between treatments and the corre-

sponding normoxic controls were established by paired t-

test at P <.05.

S.F. Martin et al. / Vascular Pharmacology 43 (2005) 148–156 151

3. Results

3.1. Both H/R and ND increase EC ROS production

ROS generation in HUVECs was monitored using

DCFH-DA and flow cytometry. Exogenously added

H2O2, and organic hydroperoxides (tBOOH and

CumOOH) were chosen as positive controls to show an

increase in intracellular ROS at concentrations and times of

incubation that were previously shown to induce lipid

peroxidation in cell membranes (Okimoto et al., 2000;

Takahashi et al., 2001). From the flow cytometry histo-

grams, it was found that H2O2 (0.5 or 5 AM), tBOOH (0.75

or 1.5 AM) and CumOOH (1 or 4 AM) increased DCF

fluorescence after 1 h of incubation in a dose-dependent

manner (Fig. 1). Stimulation of ECs preloaded with DCFH-

DA with TNF-a at 30 ng/ml for 1 h also resulted in a burst

of intracellular ROS (Fig. 1). H (2 h)/R caused a marginal

increase in DCF fluorescence which was more pronounced

at 5 min from the beginning of R and declined with

progression of R. Prolonged H (24 h)/R was much more

efficient in increasing the DCF fluorescence and, as noticed

under earlier conditions, the signal was greater at 5 min (a

5-fold increase as compared to labeled normoxic control

cells) than at 30 min following R (a 3-fold increase as

compared to labeled normoxic control cells) (Fig. 1). ND

(24 h), a component of the H (24 h) treatment, was alone

sufficient to increase the DCF fluorescence signal by 6-fold

over the labeled normoxic control cells, a level slightly

higher than that due to H (24 h)/R (5 min) (Fig. 1).

Although the cells homogeneously responded by increasing

their intracellular ROS in response to either hydroper-

100 101 102 103 104

100

labeled controlunlabeled control

tBOOH (0.75 µM)tBOOH (1.5 µM)

Fluoresce

100H (2 h)/R (5 min) H (2 h)/R (30 min)

100

Cou

nts

TNF-α (30 ng/ml)ND (24 h)

1

1

1

Fig. 1. Intracellular ROS concentration increases in HUVECs due to exposure to

either exogenously added hydroperoxides (tBOOH, CumOOH, H2O2), TNF-a, H/

to DCF fluorescence distribution (X-axis: DCF fluorescence, Y-axis: number of ce

shown.

oxides, H (2 h) /R or ND (24 h), two cell populations, one

with lower and one with higher fluorescence intensity, were

formed in response to H (24 h) /R (Fig. 1).

3.2. Both H/R and ND increase EC lipid peroxidation

In order to find out if the above treatments increase the

extent of lipid peroxidation, lipid peroxidation was moni-

tored by detection of (a) DPPP fluorescence following

reaction specifically with membrane lipid hydroperoxides

(Okimoto et al., 2000), and (b) HNE, the most prominent

aldehyde generated from hydroperoxides of polyunsaturated

fatty acids that accumulates in cell membranes following

oxidative stress (Esterbauer et al., 1991; Uchida et al.,

1993). When HUVECs preloaded with DPPP were stimu-

lated with either H2O2, tBOOH or CumOOH for 1 h, and

analyzed by flow cytometry, the results showed that

membrane lipid hydroperoxides were formed in a dose-

dependent manner, and tBOOH and CumOOH were more

efficient than H2O2 in causing lipid peroxidation (Fig. 2).

TNF-a, at 30 ng/ml for 1 h, induced a marginal increase in

DPPP fluorescence (Fig. 2). H/R was found to increase the

DPPP fluorescence signal in HUVECs with both times of H

and R. Specifically, H (2 h)/R (30 min) showed no

detectable increase in DPPP fluorescence, whereas H (24

h)/R (2 h) showed a 4-fold increase over the DPPP-

preloaded normoxic control cells (Fig. 2). A two-peak

distribution was evident in the DPPP fluorescence from cells

exposed to H (24 h) /R, indicating that the two cell

populations that raised their intracellular ROS to different

levels also experienced different levels of lipid peroxidation.

ND (24 h) alone caused a 3-fold increase in DPPP

100 101 102 103 104

nce intensity

H (24 h)/R (5 min) H (24 h)/R (30 min)

00CumOOH (1 µM) CumOOH (4 µM)

H2O2 (0.5 µM) H2O2 (5 µM)

00

00

H/R and ND. Cells were prelabeled with DCFH-DA and then exposed to

R or ND, as described in Materials and methods. Histograms corresponding

lls) from one representative experiment out of three with similar results are

104103102101100104103102101100

50

labeled controlunlabeled control

tBOOH (0.75 µM)tBOOH (1.5 µM)

50CumOOH (1 µM)CumOOH (4 µM)

Cou

nts

50H2O2 (0.5 µM)H2O2 (5 µM)

Fluorescence intensity

100

100H (2 h)/R (30 min)H (2 h)/R (2 h)

100H (24 h)/R (30 min) H (24 h)/R (2 h)

TNF-α (30 ng/ml)ND (24 h)

Fig. 2. Membrane lipid peroxidation, as measured by a fluorescent probe, increases in HUVECs following exposure to H/R and ND. Cells were prelabeled with

DPPP before they were stimulated by exposure to exogenously added hydroperoxides (tBOOH, CumOOH, H2O2), TNF-a, HR or ND, as described in

Materials and methods. Histograms corresponding to DPPP_O fluorescence distribution (X-axis: DPPP_O fluorescence, Y-axis: number of cells) from a

representative experiment out of three with similar results are shown.

S.F. Martin et al. / Vascular Pharmacology 43 (2005) 148–156152

fluorescence (Fig. 2), whereas H (24 h) in EGM-2 complete

medium did not show any detectable increase in DPPP

fluorescence (not shown), suggesting that H without ND is

not primarily responsible for inducing lipid peroxidation.

When the formation of HNE-modified proteins was

examined by Western blotting in HUVECs exposed to

different treatments of H/R and ND, it was found that the

intensity of certain bands increased following H/R as

compared to the normoxic control cells in proportion to

the H period alone (Fig. 3). As in the case of DPPP

fluorescence, ND (24 h) alone was sufficient to induce an

accumulation of HNE-modified proteins to at least the same

Fig. 3. Membrane lipid peroxidation, as measured by Western blotting of

HNE-modified proteins, increases in HUVECs following exposure to H/R

and ND. Following cell exposure to different treatments (normoxia, TNF-a,

H/R or ND), the accumulation of HNE-modified proteins was determined

by Western blotting of cell lysates using an anti-HNE polyclonal antibody,

as described in Materials and methods. Data from one representative

experiment out of two with identical results are shown.

magnitude as that caused by H (24 h)/R (2 h) (Fig. 3).

Although TNF-a caused a pronounced increase in densito-

metric intensity of HNE-protein adducts, the band profile

modified by TNF-a was different from that modified by

either H/R or ND (Fig. 3).

Fig. 4. NADPH-dependent O2� production, detected by lucigenin-enhanced

chemiluminescence in intact HUVECs, increases following cell exposure to

H/R and ND. Cells infected with either AdhGal or AdRac1N17 were

exposed to different treatments (normoxia, ND or H/R). Cells were

harvested and distributed in wells of 96-well plates where NADPH and

lucigenin were added, as described in Materials and methods. Light

emission was recorded and expressed as mean arbitrary units/min/mg

protein T S.D. for n =3 independent experiments. The data were analyzed

for statistical significance compared to their corresponding control by

paired t-test: *P <.05, **P <.01.

S.F. Martin et al. / Vascular Pharmacology 43 (2005) 148–156 153

3.3. Both H/R and ND increase the NADPH-dependent O2�

production

Since a phagocyte-type NADPH oxidase is thought to be

a major source of EC ROS (Al-Mehdi et al., 1998;

Mohazzab et al., 1994) and this oxidase generates ROS

following H/R (Kim et al., 1998), we examined if H/R and

ND can each increase the NADPH-dependent O2� produc-

tion by intact HUVECs as measured by lucigenin-enhanced

chemiluminescence. A basal level of lucigenin chemilumi-

nescence even by the normoxic control cells that had been

infected with AdhGal was detected (Fig. 4). Normoxic cells,

either uninfected or infected with the control virus, had the

same level of lucigenin chemiluminescence (not shown). In

ECs infected with AdhGal, ND (24 h) exhibited a 9-fold

increase and H (24 h)/R (2 h) a 5-fold increase, respectively,

compared to the normoxic control (Fig. 4). In cells infected

with AdRac1N17, which results in expression of the

dominant-negative form of Rac1, there was no increase in

lucigenin chemiluminescence following either ND or H/R

treatments, as compared to their corresponding normoxic

control (Fig. 4).

3.4. Rac1-dependent ROS contribute to lipid peroxidation

following either H/R or ND

To examine the role of Rac1-dependent and NADPH

oxidase-generated ROS in the lipid peroxidation following

ND or H/R, the formation of HNE-protein conjugates was

examined by Western blotting in lysates of HUVECs

exposed to ND (24 h) or H (24 h)/R (2 h) following

infection with either the control virus or AdRac1N17.

Although equal amount of protein was loaded for each cell

type based on Coomassie blue staining of the gel, the profile

of HNE-modified proteins was found to be different in

H (2

4 h)/R

(2 h

)

H (2

4 h)/R

(2 h

)

Norm

oxia

Norm

oxia

ND (24 h

)

ND (24 h

)

Mol

. Wt.

(kD)

AdβGal

AdRac1N17

113

5241

92

3529

187

HNE-modifiedproteins

Fig. 5. Rac1 inhibition attenuates the H/R-and ND-induced increases in

membrane lipid peroxidation in HUVECs. Cells infected with either

AdhGal or AdRac1N17 were exposed to different treatments (normoxia,

ND or H/R). The accumulation of HNE-modified proteins was determined

by Western blotting of cell lysates using an anti-HNE polyclonal, as

described in Materials and methods. Data from one representative experi-

ment out of two with identical results are shown.

uninfected cells versus cells infected with the control virus

as compared with cell monolayers exposed to either

normoxia (compare lane 1 of Fig. 3 to lane 1 of Fig. 5),

H/R (compare lane 3 of Fig. 3 to lane 3 of Fig. 5) or ND

(compare lane 5 of Fig. 3 to lane 2 of Fig. 5). Specifically, in

cells infected with the control virus under normoxia, bands

corresponding to ¨70–80 kD HNE-protein adducts appear

more intense compared to uninfected cells under normoxia,

possibly due to redox status changes in cells following viral

infection. In AdhGal-infected cells, ND (24 h) resulted in an

increase in lipid peroxidation at least as large as that

exhibited by cells exposed to H (24 h)/R (2 h) (Fig. 5), as in

the case of uninfected cells (Fig. 3). Rac1 inhibition

significantly reduced the increase in lipid peroxidation due

to either ND or H/R (Fig. 5), suggesting that the ROS

generated by the Rac1-dependent pathway are major

contributors to the ND- and H/R-induced lipid peroxidation.

Similarly, when lipid peroxidation was monitored by DPPP

fluorescence, Rac1 inhibition successfully suppressed both

the ND- and H/R-induced increases in fluorescence (not

shown).

4. Discussion

The present study provides evidence that targeted

inhibition of the small GTPase Rac1, which regulates

ROS production via a membrane-bound NADPH oxidase

(Abo et al., 1991), protects against the H/R-induced lipid

peroxidation in cultured HUVECs. This agrees with

findings from in vivo studies which have revealed that

ischemia-induced lipid peroxidation in perfused rat lungs,

as measured by thiobarbituric acid reactive substances, was

found significantly inhibited by pretreatment with PR-39, a

synthetic peptide that inhibits activity of the lung

endothelial NADPH oxidase by blocking its assembly

(Al-Mehdi et al., 1998). In another study, inhibition of

Rac1 by adenoviral expression of Rac1N17 was shown to

markedly decrease the production of lipid peroxidation

products (malondialdehyde and lipid hydroperoxides) in a

mouse model of hepatic ischemia/reperfusion (Ozaki et al.,

2000). Furthermore, the present study provides evidence

that ND alone, of the same duration as H, is sufficient to

cause at least to the same extent of Rac1-dependent ROS

production and lipid peroxidation as caused by H/R,

suggesting that ND may be more important than H in

activating Rac1 and inducing the production of injurious

ROS by ischemic ECs. In agreement to that, a recent study

has shown that ND rather than H induces the production of

MCP-1 by H/R-exposed cultured human aortic ECs and

MCP-1 production requires a functional Rac1 (Lopes et

al., 2002). In that study, they measured Rac1 activation by

a PAK-affinity precipitation assay and demonstrated that

ND strongly activates Rac1 after 15–30 min and then

induces a prolonged but low level of activation for up to

18 h (Lopes et al., 2002).

S.F. Martin et al. / Vascular Pharmacology 43 (2005) 148–156154

In the present study, we measured the ROS production

by cultured HUVECs at 5 and 30 min from the onset of R,

based on our earlier studies wherein O2� generation was

detected by lucigenin chemiluminescence as early as 5 min

of R (Ng et al., 2002). ND (24 h) alone resulted in at least

as high ROS production as that induced by H (24 h)/R (5

min), when measured by a probe like DCF which is non-

specific to a particular ROS. The existence of two cell

populations (different DCF fluorescence levels) in response

to H (24 h) /R (5 min) may be due to different kinds of

ROS generated by prolonged H (24 h) compared to all

other treatments tested, which, based on their intracellular

distribution and relative reactivity with the trapped DCFH,

result in a two-peak DCF fluorescence distribution. The

NADPH-dependent O2� production by intact ECs is known

to be detectable using lucigenin chemiluminescence,

minimally affected by increasing lucigenin concentration

from 5 up to 400 AM, and is abolished by the flavoprotein

inhibitor diphenylene iodonium suggesting that it can

provide a measure of cellular NADPH oxidase activity

(Li and Shah, 2001). In the present study, ND (24 h)

resulted in higher production of NADPH-dependent O2�, as

measured by lucigenin chemiluminescence, as compared to

H (24 h)/R (2 h). This may be due to additional changes in

Rac1 activity (and resultant NADPH-dependent O2�

production) during the 24 h of H and 2 h of R that are

absent during exposure to ND (24 h under normoxia). The

fact that H/R and ND each generated comparable amounts

of total ROS but different amounts of NADPH-dependent

O2� may be due to activation of other ROS sourecs, such

as xanthine oxidase, cyclooxygenase, phospholipase A2,

endothelial constitutive nitric oxide synthase, the mito-

chondrial (NADH dehydrogenase, cytochrome oxidase)

and the microsomal (cytochrome P-450) electron transport

chains, during the 24 h of H and 2 h of R that are absent

during ND treatment (24 h under normoxia). EC ROS

during H in the absence of ND, in particular, are believed

to originate primarily from the mitochondrial electron

transport chain (Pearlstein et al., 2002). In this study,

functional Rac1 was found to be required for the increase

in NADPH-dependent O2� production due to either ND or

H/R. Although the effect of Rac1 inhibition on the H/R- or

ND-induced increase in ROS production using DCF

fluorescence was not examined, it is known that intra-

cellular ROS levels following H/R or ND are markedly

attenuated in cells either infected with AdRac1N17 (Kim

et al., 1998), pretreated with diphenylene iodonium or

pretreated with the more specific NADPH oxidase inhib-

itor gp91ds-tat (Lopes et al., 2002). This suggests that the

Rac1-regulated NADPH oxidase is a major source of ROS

following H/R or ND. Therefore, ND is primarily res-

ponsible for the production of Rac1-dependent O2� and

total intracellular ROS following H/R, without ruling out

the possibility of independent effects by either H in the

absence of ND or R on Rac1 activation and ROS

production.

The times of detection of lipid peroxidation, 30 min and

2 h after R exposure, were chosen based on the time profile

of lipid hydroperoxide formation as it was monitored by

DPPP_O fluorescence in phorbol ester-stimulated poly-

morphonuclear leukocytes (Okimoto et al., 2000). In order

to show that the extent of DPPP oxidation reflects the extent

of lipid peroxidation within cell membranes, we compared

the reactivities of three hydroperoxides with different

polarities and solubilities with DPPP. Due to its high

lipophilicity, DPPP is supposed to localize within lipid

membranes and react preferentially with lipophilic hydro-

peroxides. H2O2 is highly permeable in the cell membrane

and does not stay within the membrane for long duration to

effectively react with DPPP, hence, the lowest signal

observed among the hydroperoxides tested. CumOOH gave

a higher signal than tBOOH at similar concentrations in

DPPP-labeled U937 cells (Takahashi et al., 2001), but

tBOOH was found to be the most effective stimulus in

DPPP-labeled ECs. No quantitative relationship could be

established between the level of ROS and the resultant

extent of lipid peroxidation. For example, TNF-a resulted in

a markedly greater increase in ROS but a modest increase in

DPPP_O fluorescence. It should be mentioned that

DPPP_O fluorescence provides a cumulative history of

lipid peroxidation of the cell membrane within the course of

the experiment, thus, the signal keeps on amplifying with

time. In addition, it became obvious that the two methods

used to detect lipid peroxidation had different selection

criteria. Thus, TNF-a resulted in a modest DPPP_O

fluorescence signal, but showed a big increase in densito-

metric intensity of certain bands as revealed by the Western

blotting with an anti-HNE antibody. Furthermore, stimula-

tion by TNF-a resulted in increased intensity of different

protein bands containing HNE-adducts as compared to

stimulation by H/R or ND. Both these differences could be

explained if the exact subcellular localization of ROS

following cell exposure to each treatment (as well as the

exact localization of the probe in the case of DPPP) was

known. For example, following stimulation by TNF-a, there

are mitochondrial sources of ROS that are activated

(Deshpande et al., 2000) and this may result in modification

of different proteins by HNE as compared to those modified

due to either H/R or ND. Even viral infection with a control

virus resulted in a different pattern of HNE-protein adducts,

compared to uninfected cells, possibly due to an effect of the

infection itself on the cellular redox status and in agreement

with earlier studies from our group where control infected

cells showed differences in the ROS-mediated profile of

cytosolic tyrosine phosphorylated proteins compared to

uninfected cells (Yeh et al., 1999).

Viability of HUVECs was measured by the trypan blue

exclusion method and was found not to be significantly

affected by H(24 h)/R (2 h with serum readmission) or ND

(24 h). This agrees with other reports which did not find cell

lysis following an overnight exposure of cells to H and an

additional 3 h exposure to R (Mold and Morris, 2001).

S.F. Martin et al. / Vascular Pharmacology 43 (2005) 148–156 155

Contrastingly, some studies showed that either H (16 h)/R (5

min) or ND (24 h) led to ¨30% of cell death (Kim et al.,

1998; Lopes et al., 2002). In the current study, the effect of

different treatments without or with Rac1 inhibition on EC

apoptosis was not examined. However, cellular lipid peroxyl

radicals or lipid hydroperoxides formed in ECs stimulated

by oxidants are known to initiate caspase-3 activation and to

induce an apoptotic signaling pathway (Kotamraju et al.,

2001). One peroxidation product that is involved in the

apoptotic signaling pathway is HNE itself, since it can

activate caspases such as caspase-l, -2, -3 and -8 in human

lens epithelial cells (Choudhary et al., 2002; Herbst et al.,

1999). Based on these findings and also on the fact that

Rac1 inhibition limits the H/R- and ND-induced lipid

peroxidation, it is expected that both H/R and ND will

cause EC apoptosis and Rac1 inhibition will be protective

against apoptosis. Indeed, it has been shown that H (24 h in

complete culture media)/R (3 h) causes ¨30% of human

coronary artery ECs to become apoptotic, as measured by

fragmented DNA end-labeling in situ (Li et al., 1998). ND

during H may further increase apoptosis as revealed by a

study in which bovine aortic ECs exposed to either 48 h of

H in the presence of serum or 48 h of H without serum

exhibited that H plus serum results in ¨15% apoptosis,

whereas H without serum in ¨30% apoptosis (Hogg et al.,

1999). As for the protective effect of Rac1 inhibition on

apoptosis, it was demonstrated that, in an in vivo model of

mouse hepatic ischemia/reperfusion, Rac1N17 expression

significantly suppressed the ischemia/reperfusion-induced

apoptosis in liver (Ozaki et al., 2000).

In summary, H/R induces ROS production and lipid

peroxidation in cultured HUVECs. The Rac1-dependent and

NADPH oxidase-generated ROS are major contributors to

the H/R-induced lipid peroxidation. ND, an integral

component of ischemia, is responsible for most of the H/

R-induced increase in Rac1-dependent ROS production and

resultant EC lipid peroxidation. Thus, targeted inhibition of

the small GTPase Rac1 may be an effective approach in

clinical conditions in which ischemia occurs.

Acknowledgements

We thank K. Irani for the generous gift of AdhGal andAdRac1N17, and A. Kolmakova and S. Deshpande for

technical assistance. This work was supported by National

Institutes of Health grant RO1-HL67027 (B. R. Alevriadou).

S. F. Martin was partially supported by a fellowship from

Ministerio de Educacion y Cultura, Spain.

References

Abo, A., Pick, E., Hall, A., Totty, N., Teahan, C.G., Segal, A.W., 1991.

Activation of the NADPH oxidase involves the small GTP-binding

protein p21rac1. Nature 353, 668–670.

Acosta, D., Li, C.P., 1979. Injury to primary cultures of rat heart endothelial

cells by hypoxia and glucose deprivation. In Vitro 15, 929–934.

Al-Mehdi, A.B., Zhao, G., Dodia, C., Tozawa, K., Costa, K., Muzykantov,

V., Ross, C., Blecha, F., Dinauer, M., Fisher, A.B., 1998. Endothelial

NADPH oxidase as the source of oxidants in lungs exposed to ischemia

or high K+. Circ. Res. 83, 730–737.

Choudhary, S., Zhang, W., Zhou, F., Campbell, G.A., Chan, L.L.,

Thompson, E.B., Ansari, N.H., 2002. Cellular lipid peroxidation end-

products induce apoptosis in human lens epithelial cells. Free Radic.

Biol. Med. 32, 360–369.

Deshpande, S.S., Angkeow, P., Huang, J., Ozaki, M., Irani, K., 2000. Rac1

inhibits TNF-alpha-induced endothelial cell apoptosis: dual regulation

by reactive oxygen species. FASEB J. 14, 1705–1714.

Esterbauer, H., Schaur, R.J., Zollner, H., 1991. Chemistry and biochemistry

of 4-hydroxynonenal, malonaldehyde and related aldehydes. Free

Radic. Biol. Med. 11, 81–128.

Girotti, A.W., 1998. Lipid hydroperoxide generation, turnover, and effector

action in biological systems. J. Lipid Res. 39, 1529–1542.

Granger, D.N., Korthuis, R.J., 1995. Physiologic mechanisms of post-

ischemic tissue injury. [Review]. Annu. Rev. Physiol. 57, 311–332.

Grisham, M.B., Granger, D.N., Lefer, D.J., 1998. Modulation of leukocyte-

endothelial interactions by reactive metabolites of oxygen and nitrogen:

relevance to ischemic heart disease [Review]. Free Radic. Biol. Med.

25, 404–433.

Guzman, R.J., Hirschowitz, E.A., Brody, S.L., Crystal, R.G., Epstein, S.E.,

Finkel, T., 1994. In vivo suppression of injury-induced vascular

smooth muscle cell accumulation using adenovirus-mediated transfer

of the herpes simplex virus thymidine kinase gene. Proc. Natl. Acad.

Sci. U. S. A. 91, 10732–10736.

Gyllenhammar, H., 1987. Lucigenin chemiluminescence in the assess-

ment of neutrophil superoxide production. J. Immunol. Methods 97,

209–213.

Herbst, U., Toborek, M., Kaiser, S., Mattson, M.P., Hennig, B., 1999. 4-

Hydroxynonenal induces dysfunction and apoptosis of cultured endo-

thelial cells. J. Cell. Physiol. 181, 295–303.

Hogg, N., Browning, J., Howard, T., Winterford, C., Fitzpatrick, D., Gobe,

G., 1999. Apoptosis in vascular endothelial cells caused by serum

deprivation, oxidative stress and transforming growth factor-h. Endo-thelium 7, 35–49.

Ischiropoulos, H., Gow, A., Thom, S.R., Kooy, N.W., Royall, J.A., Crow,

J.P., 1999. Detection of reactive nitrogen species using 2,7-dichlorodi-

hydrofluorescein and dihydrorhodamine 123. Methods Enzymol. 301,

367–373.

Kim, K.S., Takeda, K., Sethi, R., Pracyk, J.B., Tanaka, K., Zhou, Y.F., Yu,

Z.X., Ferrans, V.J., Bruder, J.T., Kovesdi, I., Irani, K., Goldschmidt-

Clermont, P., Finkel, T., 1998. Protection from reoxygenation injury by

inhibition of rac1. J. Clin. Invest. 101, 1821–1826.

Kotamraju, S., Hogg, N., Joseph, J., Keefer, L.K., Kalyanaraman, B., 2001.

Inhibition of oxidized low-density lipoprotein-induced apoptosis in

endothelial cells by nitric oxide. J. Biol. Chem. 276, 17316–17323.

Kriska, T., Korytowski, W., Girotti, A.W., 2002. Hyperresistance to

photosensitized lipid peroxidation and apoptotic killing in 5-amino-

levulinate-treated tumor cells overexpressing mitochondrial GPX4. Free

Radic. Biol. Med. 33, 1389–1402.

Li, J.M., Shah, A.M., 2001. Differential NADPH-versus NADH-dependent

superoxide production by phagocyte-type endothelial cell NADPH

oxidase. Cardiovasc. Res. 52, 477–486.

Li, D., Yang, B., Mehta, J.L., 1998. Ox-LDL induces apoptosis in human

coronary artery endothelial cells: role of PKC, PTK, bcl-2, and Fas. Am.

J. Physiol. 275, H568–H576.

Lopes, N.H., Vasudevan, S.S., Gregg, D., Selvakumar, B., Pagano, P.J.,

Kovacic, H., Goldschmidt-Clermont, P.J., 2002. Rac-dependent mono-

cyte chemoattractant protein-1 production is induced by nutrient

deprivation. Circ. Res. 91, 798–805.

McLeod, L.L., Alayash, A.I., 1999. Detection of a ferrylhemoglobin

intermediate in an endothelial cell model after hypoxia-reoxygenation.

Am. J. Physiol. 277, H92–H99.

S.F. Martin et al. / Vascular Pharmacology 43 (2005) 148–156156

McLeod, L.L., Sevanian, A., 1997. Lipid peroxidation and modification of

lipid composition in an endothelial cell model of ischemia and

reperfusion. Free Radic. Biol. Med. 23, 680–694.

Michiels, C., Arnould, T., Houbion, A., Remacle, J., 1992. Human

umbilical vein endothelial cells submitted to hypoxia-reoxygenation in

vitro: implication of free radicals, xanthine oxidase, and energy

deficiency. J. Cell. Physiol. 153, 53–61.

Mohazzab, K.M., Kaminski, P.M., Wolin, M.S., 1994. NADH oxidoreduc-

tase is a major source of superoxide anion in bovine coronary artery

endothelium. Am. J. Physiol. 266, H2568–H2572.

Mold, C., Morris, C.A., 2001. Complement activation by apoptotic

endothelial cells following hypoxia/reoxygenation. Immunology 102,

359–364.

Ng, C.K., Deshpande, S.S., Irani, K., Alevriadou, B.R., 2002. Adhesion of

flowing monocytes to hypoxia-reoxygenation-exposed endothelial cells:

role of Rac1, ROS, and VCAM-1. Am. J. Physiol. 283, C93–C102.

Okimoto, Y., Watanabe, A., Niki, E., Yamashita, T., Noguchi, N., 2000. A

novel fluorescent probe diphenyl-1-pyrenylphosphine to follow lipid

peroxidation in cell membranes. FEBS Lett. 474, 137–140.

Ozaki, M., Deshpande, S.S., Angkeow, P., Bellan, J., Lowenstein, C.J.,

Dinauer, M.C., Goldschmidt-Clermont, P.J., Irani, K., 2000. Inhibition

of the Rac1 GTPase protects against nonlethal ischemia/reperfusion-

induced necrosis and apoptosis in vivo. FASEB J. 14, 418–429.

Pak, J.H., Manevich, Y., Kim, H.S., Feinstein, S.I., Fisher, A.B., 2002.

An antisense oligonucleotide to 1-cys-peroxiredoxin causes lipid

peroxidation and apoptosis in lung epithelial cells. J. Biol. Chem.

277, 49927–49934.

Pearlstein, D.P., Ali, M.H., Mungai, P.T., Hynes, K.L., Gewertz, B.L.,

Schumacker, P.T., 2002. Role of mitochondrial oxidant generation in

endothelial cell responses to hypoxia. Arterioscler. Thromb. Vasc. Biol.

22, 566–573.

Ratych, R.E., Chuknyiska, R.S., Bulkley, G.B., 1987. The primary

localization of free radical generation after anoxia/reoxygenation in

isolated endothelial cells. Surgery 102, 122–131.

Russell, W.J., Matalon, S., Jackson, R.M., 1994. Manganese superoxide

dismutase expression in alveolar type II epithelial cells from non-

ventilated and hypoperfused lungs. Am. J. Respir. Cell Mol. Biol. 11,

366–371.

Sulciner, D.J., Irani, K., Yu, Z.X., Ferrans, V.J., Goldschmidt-Clermont, P.,

Finkel, T., 1996. Rac1 regulates a cytokine-stimulated, redox-dependent

pathway necessary for NF-kappaB activation. Mol. Cell. Biol. 16,

7115–7121.

Sundaresan, M., Yu, Z.X., Ferrans, V.J., Irani, K., Finkel, T., 1995.

Requirement for generation of H2O2 for platelet-derived growth factor

signal transduction. Science 270, 296–299.

Takahashi, M., Shibata, M., Niki, E., 2001. Estimation of lipid peroxidation

of live cells using a fluorescent probe, diphenyl-1-pyrenylphosphine.

Free Radic. Biol. Med. 31, 164–174.

Terada, L.S., 1996. Hypoxia-reoxygenation increases O2� efflux which

injures endothelial cells by an extracellular mechanism. Am. J. Physiol.

270, H945–H950.

Uchida, K., Szweda, L.I., Chae, H.Z., Stadtman, E.R., 1993. Immuno-

chemical detection of 4-hydroxynonenal protein adducts in oxidized

hepatocytes. Proc. Natl. Acad. Sci. U. S. A. 90, 8742–8746.

Yang, B., Graham, L., Dikalov, S., Mason, R.P., Falck, J.R., Liao, J.K.,

Zeldin, D.C., 2001. Overexpression of cytochrome P450 CYP2J2

protects against hypoxia-reoxygenation injury in cultured bovine aortic

endothelial cells. Mol. Pharmacol. 60, 310–320.

Yeh, L.H., Park, Y.J., Hansalia, R.J., Ahmed, I.S., Deshpande, S.S.,

Goldschmidt-Clermont, P.J., Irani, K., Alevriadou, B.R., 1999. Shear-

induced tyrosine phosphorylation in endothelial cells requires Rac1-

dependent production of ROS. Am. J. Physiol. 276, C838–C847.

Zweier, J.L., Kuppusamy, P., Lutty, G.A., 1988. Measurement of endothe-

lial cell free radical generation: evidence for a central mechanism of free

radical injury in postischemic tissues. Proc. Natl. Acad. Sci. U. S. A. 85,

4046–4050.