Antiatherogenic effect of bisvanillyl-hydralazone, a new hydralazine derivative with antioxidant, carbonyl scavenger, and antiapoptotic properties

Article abstract of DOI:10.1089/ars.2010.3321

Reactive oxygen species (ROS) generated within the vascular wall trigger low-density lipoprotein (LDL) oxidation, lipid peroxidation, and carbonyl stress that are involved in atherogenesis. We recently reported that the antihypertensive drug, hydralazine, exhibits carbonyl scavenger and antiatherogenic properties, but only moderate antioxidant activity, so that high concentrations are required for inhibiting LDL oxidation. We aimed to develop agents sharing both antioxidant and carbonyl scavenger properties. We have synthesized a new hydralazine derivative, the bisvanillyl-hydralazone (BVH). BVH strongly inhibited LDL oxidation induced by copper and by human endothelial cells (HMEC-1), and prevented the formation of macrophagic foam cells. BVH reduced both the extracellular generation of ROS (superoxide anion and hydrogen peroxide) induced by oxidized LDL (oxLDL), as well as intracellular oxidative stress and proteasome activation, NFkappaB activation, and oxLDL-mediated proinflammatory signaling. In parallel, BVH prevented the carbonyl stress induced by oxLDL on cellular proteins, and blocked the apoptotic cascade as assessed by the inhibition of Bid cleavage, cytochrome C release, and DEVDase activation. Lastly, BVH prevented atherogenesis and carbonyl stress in apoE ^-/- mice. In conclusion, BVH is the prototype of a new class of antioxidant and carbonyl scavenger agents designed for new therapeutical approaches in atherosclerosis. Mary Ann Liebert, Inc.

Full text of DOI:10.1089/ars.2010.3321

ANTIOXIDANTS & REDOX SIGNALING
Volume 14, Number 11, 2011
ª Mary Ann Liebert, Inc.
ORIGINAL RESEARCH COMMUNICATION
DOI: 10.1089=ars.2010.3321
Antiatherogenic Effect of Bisvanillyl-Hydralazone,
a New Hydralazine Derivative with Antioxidant,
Carbonyl Scavenger, and Antiapoptotic Properties
1–3
1,3
Benaissa Bouguerne, Nadji Belkheiri,2,3 Florence Bedos-Belval,2,3 Ce´cile Vindis,
Koji Uchida,4 Hubert Duran,2,3 Marie-He´le`ne Grazide, Michel Baltas,2,3
1,3
Robert Salvayre, and Anne Ne`gre-Salvayre1,3
1,3
Abstract
Reactive oxygen species (ROS) generated within the vascular wall trigger low-density lipoprotein (LDL) oxi-
dation, lipid peroxidation, and carbonyl stress that are involved in atherogenesis. We recently reported that the
antihypertensive drug, hydralazine, exhibits carbonyl scavenger and antiatherogenic properties, but only
moderate antioxidant activity, so that high concentrations are required for inhibiting LDL oxidation. We aimed
to develop agents sharing both antioxidant and carbonyl scavenger properties. We have synthesized a new
hydralazine derivative, the bisvanillyl-hydralazone (BVH). BVH strongly inhibited LDL oxidation induced by
copper and by human endothelial cells (HMEC-1), and prevented the formation of macrophagic foam cells. BVH
reduced both the extracellular generation of ROS (superoxide anion and hydrogen peroxide) induced by oxi-
dized LDL (oxLDL), as well as intracellular oxidative stress and proteasome activation, NFkappaB activation,
and oxLDL-mediated proinflammatory signaling. In parallel, BVH prevented the carbonyl stress induced by
oxLDL on cellular proteins, and blocked the apoptotic cascade as assessed by the inhibition of Bid cleavage,
cytochrome C release, and DEVDase activation. Lastly, BVH prevented atherogenesis and carbonyl stress in
apoEꢀ=ꢀ mice. In conclusion, BVH is the prototype of a new class of antioxidant and carbonyl scavenger agents
designed for new therapeutical approaches in atherosclerosis. Antioxid. Redox Signal. 14, 2093–2106.
Introduction
products oxidized phospholipids, oxysterols (22, 30), and lipid
peroxidation-derived aldehydes (4-hydroxynonenal [4-HNE],
xidative stress plays a major role in atherogenesis malondialdehyde [MDA], and acrolein) (13, 14, 59). These al-
and cardiovascular diseases (50), via its implication in the dehydes react with free amino groups and thiol residues of
O
oxidative modifications of low-density lipoproteins (LDLs), proteins, thereby forming protein adducts (‘‘carbonyl stress’’),
that render them highly atherogenic (4, 7, 25, 54, 64). Oxidized which modify and alter the function of circulating (e.g., LDL)
LDL (oxLDL) are metabolically deviated toward macrophagic and cellular proteins (14, 40). Carbonyl stress occurs as a
cells and taken-up via the scavenger receptor system, leading to downstream consequence of oxidative stress, and it is inhibited
the formation of foam cells that accumulate as fatty streaks in when blocking lipid oxidation by antioxidants (52, 40). How-
the arterial intima (43, 54, 55). Early lesions may be complicated ever, antioxidants fail to neutralize carbonyl stress, once ad-
by chronic inflammatory processes and local remodeling of the ducts are formed on proteins (40, 42). Carbonyl stress is
arterial wall, leading finally to more complex atherosclerotic involved in cell growth arrest, inhibition of cell migration, in-
plaques (33). Besides, oxLDLs exert various biological proper- flammation, and (at a lesser extent) apoptosis, and is thought to
ties (including inflammation, cell proliferation, migration, contribute to the progression of atherosclerosis (40, 42).
growth arrest, and apoptosis) that play a role in the formation
The ability of antioxidants to reduce the coronary risk
of complicated lesions, and in plaque erosion and rupture (31, in humans is largely debated. The antiatherogenic effect of
39, 44, 47, 48). These responses depend on the local concen- antioxidants has been extensively studied through in vivo
tration of oxLDLs, and on their content in lipid peroxidation preclinical studies on animal models for atherosclerosis and in
1Biochimie, INSERM U858, Team ‘‘Atherosclerosis,’’ Toulouse, France.
2CNRS, Laboratoire de Synthe`se et Physico-Chimie de Molecules d’Inte´reˆt Biologique, Toulouse, France.
3Universite´ Paul Sabatier, Toulouse Cedex, France.
4Laboratory of Food and Biodynamics, Graduate School of Bioagricultural Sciences, Nagoya University, Nagoya, Japan.
2093

2094
BOUGUERNE ET AL.
human trials (3, 5, 15, 23, 57, 63). Most antioxidants are highly (IR) spectra were recorded on a Perkin-Elmer 1725 IR spec-
efficient in preventing atherosclerosis development in ani- trophotometer and the data are reported in inverse centime-
mals (which mainly allow to investigate the first steps in ters. UV spectra were recorded on a Perkin-Elmer lambda 17
atherogenesis, i.e., foam cell formation), but fail to prevent the UV-vis spectrophotometer. Proton nuclear magnetic reso-
late athero-thrombotic complications of atherosclerosis and nance (1H NMR) spectra were obtained with a Bruker AC-
the subsequent cardiovascular events in human trials (3, 23, 300 MHz spectrometer. Chemical shifts were reported in parts
63). The low apparent benefit of antioxidant therapy does not per million (ppm) and signals are given as follows: s, singlet;
question the implication of oxidative stress and LDL oxida- d, doublet; t, triplet; m, multiplet. Mass spectra were recorded
tion in atherogenesis, as it may result from diverse mecha- on an R 10-10 C Nermag (70 eV) quadripolar spectrome-
nisms, such as difficulties in delivering effective in situ ter using desorption chemical ionization or electrospray
antioxidant therapy, including a lack of antioxidant avail- techniques.
ability, or a lack of effect on reactive oxygen species (ROS)
involved in vivo in LDL oxidation and inflammatory signal-
Bisvanilline 2 synthesis. To vanillin (10.64 g, 70 mmol) in
ing. Inhibiting carbonyl stress (which is a consequence of 700 ml of water was added FeSO4 (0.4 g, 1.4 mmol) under
oxidative stress) may represent a new therapeutic strategy for stirring. After heating during 10 min at 508C, Na2S2O8 (8.93 g,
patients, but so far, it is not known if the use of carbonyl 37.5 mmol) was added. The reaction mixture was stirring at
scavengers could be beneficial or not on the late steps of 508C for 5 days. The brown precipitate formed was filtered
atherosclerosis (40, 42). Most carbonyl scavengers usually off. The solid was dissolved in NaOH (2 M) aqueous solution.
exhibit poor antioxidant activity and inhibit neither LDL ox- HCl (2 M) aqueous solution was added. Brown solid formed
idation nor the generation of toxic lipid oxidation products was isolated by filtration (10 g, 95%). mp > 2708C. IR (KBr) n
(oxysterols and oxidized phospholipids), which are involved cmꢀ1: 3186 (O-H), 1672 (C ¼ O), 1587 (C ¼ C arom.), 1454
in the inflammatory and apoptotic events characterizing (C ¼ C arom.). 1H NMR (DMSO, 300 MHz) d ppm: 3.94 (s, 6H,
atherosclerosis progression (40). Hydrazine derivatives, such OCH3), 7.42 (s, 4H), 9.80 (s, 2H, CHO), 9.89 (s, 2H, OH). 13C
as the antihypertensive drug hydralazine, exhibit potent car- NMR (DMSO, 75 MHz) d ppm: 56.50 (s, 2C, OCH3), 109.70
bonyl scavenger efficiency associated with anti-atherogenic (s, 2C), 125 (s, 2C), 128.23 (s, 2C), 128.62 (s, 2C), 148.60 (s, 2C),
properties in apoEꢀ=ꢀ mice, but present poor or moderate 150.90 (s, 2C), 191.60 (s, 2C, CHO). UV (EtOH=0.2% DMSO,
antioxidant properties, so that high concentrations (50– 25 mM, 258C): l ¼ 303 nm, e ¼ 14,300 molꢀ1.L.cmꢀ1.MS (FAB,
100 mM) are required for inhibiting cell-mediated LDL oxi- MNBA) m=z: 303.1 [MþHþ].
dation (16). Theoretically, agents sharing both antioxidant
and carbonyl scavenger activities should inhibit each step of
BVH synthesis. 5,50-Bisvanilline (BV) 2 (0.2 g, 0.66 mmol)
atherosclerosis progression, including LDL oxidation, oxida- was suspended in absolute ethanol (26 ml). Hydralazine hy-
tive stress, and carbonyl stress as well, and their consequences drochloride (0.26 g, 1.32 mmol) was added. The reaction
in the evolution toward advanced lesions.
mixture was refluxed for 6h30 and filtrated to give pure BVH
We have designed and synthesized a hydralazine derivative, (0.38 g, 87%) as a yellow solid. mp > 2708C; IR (KBr) n cmꢀ1
:
the bisvanillyl-hydralazone (BVH), and we have characterized 3423 (O-H), 1675 (C ¼ N), 1617 (C ¼ C arom.), 1594 (C ¼ C
its antioxidant, carbonyl scavenger, and antiatherogenic prop- arom.), 1473 (C ¼ C arom.). 1H NMR (DMSO, 300 MHz)
erties, more precisely its ability to (i) block oxidative stress, LDL d ppm: 3.94 (s, 6H, OCH3), 7.41 (d, 2H, 4J ¼ 1.8 Hz), 7.78 (d, 2H,
oxidation, carbonyl stress, and foam cell formation, (ii) inhibit 4J ¼ 1.8 Hz), 8.10 (m, 6H), 8.68 (s, 2H, HC ¼ N); 8.95 (s, 2H);
the inflammatory signaling of oxLDL, (iii) neutralize the apo- 9.03 (s, 2H); 9.20 (d, 3J ¼ 7.3 Hz, 2H); 9.29 (bs, 2H, OH); 14.31
ptotic effect of oxLDL in cultured vascular cells, and (iv) slow (bs, 2H, HCl).13C NMR (DMSO, 75 MHz) d ppm: 56.55 (s, 2C,
down or inhibit the formation of atherosclerotic lesions in OCH3), 109.70 (s, 2C, C2), 119.35 (s, 2C, C8), 123.91 (s, 2C),
apoEꢀ=ꢀ mice.
124.26 (s, 2C), 125.12 (s, 2C), 125.38 (s, 2C), 126.61 (s, 2C),
128.25 (s, 2C),128.62 (s, 2C), 134.12 (s, 2C), 136.27 (s, 2C),
147.70 (s, 2C), 148.29 (s, 2C), 148.46 (s, 2C), 157.75 (s, 2C,
HC ¼ N). UV (EtOH=0.2% DMSO, 25 mM, 258C): l ¼ 376 nm,
e ¼ 34,750 molꢀ1.L.cmꢀ1, l ¼ 296 nm, e ¼ 21,500 molꢀ1.L.cmꢀ1
Materials and Methods
Chemicals
Antibodies directed against MCP-1, Bid, gp91phox, and MS (IC, NH3, pos) m=z: 587.2 [MþHþ].
cytochrome C were from Santa Cruz (Tebu-Bio SA). Anti
phospho-NF-kB was from Cell Signaling. AntiCD3 and anti Cell culture
a-actin antibodies were from R&D System. Anti 4-HNE
Human microvascular endothelial cells (HMEC-1) (CDC,
antibody was prepared as reported (58). Acrylamide-4X=
bisacrylamide-2ꢁsolutions were from Euromedex; SYTO-13,
6-carboxy-20,70-dichlorodihydrofluorescein diacetate, di (acet-
oxymethyl ester) (H2DCFDA-AM), dihydroethidium (DHE), and
propidium iodide were from Molecular Probes (InVitrogen); cell
culture reagents and other reagents were from WWR or Sigma.
Dr. Candal) were grown in MCDB131 culture medium sup-
plemented with 10% heat inactivated fetal calf serum,
100 U=ml penicillin, and 100 mg=ml streptomycin. Raw 264.7
murine macrophages were obtained from ATCC and were
grown in DMEM. Twenty-four hours before LDL incorpora-
tion, cells were starved in serum-free RPMI medium.
Synthesis
LDL isolation and oxidation
Organic solvents were purified as described (41) or were
LDL from human pooled sera were prepared by ultracen-
purchased from Aldrich Chemie. Melting points (mp) were trifugation, and dialyzed against phosphate-buffered saline
obtained on a Buchi apparatus and are uncorrected. Infrared (PBS) containing 100 mM EDTA, as previously indicated (62).

ANTIATHEROGENIC EFFECT OF BISVANILLYL-HYDRALAZONE
2095
The mechanism by which BVH protects against LDL oxida- H2DCFDA, and 518=605 nm for DHE, respectively) (18). The
tion has been compared in vitro on LDL oxidation mediated data are expressed as ratio of fluorescence=fluorescence of the
by transition metal-dependent (copper) and -independent unstimulated control (12).
(AAPH) systems. Native LDLs (1 mg=ml) were incubated in
the presence of BVH (5 mM), and either with freshly prepared Proteasome activation
AAPH [2,2’-azobis-(2-amidinopropane) hydrochloride] (at
The in vitro activity of 20S proteasome was determined
the indicated concentrations, 2 h at 378C) (49), or with copper
according to Grune et al. (21). Cells were homogeneized in PBS
(CuSO4 5 mM, 4 h incubation at 378C). LDL oxidation was
containing 0.1% Triton X-100 and 0.5 mM dithiotreitol, and
evaluated by determining the thiobarbituric acid-reactive
the enzymatic activity was measured. The assay mixture
substance (TBARS) content (16). To evaluate cell-mediated
contained 50 ml of buffer (50 mM Tris-HCl pH 7.8, 20 mM KCl,
LDL oxidation, HMEC-1 were seeded in 12-multiwell plates.
5 mM MgCl2, and 0.1 mM DTT), 250 mM of sLLVY-MCA, and
The standard culture medium was removed and replaced
50 ml of cell homogenates (30 mg protein). After 30 min incu-
by serum-free RPMI 1640 containing native LDLs (100 mg
bation at 378C, the reaction was stopped by adding 1 ml of
apoB=ml). BVH, hydralazine, and BV were solubilized in
0.2 M glycine buffer pH 10 and the fluorescence of the liber-
DMSO, and added to the culture medium at the indicated
ated 7-amino-4 methylcoumarin was measured (exc=em
concentrations (and 0.1% DMSO final concentration), simul-
365=460 nm, respectively).
taneously with LDLs. After incubation (14 h at 378C), the LDL-
containing medium was recovered for determining the TBARS
Monocyte adhesion assay and MCP1 release
and protein carbonyl content (16, 26). For all the other exper-
in the culture medium
iments, LDLs were mildly oxidized by UV=copper (in the
absence of drug), as previously described (62). Under standard
Monocyte adhesion assay was performed according to
conditions, UV=Cu-oxLDLs contained 71–104 nmol lipid hy- Srinivasan et al. (51). Briefly, Raw 264.7 macrophages were
droperoxide=mg apoB and 6.4–9.7nmol TBARS=mg apoB.
labeled with Calcein AM (5 mM) for 10 min at 378C, rinsed
three times in PBS, and resuspended in PBS. Labeled fluo-
rescent monocytes were added (100,000=well) to HMEC-1
either unstimulated or stimulated (at the indicated times)
with oxLDL (200 mg=ml) in the presence or absence of the
different agents (5 mM), for 30 min. At the end, the cells were
carefully washed three times with PBS for removing unbound
monocytes, and then lysed in water by sonication; the fluo-
rescence of the cell lysate was immediately determined by
fluorometry (ex=em 490=525). MCP-1 released by cells was
measured by ELISA in the culture medium of control and
LDL-stimulated HMEC-1, with or without the tested agents,
according to the indications of the manufacturer (Pierce).
Protein carbonyl content and foam cell formation
The protein carbonyl content was determined using DNPH,
according to Ichihashi et al. (26). The formation of foam cells
was evaluated on Raw 264.7 macrophages incubated for 14 h
in serum-free RPMI containing UV-oxLDL (100 mg=ml) and
the agents at the indicated concentrations. After 14 h incuba-
tion, the LDL-containing culture medium was removed, cells
were fixed in 0.4% paraformaldehyde=PBS for 10 min, washed
in PBS, and stained with oil Red O, or fluorometrically quan-
tified by Nile Red, as described (16).
Extracellular and intracellular ROS determination
Evaluation of cytotoxicity, necrosis and apoptosis
Extracellular superoxide anion (O28ꢀ) generation was
measured as the superoxide dismutase (SOD)-inhibitable
Cytotoxicity was evaluated using the MTT (3-(4,5-
reduction of cytochrome C. HMEC-1 preincubated for 1 h dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) test.
with native LDL (100 mg=ml), CuSO4 (1 mM), and the mole- Necrosis and apoptosis were determined by counting the cells
cules tested at 5 mM), were incubated in 1 ml phenol red-free (inverted fluorescence microscope Fluovert FU; Leitz), after
RPMI containing cytochrome C (20 mM) with or without staining with 2 vital fluorescent dyes, 0.6 mM SYTO-13 (a
SOD (100 mU=l), for 30 min. The reduction of cytochrome C permeant DNA intercalating green-colored probe) and 1.5 mM
8ꢀ
specific of O2 (e.g., SOD inhibitable) was determined propidium iodide (a non permeant intercalating red probe)
spectrophotometrically at 550 nm (difference in absorbance (16). DEVDase activity (caspase 3) was determined using the
with or without SOD), according to McCord and Fridovich fluorogenic substrate Ac-DEVD-AMC (N-acetyl-Asp-Glu-
(35). Extracellular hydrogen peroxide (H2O2) generation was Val-Asp-7-amino-4-methylcoumarin) (40 mM), on cell lysates
determined as horseradish peroxidase (HRP)-dependent from HMEC-1 stimulated with oxLDLs (200 mg=ml) and the
quenching of scopoletin fluorescence according to Loschen et agents, as reported (16). After 30 min incubation at 378C, the
al. (32), modified as follows: cells were incubated at 378C in released fluorescent product AMC (aminomethylcoumarin)
1 ml of PBS containing 5 mM scopoletin and 30 nM HRP, and was determined by fluorometry (ex=em 351=430 nm).
the fluorescence of the medium was read after 0, 30, and
60 min (ex=em 358=448 nm).
Intracellular ROS generated in cells treated by oxLDL
Western blot analysis
(200 mg=ml) and the tested agents (1–5 mM) were evaluated by
Cells extracts were subjected to SDS-PAGE and revealed
measuring the oxidation of H2DCFDA-AM (for hydroperox- using the indicated primary and secondary antibodies con-
ides) and DHE (for O28ꢀ) (10, 18). The probes were added to jugated with horseradish peroxidase, and developed by ECL
the culture medium (5 mM final concentration) 30 min before (Amersham). Same protocol was applied on aortas from
the end of the experiment. At the end, the cells were washed apoEꢀ=ꢀ mice (16).
three times in PBS, and the fluorescence of the cell homoge-
Western blot was quantified using NIH ImageJ software
nate was measured as reported (ex=em 495=520 nm for (http:==rsbweb.nih.gov=ij=index.html).

2096
BOUGUERNE ET AL.
Animal studies
(45% of lesion reduction in apoEꢀ=ꢀ mice treated by
10 mg=kg=day hydralazine in drinking water during 16
weeks) (16). However, hydralazine exhibits only mild anti-
oxidant properties, and relatively high concentrations are re-
quired to prevent LDL oxidation and subsequent formation of
foam cells that accumulate in early atherosclerotic lesions. We
aimed to improve the antioxidant properties of hydralazine,
without altering its carbonyl scavenger properties. For this
purpose we designed and synthesized a hydralazine deriva-
tive, and we report its ability to prevent LDL oxidation and
oxLDL cytotoxicity, and to delay the formation of athero-
sclerotic lesions in apoEꢀ=ꢀ mice.
ApoE KO (apoEꢀ=ꢀ) mice were from Transgenic Alliance
(IFFA Credo; Les Oncins). Three groups of 20 male mice
(4 weeks old), housed under specific pathogen-free conditions,
were fed a standard diet supplemented (or not) with BVH or
hydralazine (10 mg=kg in drinking water), as reported in (16,
38). Mice were killed at age 20 weeks. Mouse experiments were
approved by the Committee on Research Animal Care of In-
serm U-858 Center. Plasma cholesterol, triglycerides, and gly-
cemia were determined immediately after sacrifice (AU 2700
Biochemistry Analyzer Olympus, CHU Rangueil Toulouse).
Quantification and characterization
of atherosclerotic lesions
BVH synthesis
BV was obtained through oxidative coupling reaction of 2
vanilline moieties. The reaction conditions were optimized
and the best results were obtained when performing the
coupling reaction in the presence of sodium persulfate and
catalytic amount iron sulfate in water at 508C (Fig. 1). The
reaction was followed by thin layer chromatography till
completion after 5 days, affording after work up BV with an
excellent yield (95%). BV was then allowed to react with two
equivalents of hydralazine hydrochloride in refluxing ethanol
for 6h30 affording pure chlorhydrated BVH as a precipitated
solid which was filtered off (87% yield).
The lesions were estimated as described by Bucciarelli et al.
(6). Hearts were washed in PBX, frozen on a cryostatmount
with OCT compound (Tissue-Tek), and stored at ꢀ808C. Se-
rial sections 10 mm thick of aortic sinus were stained with oil
Red O, counterstained with hematoxylin-eosin, and analyzed
for morphometric evaluation of the lesion size, using a com-
puterized Biocom morphometry system. The mean lesion size
in aortic sinus was expressed in mm2 ꢂ SEM.
Immunohistochemistry
Fixed frozen cryo-sections (10mm thick) of aortic sinus from
control and BVH-treated mice were incubated with the anti-
bodies, anti-CD3 (lymphocytes), anti-a-actin (SMC), anti-4-
HNE-adducts (lipid peroxidation), and anti-MCP1. After incu-
BVH copper chelating properties
Transition metals are strongly implicated in the reductive
bation with appropriate biotin-labeled antibody, sections were decomposition of lipid hydroperoxides, to give alkoxyl and
revealed using avidin–biotin horseradish peroxidase visuali- lipid peroxyl radicals during the propagation step. As some
zation system (Vectastain, ABC kit Elite; Vector Laboratories).
polyphenolic compounds block the oxidation induced by
transition metal ion-induced by chelating copper or iron ions
(36), we evaluated whether this mechanism was implicated in
the protective effect of BVH against LDL oxidation. To clarify
the ability of BVH to react with cupric ion, its interaction was
assessed by UV-vis spectroscopy. The wavelength spectra of
the complex formed between BVH (25 mM) and increased
concentration of copper (10–80 mM) revealed absorbance
peaks at 296 and 376 nm related to BVH alone (Fig. 2A). The
Statistical analysis
Data are given as mean ꢂ SEM. Estimates of statistical
significance were performed by analysis of variance (one-way
ANOVA, Tukey test, SigmaStat software).
Results
We recently reported that hydralazine is an efficient car- increase in Cu(II) concentration caused a linear decrease in
bonyl scavenger agent with potent antiatherogenic effect the absorbance at 296 and 376 nm and a linear increase in
FIG. 1. Chemical structures of bisvanillyl-hydralazone (BVH), hydralazine (Hdz), and bisvanilline (BV), and scheme of
BVH synthesis.

ANTIATHEROGENIC EFFECT OF BISVANILLYL-HYDRALAZONE
2097
FIG. 2. BVH
copper-dependent
prevents
low-
density lipoprotein (LDL)
oxidation. (A) Wavelength
spectra of the complex
formed
(25 mM) and increasing cop-
per concentrations (10–
between
BVH
80 mM). The absorbance of
the complex was measured
at 296 and 376 nm related to
BVH alone, and at 456 nm
(absorbance of the complex).
(B) Dose effect of BVH, Hdz,
and BV on LDL oxidation
(1 mg apoB=ml in phosphate-
buffered saline [PBS]) in-
duced by CuSO4 (5 mM, 2 h
incubation at 378C). (C) Ef-
fect of BVH (5 mM) on LDL
oxidation (1 mg=ml) incu-
bated for 2 h at 378C with
variable concentrations of
AAPH (2,20-azobis-(2-amidi-
no propane) hydrochloride)
(0.1–1 mM as indicated).
LDL oxidation was evalu-
ated by the thiobarbituric
acid-reactive
substance
(TBARS) content, as de-
scribed in the Materials and
Methods section, and the re-
sults were expressed as nmol
TBARS=mg apoB in LDL
oxidized in the absence of
antioxidant. In (B, C), the
data are meanꢂ SEM of four
separate experiments. *p< 0.05.
absorbance at 456 nm. Figure 2A also shows the formation of 2B, low concentrations of BVH (1–5 mM) prevented copper-
four isosbestic points at 274, 306, 330, and 415 nm in the mediated LDL oxidation (monitored by TBARS formation), in
presence of concentration of Cu(II) between 0 and 20 mM. The contrast to hydralazine and BV, which had no (or only minor)
isosbestic points at 306, 330, and 415 nm shifted, respectively, effect at these concentrations. BVH efficiently inhibited LDL
to 299, 323, and 402 nm in the presence of 23.33 to 80 mM oxidation mediated by AAPH (100 and 300 mM), indicating
Cu(II). Altogether, these results indicate the formation of that it also exhibits free radical scavenger properties in metal-
copper-BVH complexes.
independent oxidation systems (Fig. 2C). Likewise, BVH
inhibited cell-mediated LDL oxidation and subsequent cyto-
toxicity, whereas hydralazine and BV had no effect (Fig. 3A, B).
Since cell-mediated LDL oxidation depends on the secre-
tion of ROS in the extracellular medium, we evaluated the
BVH prevents metal- and AAPH-dependent
LDL oxidation
The antioxidant properties of BVH were evaluated on its ability of BVH to reduce the level of extracellular ROS,
ability to block LDL oxidation induced in vitro by metal- namely, superoxide anion (O28ꢀ) and hydrogen peroxide
dependent (copper) and -independent (AAPH) oxidation sys- (H2O2). BVH (1 and 5 mM) reduced the generation of O28ꢀ and
tems or by contact of LDL with HMEC-1. As shown on Figure H2O2 (Fig. 3C, D), in agreement with its inhibitory effect on

2098
BOUGUERNE ET AL.
FIG. 3. BVH inhibits cell-induced LDL
oxidation and the extracellular generation
of reactive oxygen species (ROS). (A) Cell-
mediated LDL oxidation was evaluated on
human microvascular endothelial cell-1
(HMEC-1) seeded in 12-multiwell plates. The
standard culture medium was removed and
replaced by serum-free RPMI 1640 contain-
ing native LDL (100 mg apoB=ml) and the
agents (BVH, Hdz, and BV) used at the in-
dicated concentrations. After 14 h incubation
at 378C, the TBARS content was determined
in the LDL-containing medium. Results are
expressed as percent of the control incubated
with LDLs in the absence of drug (the TBARS
content of the oxidized LDL [oxLDL]-
containing culture medium was 15 nmol=mg
apoB ꢂ 2.5). (B) Determination of HMEC-1
viability after 14 h of incubation in the pres-
ence of native LDLs and copper (1 mmol), as
indicated in (A). Cell viability was deter-
mined by the MTT test. Results are expressed
as percent of the untreated control. (C) De-
8ꢀ
termination of O2 generation by HMEC-1
incubated for 1 h with oxLDL (100 mg=ml) ꢂ
8ꢀ
the indicated agents (1 and 5 mM). O2 was
measured by the SOD-inhibitable reduction
of cytochrome C, as reported in the Materials
and Methods section. Results are expressed as
percent of the control treated with oxLDLs.
(D) Determination of H2O2 generation in the
extracellular medium of HMEC-1 incubated
for 30 min with oxLDL (100 mg=ml)ꢂ the in-
dicated agents (5mM). H2O2 was evaluated by
the HRP-dependent quenching of scopoletin
fluorescence (ex=em 358=448 nm), measured
as indicated in the Materials and Method
section. Results are expressed as percent of the
control treated with oxLDLs. (E) Left panel:
Expression of the gp91phox subunit of
NAD(P)H oxidase, evaluated by Western blot
experiments from HMEC-1 incubated for 5 h
with oxLDLs (100 mg=ml)ꢂ the agents (BVH,
Hdz, and BV), and revealed with a gp91phox
anti antibody and anti EGF-receptor (EGF-R)
(Santa-Cruz), as control. Right panel: densito-
metric analysis of gp91phox Western blot experiments by ImageJ software. Gp91phox=EGF-R ratio was calculated and expressed
in arbitray units (au), as meanꢂ SEM of four separate experiments. *p < 0.05 (values compared to the unstimulated control). In
(A–E), the data are meanꢂ SEM of four separate experiments. *p < 0.05. In (C, D), compared to the oxLDL-treated control, black
bar. The yellow bars represent the untreated control. (To see this illustration in color the reader is referred to the web version of
this article at www.liebertonline.com=ars).
lipid peroxidation and subsequent toxicity (Fig. 3A, B). It may involved in the generation of extracellular ROS in HMEC-1
be noted that hydralazine was less efficient than BVH (Fig. (11). We thus checked whether oxLDLs (preoxidized by UV
3A–D). BV did not exert any protective effect at the used irradiation=Cu in the absence of drug) increase the expression
concentrations.
of gp91phox at the plasma membrane, and whether these re-
The generation of extracellular ROS by NAD(P)H oxidase sponses are regulated by BVH and the other tested com-
and its implication in LDL oxidation and atherosclerosis have pounds. As shown in Figure 3E, oxLDLs increased the
been largely investigated (19, 20, 46). This functionally active expression of gp91phox (the data in Fig. 3E were observed after
NAD(P)H oxidase contains the membrane-bound subunits 5 h incubation). All these responses were inhibited by BVH
p22phox and gp91phox (20), and is required for the proliferation (and at a lesser extent by hydralazine) (both being tested at
and migration of these cells (20, 60). The expression of 5 mM), as well as by the metal chelator desferal, and by trolox
gp91phox is increased and regulated by oxidative stress (20, (data not shown). BV had no inhibitory effect. Thus, it can be
46), and is blocked by antioxidants such as resveratrol (8). speculated that BVH inhibits the generation of extracellular
Moreover, we previously reported that NAD(P)H oxidase is ROS via a direct mechanism including its radical scavenger

ANTIATHEROGENIC EFFECT OF BISVANILLYL-HYDRALAZONE
2099
FIG. 4. BVH inhibits the
intracellular oxidative stress
and the release of MCP-1.
(A, B) Rise of intracellular ROS
triggered by oxLDLs (200 mg=
ml, 3 h incubation) and inhibi-
tion by BVH, Hdz, and BV.
HMEC-1 were incubated with
or without oxLDLs (200 mg=ml
for 3 h) and the indicated
agent, used at 1 or 5 mM (A) or
at 5 mM (B). Thirty minutes
before the end of the experi-
ment, cells were loaded with
the ROS-sensitive fluorescent
probes
6-carboxy-20,70-
dichlorodihydrofluorescein dia-
cetate, di(acetoxymethyl ester)
(H2DCFDA-AM) (A) or dihy-
droethidium (DHE) (B) and
the fluorescence of the cells
was measured as indicated in
the Materials and Methods
section. (C) Activation of the
20S proteasome determined by
measuring the hydrolysis of
the sLLVY-MCA substrate, by
lysates from cells incubated for
2 h with 200 mg=ml of oxLDL
(as described in the Materials
and Methods section), and ef-
fect of BVH, Hdz, BV, and
trolox (Tx) as control (all were
tested at 5 mM, except trolox, 10 mM). (D) Upper panel: Western blot experiments showing the phosphorylation of NF-kB, upon 2 h
stimulation of HMEC-1 with toxic oxLDL concentrations (200 mg=ml), and effect of BVH, Hdz and BV (all the agents were tested at
5mM). Lower panel: densitometric analysis of phospho-NF-kB Western blot experiments by ImageJ software. Phospho-NF-kB=b-
actin ratio was calculated and expressed in arbitray units (au), as meanꢂ SEM of four separate experiments. *p< 0.05 (values
compared to the unstimulated control). (E) Monocytes adhesion to HMEC-1 activated by oxLDLs. HMEC-1 were grown on 12-
multiwell culture plaques, and stimulated by oxLDLs for 1–4 h. The adhesion assay was performed using Raw-267.4 macrophages
previously loaded with calcein (5 mM) for 10 min and washed three times in warm PBS. The fluorescent calcein-loaded macro-
phages (50,000 cellules=well) were added to HMEC-1 culture for 30 min, and then the cells were carefully washed with warm PBS
and lyzed in water, and the fluorescence of the mixture was immediately measured (ex=em 490=525). The results are expressed as
percentage of the unstimulated control. (F) MCP-1 release by HMEC-1 activated by oxLDLs. HMEC-1 were incubated with
oxLDLs (200 mg=ml) for 1 to 5 h, and MCP-1 released in the culture medium was determined by ELISA. The data are expressed as
percentage of the unstimulated control. In (A–F), the data are meansꢂ SEM of four separate experiments. *p< 0.05; compared to
the oxLDL-treated control, black bar. The yellow bars represent the untreated control. (To see this illustration in color the reader is
referred to the web version of this article at www.liebertonline.com=ars).
and metal chelator properties, and indirectly, by blocking the nuclear translocation of NF-kB depends on its phosphoryla-
increase in gp91phox expression.
tion, and on the proteasomal degradation of IkBa, the cyto-
We then investigated whether BVH is able to inhibit the solic constitutive inhibitor of NF-kB (44). The incubation of
intracellular ROS rise generated by oxLDLs, and involved in HMEC-1 with oxLDLs stimulated the in vitro hydrolysis of
vascular dysfunction, inflammatory response, and cytotoxic- sLLVY-MCA (a fluorogenic synthetic peptide substrate for
ity (1, 44, 45). For this purpose, HMEC-1 were loaded with 20S proteasome) (Fig. 4C) and the phosphorylation of NF-kB
two permeant ROS-sensitive probes, the H2DCFDA-AM (Fig. 4D) (the reported data were observed at 2 h), in agree-
probe that reacts with hydroperoxides, and DHE that is more ment with our previous report (61). These responses were
8ꢀ
specific for O2 (18). Incubation of HMEC-1 with toxic con- completely inhibited by BVH, whereas hydralazine and BV
centration of oxLDLs (200 mg=ml) resulted in increased level had no effect. The proinflammatory signaling resulting from
in cellular hydroperoxides and O28ꢀ, which were completely NF-kB activation is implicated in the expression of adhesion
prevented by BVH (1–5 mM) (Fig. 4A, B), whereas no or only molecules and the secretion of chemoattractants (9, 51, 56). To
very weak protective effects were obtained with similar con- check whether BVH inhibits the proinflammatory signaling of
centrations of hydralazine and BV.
oxLDLs, we tested its ability to inhibit the monocyte recruit-
Oxidative stress and oxLDLs trigger a proinflammatory ment to the endothelium activated by oxLDLs. As shown in
signaling, which results (at least in part) from the activation of Figure 4E, a significant increase in monocyte adhesion to the
the redox-sensitive NF-kB transcription factor (44). The endothelium was observed after incubation of cells with

2100
BOUGUERNE ET AL.
FIG. 5. BVH inhibits foam cell formation
and the carbonyl stress. (A, B) Inhibition of
foam cell formation by BVH. Raw264.7 mu-
rine macrophages were incubated for 14 h
with oxLDLs (100 mg=ml) and the tested
agents BVH, Hdz, or BV (each used at the final
concentration of 5 mM). In (A), the cells were
washed twice, fixed in 4% paraformaldehyde,
and stained with oil Red O, and observed
microscopically. Representative of three sepa-
rate experiments. In (B) Raw264.7 were
stained with Nile Red, scrapped, and the
fluorescence of the cells was measured as in-
dicated in the Materials and Methods section.
Results are expressed as percentage of the
control treated by oxLDLs (without added
drug, black bars). (C) Protein carbonyl content
increase induced by oxLDLs, and inhibition
by BVH. The protein carbonyl content was
determined on cell extracts after 14 h incuba-
tion of HMEC-1 with oxLDLs (200 mg=ml)
used either alone or in the presence of BVH,
Hdz, and BV (5 mM final concentration). Car-
bonylated proteins were determined spectro-
photometrically using DNPH as described in
the Materials and Methods section. The results
are expressed as percent of the control treated
by oxLDLs (without drug, black bars). (D)
Protein carbonyl content increase induced by
4-hydroxynonenal (4-HNE) (10 mM, 14 h) and
inhibitory effect of BVH and Hdz. These data
are meansꢂ SEM of four separate experiments
(*p < 0.05; compared to the oxLDL-treated
control, black bar). The yellow bars represent
the untreated control. (To see this illustra-
tion in color the reader is referred to the web
version of this article at www.liebertonline
.com=ars).
oxLDLs. This effect was prevented by BVH, but not by hy- 5C, BVH prevented the cellular increase in protein carbonyl
dralazine nor BV. Likewise, the secretion of MCP-1 was in- content induced by oxLDLs (200 mg=ml, 14 h incubation),
hibited by BVH, but not or only weakly by hydralazine or BV whereas hydralazine and BV were unefficient at the used
(Fig. 4F). Altogether, these data indicate that BVH exerts a concentration (5 mM). As shown in Figure 5D, 4-HNE (10 mM)
potent antioxidant activity able to prevent the rise of extra- directly added to the culture medium of HMEC-1 triggered an
and intracellular ROS. Moreover, BVH inhibits the inflam- increase in protein carbonyl content, which was blocked by
matory signaling mediated by oxLDLs, as assessed by the BVH and hydralazine, but not by BV. The protection against
inhibition of NF-kB, and the reduced expression of gp91phox
adhesion molecules, and chemoattractant factors.
,
the carbonyl stress induced by 4-HNE necessited higher BVH
and hydralazine concentrations (25 mM), whereas the antiox-
idant properties of BVH were observed from 1 to 5 mM. Taken
together, it is suggested that the inhibitory effect of BVH on
the carbonyl stress elicited by oxLDLs mainly results from its
antioxidant properties, whereas its intrinsic carbonyl scav-
enger properties are comparable with those exerted by
hydralazine.
BVH inhibit foam cell and carbonyl stress
induced by oxLDLs
The formation of foam cells results from the uptake by
macrophages of modified and oxLDLs (55). Foam cell for-
mation was evaluated in murine Raw 264.7 macrophages
incubated for 14 h with 100 mg=ml oxLDLs and the tested
agents (5 mM). The lipid accumulation was evidenced by oil
red O staining and was fluorometrically quantified by Nile
BVH prevents oxLDL-mediated apoptosis
The antiapoptotic properties of BVH, hydralazine, and
Red. Foam cell formation was strongly inhibited by BVH, but BV were evaluated on HMEC-1 incubated with toxic
not by BV, whereas hydralazine exhibited only weak protec- concentrations of oxLDLs. After 24 h incubation, the re-
tion (Fig. 5A, B) (higher concentrations being required for sidual viability of cells (evaluated by the MTT test)
preventing foam cell formation) (16).
was < 20% of the control (cells incubated without oxLDLs)
We then investigated whether BVH may block the modi- (Fig. 6A). Dose–response experiments indicated that BVH
fication by carbonylation of cellular proteins (53), which can was strongly efficient even at low concentrations (1 mM)
be protected by carbonyl scavengers (16). As shown in Figure for preventing the toxic effect of oxLDLs, whereas 5 mM

ANTIATHEROGENIC EFFECT OF BISVANILLYL-HYDRALAZONE
2101
FIG. 6. BVH prevents oxLDL-mediated
apoptotic signaling. (A–C) Cytoprotective
effect of BVH. Sub-confluent HMEC-1 were
incubated in serum-free culture medium with
oxLDLs (200mg apoB=ml) and the tested
agents. (A) Cytotoxicity and cytoprotection
were evaluated by the MTT assay, after 24h
incubation with oxLDLs and with=without the
tested agents (used at the indicated concentra-
tions). The results are expressed as percentage
of the untreated control. Fluorescence micros-
copy (B) and count of live and dead cells (C)
incubated for 16 h with=without oxLDLs
(200 mg=ml) and with=without BVH (5 mM),
Hdz (5 mM), BV (5 mM), and trolox (Trol,
10 mM). Cells were stained by the fluorescent
DNA intercalating agents Syto-13 (permeant,
green) and propidium iodide (nonpermeant,
red), and examined by fluorescence microscopy
that reveals live cells (green with normal nu-
cleus) and dead cells with apoptotic nucleus
(green apoptotic or red postapoptotic necrosis,
arrows) or primary necrosis with red nucleus
(without apoptotic features). In (B), represen-
tative of three separate experiments. (D) Cyto-
protective effect of BVH, Hdz, and BV (5 and
25 mM) on the toxicity of 4-HNE (10 mM, 14h
incubation), and evaluated by the MTT test as
percentage of the untreated control. (E, F) BVH
blocks the proapoptotic cascade elicited by
oxLDLs in HMEC-1 (similar experimental con-
ditions as in (C). (E) Western blots of cyto-
chrome C (cyto C) on cytosolic proteins, and
Bid and b-actin on total cell extracts. Re-
presentative of four separate experiments. (F).
Evaluation of the DEVDase activity on cell ex-
tracts using the fluorogenic substrate Ac-
DEVD-AMC, as indicated in Materials and
Methods section. The results are expressed as
percentage of the untreated controls. These data
are meansꢂ SEM of at least four separate ex-
periments. *p< 0.05 (comparison to the oxLDL-
treated controls, black bars). The yellow bars
represent the untreated control. (To see this
illustration in color the reader is referred to the
web version of this article at www.liebertonline
.com=ars).
hydralazine and BV had no protective effect (Fig. 6A). Cell The protective effect of BVH on 4-HNE-induced cell death
death was characterized by counting apoptotic cells after necessited higher concentrations (25 mM), and was compa-
double staining with two fluorescent intercalating agents, rable with that exerted by hydralazine (Fig. 6D). Trolox, an
the permeant DNA probe syto-13 that stains green all the antioxidant deriving from vitamin E, and used as control,
nuclei, and propidium iodide (PI) that stains red only cells blocked oxLDL-mediated apoptosis in agreement with
with permeabilized membrane (i.e., necrotic cells affected by previous studies (33), but had no effect on 4-HNE-induced
either primary necrosis or postapoptotic necrosis) (47, 62). cytotoxicity.
As shown in Figure 6B and C, oxLDLs increased the number
of apoptotic cells. BVH (1–5 mM) prevented the apoptosis of
BVH prevents the development
HMEC-1 cells, whereas hydralazine and BV (1–5 mM) were
of atherosclerosis in apoEꢀ=ꢀ mice
unefficient. BVH (5 mM) blocked the apoptotic signaling
elicited by oxLDLs (Fig. 6E, F), by inhibiting the cleavage of
The deletion of the apoE gene in apoEꢀ=ꢀ mice leads to
Bid and the cytosolic release of cytochrome C, as well as the severe hypercholesterolemia and spontaneous atherosclero-
activation of DEVDase (caspase 3). BVH and hydralazine sis, even on a standard chow diet (65). Three groups of 20
(25 mM), but not BV, blocked the carbonyl stress-induced apoEꢀ=ꢀ mice (male) received a standard diet chow, and were
toxicity, by inhibiting cell death elicited by pure 4-HNE either untreated (control) or treated for 16 weeks with BVH or
(10 mM) added to the culture medium of HMEC-1 (Fig. 6D). hydralazine added to the drinking water. Blood pressure,

2102
BOUGUERNE ET AL.
Table 1. Plasma Cholesterol, Triglyceride, and Glycemia Concentrations and Blood Pressure
in apoEꢀ=ꢀ Mice After 16 Weeks of Treatment with Bisvanillyl-Hydralazone and Hydralazine
ApoEꢀ=ꢀ mice treatment
Glucose (mM)
Total cholesterol (mM)
TG (mM)
BP (mmHg)
Untreated
BVH
Hydralazine
8.6 ꢂ 0.4a
8.2 ꢂ 0.3a
8.5 ꢂ 0.3a
13.8 ꢂ 1.6a
13.4 ꢂ 1.4a
13.7 ꢂ 1.5a
1.58 ꢂ 0.4a
1.76 ꢂ 0.3a
1.45 ꢂ 0.3a
108 ꢂ 5
110 ꢂ 6
107 ꢂ 4
Results are given as mean ꢂ SD.
ap < 0.05 for comparison between untreated and drug-treated apoEꢀ=ꢀ mice, 20 animals per group.
BVH, bisvanillyl-hydralazone; BP, blood pressure; TG, triglycerides.
plasmatic levels of total cholesterol, triglycerides, and glucose positive cells (macrophages) (data not shown). BVH also re-
were comparable in treated and untreated animals (Table 1), duced lymphocyte infiltration (as assessed by a decrease in
thereby indicating that neither BVH nor hydralazine modify CD3-positive T cells), and inhibited the accumulation of 4-
the blood level of circulating lipids. In untreated (control) HNE adducts on tissular proteins (carbonyl stress) in aortic
apoEꢀ=ꢀ mice, the area of atherosclerotic lesions in aortic sinus of apoEꢀ=ꢀ mice (Fig. 7B). These data were confirmed by
sinus was 24,550 ꢂ 6340 mm2, in agreement with (16). BVH the reduction by BVH (and hydralazine) of the tissular in-
administration resulted in a marked reduction (>55%) of crease in protein carbonyl content (Fig. 7C), and by the de-
the atherosclerotic lesion area (Fig. 7A). By comparison, hy- crease in 4-HNE-adduct accumulation on the PDGF receptor
dralazine reduced by 40% the size of the lesions, as previ- (a target of carbonyl stress) (13, 62), in aortas of treated versus
ously reported (16). In the light of the antioxidant and anti- untreated animals (Fig. 7D).
inflammatory properties of BVH, we checked the composition
of atherosclerotic lesions from BVH-treated and control ani-
Discussion
mals. As shown in Figure 7B, BVH reduced the tissular ex-
pression of the chemoattractant factor MCP-1, and the
Oxidative stress, LDL oxidation, and carbonyl stress are
number of a-actin-positive cells (SMC), as well as CD68- involved in the initiation of early atherosclerotic lesions, and
FIG. 7. Antiatherogenic effect of BVH on
ApoEꢀ=ꢀ mice lesions. ApoEꢀ=ꢀ male mice
aged 4 weeks (20 in each group) were fed a
regular mouse chow diet and treated or not by
BVH or Hdz (30 mg=l, i.e., 10 mg=kg=day added
to the drinking water) for 16 weeks. Hearts were
embedded in OCT (Tissue-Tek) and aortic sinus
were cut into 10 mm sections, stained with oil
Red O, and counterstained with hematoxylin–
eosin, and a morphometric quantification of the
lesion size was performed, as indicated in the
Materials and Methods section. (A) Fatty streak
area quantification. Means ꢂ SEM of fatty streak
area in aortic sinus of untreated male mice was
24,550 ꢂ 6340 mm2. (B) Immunohistochemistry
of cryosections of aortic sinus from untreated
(left) and BVH-treated (right) apoEꢀ=ꢀ mice.
From top to bottom, a-actin (SMC), CD3 (T
lymphocytes), MCP-1, and 4-HNE-adduct
(HNE) immunostaining. The pictures showed
here are representative of five separate un-
treated and BVH-treated mice. (C) Determina-
tion of the protein carbonyl content in protein
extracts from aortas of control untreated, BVH-,
and Hdz-treated mice. (D) Western blot exper-
iments of HNE-adduct accumulation in aortas
of control and BVH-treated mice. These results
are representative of three separate experi-
ments. In (A, C), the data are means ꢂ SEM of 20
mice in each group. *p< 0.05 (comparison to the
untreated controls, black bars). The yellow bars
represent the untreated control. (To see this
illustration in color the reader is referred to the
web version of this article at www.liebertonline
.com=ars).

ANTIATHEROGENIC EFFECT OF BISVANILLYL-HYDRALAZONE
2103
in the inflammatory and apoptotic events involved in the chemoattractant MCP-1 involved in mononuclear cell re-
progression of atherosclerosis (3, 4, 23, 25). Antioxidants cruitment (9, 44). Interestingly, BVH prevented oxLDL
prevent the development of fatty streaks in animals (3, 24, 37), cytotoxicity, by blocking the apoptotic signaling resulting
but their efficacy on advanced atherosclerotic lesions and on in the activation of the intrinsic mitochondrial apoptotic
cardiovascular events in humans is debated (3, 5, 63). Al- cascade (characterized by the cleavage of the proapoptotic
though the efficiency of carbonyl scavengers on LDL oxida- factor Bid, the release of cytochrome C, and the activation
tion (27) and on the late steps of atherosclerosis is not clarified, of caspase 3) (47). These data are in agreement with the
inhibiting both oxidative stress and carbonyl stress may rep- generally known cytoprotective properties of antioxidants
resent a new therapeutic strategy for atherosclerosis. Carbo- (vitamin E, Trolox, probucol, phenolic acids, and flavo-
nyl scavengers are usually poor antioxidants, whereas noids) (34, 37, 57). The antioxidant effect of BVH was not
antioxidants fail to neutralize carbonyl compounds once ad- reproduced by co-incubating BV and hydralazine, thereby
ducts are formed on proteins (40, 42). We recently reported suggesting that the effect of BVH results from a confor-
that hydrazine derivatives used in medicine inhibit the car- mational change in the chemical structure of its compo-
bonyl stress induced by oxLDL, and slow down the devel- nents, which confers to the molecule a better access to its
opment of atherosclerotic lesions in apoEꢀ=ꢀ mice (this being targets (ROS, copper), and its subsequent antioxidant and
correlated with a decrease in foam cell formation). Among cytoprotective properties.
these agents, the antihypertensive drug hydralazine is a car-
The second observation is that BVH exhibited carbonyl
bonyl scavenger agent with (moderate) antioxidant activity scavenger properties similar to those observed for hydral-
(16). To improve the antioxidant properties of hydralazine, we azine, and evidenced by the inhibitory effect of BVH
designed and synthesized BVH, a hydralazine derivative, on protein carbonylation and toxicity elicited by 4-HNE di-
exhibiting both antioxidant (presence of a phenolic moiety) rectly added to the culture medium of HMEC-1 (which
and carbonyl scavenger properties (presence of hydralazine). triggers a carbonyl stress independent of lipid peroxidation).
The linker between BV and hydralazine is an acid-labile hy- Trolox used as control cytoprotective antioxidant was un-
drazone group that was designed to undergo hydrolysis, al- able to block the carbonyl stress induced by 4-HNE, and this
lowing the release of hydralazine under physiological suggests that the carbonyl scavenger properties of BVH (on
conditions (28).
the direct formation of 4-HNE-adducts) are independent of
A first observation is that BVH is a potent antioxidant, its antioxidant activity. These effects of BVH necessite rela-
associating both metal chelator and radical scavenger tively higher concentrations (25 mM), and are comparable
properties. BVH (even at low concentrations) formed with (thus probably result from) hydralazine, which is an
complexes with copper, and significantly blocked the hydrazine derivative, and a strong carbonyl scavenger (45).
copper-induced and AAPH-induced LDL oxidation, One hypothesis is that BVH undergoes hydrolysis of its acid-
(which is metal-independent), as well as LDL oxidation labile hydrazone group, thereby releasing hydralazine at
mediated by vascular cells (HMEC-1), which depends on acidic pH under physiological conditions (28). In contrast,
the presence of transition metal. The later mimicks the the strong inhibitory effect of low BVH concentrations (1–
mechanism of LDL oxidation occurring in vivo in the 5 mM) on the formation of protein carbonyls and foam cell
vascular wall, since copper ions are present in the vessels, accumulation induced by oxLDLs likely resulted from its
and contribute in the generation of oxLDLs and athero- antioxidant and cytoprotective rather than carbonyl scav-
genesis (3, 25, 64). The metal chelator and radical scaven- enger properties. Nevertheless, it can be speculated that
ger properties of BVH allow to block the generation of BVH is protective through its antioxidant properties (inhi-
extracellular ROS that are involved in cell-mediated LDL bition of oxidative stress, inflammation, and apoptosis in-
8ꢀ
oxidation (O2 initiates the generation of H2O2 and OH8, duced by oxLDLs), and through its carbonyl scavenger
formed through the Fenton reaction in the presence of iron properties at least in the vascular wall of apoEꢀ=ꢀ mice,
or copper) (29). The copper-chelating properties of BVH where it prevented the accumulation of HNE-adducts on
should help to inhibit the Fenton reaction as assessed by PDGF-R, a known target of carbonyl stress (13, 62). The
the inhibition of H2O2 secretion in the culture medium. antiatherogenic effect of BVH in apoEꢀ=ꢀ mice was superior
These effects are reinforced by the inhibitory effect of BVH to that of hydralazine (55%–60% of lesion area decrease vs.
on the expression of the gp91phox subunit of NADPH- 40%–45% for hydralazine), perhaps because of its better
oxidase, in agreement with reports showing the role of antioxidant properties. BVH did not reduce the plasma lipid
gp91phox in extracellular ROS generation (19, 20). The in- content of apoEꢀ=ꢀ mice, but prevented in part the accu-
hibitory effect of BVH on gp91phox expression could result mulation of macrophagic foam cells, CD3 T-cell infiltration,
from an inhibition by BVH, of the redox sensitive proin- and MCP-1 expression. The protein carbonyl content and the
flammatory NF-kB transcription factor, which regulates accumulation of 4-HNE adducts on tissular proteins (such as
the expression of NAD(P)H oxidase (2, 17), and is acti- PDGFR) were reduced in BVH-treated apoEꢀ=ꢀ mice, indi-
vated by oxLDLs (1). These data are in agreement with the cating a good correlation of these parameters with the de-
classical inhibitory effect of antioxidants on the intracel- crease in fatty streaks accumulation and plaque size
lular oxidative stress, the proteasome activation (as re- reduction.
ported here with BVH), and the proteasomal degradation
In conclusion, BVH is a potent antiatherogenic agent, and
of I-kB, the natural cytosolic inhibitor of NF-kB (44). Be- represents the prototype of new antioxidant and carbonyl
sides, BVH inhibited the proinflammatory signaling me- scavenger molecules, which should be of interest for slowing
diated by oxLDLs (44, 47), leading to stress-mediated down the progression of early atherosclerotic lesions forma-
cellular responses, such as the adhesion of monocytes to tion, and more generally of pathologies associated with oxi-
HMEC-1 stimulated by oxLDL, and the secretion of the dative and carbonyl stress.

2104
BOUGUERNE ET AL.
Financial Support and Acknowledgments
ing of vascular cells trigger the generation of extracellular
reactive oxygen species and LDL oxidation. Free Radic Biol
Med 35: 1589–1598, 2003.
The authors acknowledge the financial support by
INSERM, CNRS, Universite´ Paul Sabatier, Fondation Coeur et
Arte`res (FCA-06T6), Fondation pour la Recherche Me´dicale
(DCV2007040927), and European COST-B35. Dr. Gallet
(SNCF Laboratory, Toulouse) is gratefully acknowledged for
providing human sera. The authors thank Mrs. Corinne Ber-
nis for her excellent technical assistance.
13. Escargueil-Blanc I, Salvayre R, Vacaresse N, Ju¨rgens G,
Darblade B, Arnal JF, Parthasarathy S, and Ne`gre-Salvayre
A. Mildly oxidized LDL induces activation of platelet-
derived growth factor beta-receptor pathway. Circulation
104: 1814–1821, 2001.
14. Esterbauer H, Schaur RJ, and Zollner H. Chemistry and
biochemistry of 4-hydroxy nonenal, malonaldehyde and
related aldehydes. Free Radic Biol Med 11: 81–128, 1991.
Benaissa Bouguerne is recipient of support from the
Ministe`re de l’Enseignement Supe´rieur et de la Recherche.
Nadji Belkheiri is recipient of a grant from Algerian 15. Fang YZ, Yang S, and Wu G. Free radicals, antioxidants, and
nutrition. Nutrition 18: 872–879, 2002.
government.
16. Galvani S, Coatrieux C, Elbaz M, Grazide MH, Thiers JC,
Parini A, Uchida K, Kamar N, Rostaing L, Baltas M, Salvayre
R, and Ne`gre-Salvayre A. Carbonyl scavenger and anti-
atherogenic effects of hydrazine derivatives. Free Radic Biol
Med 45: 1457–1467, 2008.
Author Disclosure Statement
No competing financial interests exist.
References
17. Gauss KA, Nelson-Overton LK, Siemsen DW, Gao Y, DeLeo
FR, and Quinn MT. Role of NF-kappaB in transcriptional
regulation of the phagocyte NADPH oxidase by tumor ne-
crosis factor-alpha. J Leukoc Biol 82: 729–741, 2007.
18. Georgiou CD, Papapostolo I, Patsoukis N, Tsegenidis T, and
Sideris T. An ultrasensitive fluorescent assay for the in vivo
quantification of superoxide radical in organisms. Anal Bio-
chem 347: 144–151, 2005.
19. Go¨rlach A, Brandes RP, Nguyen K, Amidi M, Dehghani F,
and Busse R. A gp91phox containing NADPH oxidase se-
lectively expressed in endothelial cells is a major source of
oxygen radical generation in the arterial wall. Circ Res 87:
26–32, 2000.
1. Andalibi A, Liao F, Imes S, Fogelman AM, and Lusis
AJ. Oxidized lipoproteins influence gene expression by
causing oxidative stress and activating the transcription
factor NF-kappa B. Biochem Soc Trans 21(Pt 3): 651–655,
1993.
2. Anrather J, Racchumi G, and Iadecola C. NF-kappaB regu-
lates phagocytic NADPH oxidase by inducing the expres-
sion of gp91phox. J Biol Chem 281: 5657–5667, 2006.
3. Aviram M. Review of human studies on oxidative damage
and antioxidant protection related to cardiovascular dis-
eases. Free Radic Res 33, Suppl: S85–S97, 2001.
4. Berliner JA and Heinecke JW. The role of oxidized lipopro-
teins in atherogenesis. Free Radic Biol Med 20: 707–727, 1996.
5. Brigelius-Flohe R, Kluth D, and Banning A. Is there a future
for antioxidants in atherogenesis? Mol Nutr Food Res 49:
1083–1089, 2005.
6. Bucciarelli LG, Wendt T, Qu W, Lu Y, Lalla E, Rong LL,
Goova MT, Moser B, Kislinger T, Lee DC, Kashyap Y, Stern
DM, and Schmidt AM. RAGE blockade stabilizes estab-
lished atherosclerosis in diabetic apolipoprotein E-null mice.
Circulation 106: 2827–2835, 2002.
20. Griendling KK, Sorescu D, and Ushio-Fukai M. NAD(P)H
oxidase: role in cardiovascular biology and disease. Circ Res
86: 494–501, 2000.
21. Grune T, Reinheckel T, and Davies KJA. Degradation of
oxidized proteins in K562 human hematopoietic cells by
proteasome. J Biol Chem 271: 15504–15509, 1996.
22. Harkewicz R, Hartvigsen K, Almazan F, Dennis EA,
Witztum JL, and Miller YI. Cholesteryl ester hydroperox-
ides are biologically active components of minimally oxi-
dized low density lipoprotein. J Biol Chem 283: 10241–10251,
2008.
7. Chait A and Heinecke JW. Lipoprotein modification: cellular
mechanisms. Curr Opin Lipidol 5: 365–370, 1994.
8. Chow SE, Hshu YC, Wang JS, and Chen JK. Resveratrol
attenuates oxLDL-stimulated NADPH oxidase activity and
protects endothelial cells from oxidative functional damages.
J Appl Physiol 102: 1520–1527, 2007.
9. Cominacini L, Garbin U, Pasini AF, Davoli A, Campagnola
M, Contessi GB, Pastorino AM, and Lo Cascio V. Anti-
oxidants inhibit the expression of intercellular cell adhesion
molecule-1 and vascular cell adhesion molecule-1 induced
by oxidized LDL on human umbilical vein endothelial cells.
Free Radic Biol Med 22: 117–127, 1997.
10. Cossarizza A, Ferraresi R, Troiano L, Roat E, Gibellini L,
Bertoncelli L, Nasi M, and Pinti M. Simultaneous analysis of
reactive oxygen species and reduced glutathione content in
living cells by polychromatic flow cytometry. Nat Protoc 4:
1790–1797, 2009.
11. de Souza JA, Vindis C, Ne`gre-Salvayre A, Rye KA, Cou-
turier M, Therond P, Chantepie S, Salvayre R, Chapman MJ,
and Kontush A. Small, dense HDL 3 particles attenuate
apoptosis in endothelial cells: pivotal role of apolipoprotein
A-I. J Cell Mol Med 14: 608–620, 2010.
23. Heinecke JW. Oxidants and antioxidants in the pathogenesis
of atherosclerosis: implications for the oxidized LDL hy-
pothesis. Atherosclerosis 141: 1–15, 1998.
24. Hishikawa K, Nakaki T, and Fujita T. Oral flavonoid sup-
plementation attenuates atherosclerosis development in
apolipoprotein E-deficient mice. Arterioscler Thromb Vasc Biol
25: 442–446, 2005.
25. Holvoet P and Collen D. Oxidized lipoproteins in athero-
sclerosis and thrombosis. FASEB J 8: 1279–1284, 1994.
26. Ichihashi K, Osawa T, Toyokuni S, and Uchida K. En-
dogenous formation of protein adducts with carcinogenic
aldehydes: implications for oxidative stress. J Biol Chem 276:
23903–23913, 2001.
27. Jedidi I, Therond P, Zarev S, Garde`s-Albert M, Legrand A,
Barouki R, Bonnefont-Rousselot D, and Aggerbeck M.
Paradoxical protective effect of aminoguanidine toward
LDL oxidation: inhibition of apolipoprotein B fragmenta-
tion without preventing its carbonylation. Mechanism of
action of aminoguanidine. Biochemistry 42: 11356–11365,
2003.
12. Duval C, Cantero AV, Auge N, Mabile L, Thiers JC, Negre-
Salvayre A, and Salvayre R. Proliferation and wound heal-
28. Kalia J and Raines RT. Hydrolytic stability of hydrazones
and oximes. Angew Chem 120: 7633–7636, 2008.

ANTIATHEROGENIC EFFECT OF BISVANILLYL-HYDRALAZONE
2105
29. Kritharides L, Jessup W, and Dean RT. Macrophages re-
quire both iron and copper to oxidize LDL in Hanks’ bal-
anced salt solution. Arch Biochem Biophys 323: 127–136,
doplasmic reticulum stress in vascular cells: prevention by
oxygen-regulated protein 150 expression. Circ Res 104: 328–
336, 2009.
1995.
49. Seccia M, Albano E, and Bellomo G. Suitability of chemical
in vitro models to investigate LDL oxidation: study with
different initiating conditions in native and alpha-
tocopherol-supplemented LDL. Clin Chem 43: 1436–1441,
1997.
30. Leonarduzzi G, Arkan MC, Basaga H, Chiarpotto E, Seva-
nian A, and Poli G. Lipid oxidation products in cell signal-
ing. Free Radic Biol Med 28: 1370–1378, 2000.
31. Libby P, Ridker PM, and Maseri A. Inflammation and ath-
erosclerosis. Circulation 105: 1135–1143, 2002.
50. Sies H. Oxidative stress: oxidants and antioxidants. Exp
Physiol 82: 291–295, 1997.
32. Loschen G, Azzi A, and Flobe A. Mitochondrial H2O2 for-
mation: relationship with energy conservation. FEBS Lett 33: 51. Srinivasan S, Hatley ME, Reilly KB, Danziger EC, and He-
84–87, 1973.
drick CC. Modulation of PPARalpha expression and in-
flammatory interleukin-6 production by chronic glucose
increases monocyte=endothelial adhesion. Arterioscler
Thromb Vasc Biol 24: 851–857, 2004.
33. Lusis AJ. Atherosclerosis. Nature 407: 233–241, 2000.
34. Mabile L, Fitoussi G, Periquet B, Schmitt A, Salvayre R, and
Ne`gre-Salvayre A. Alpha-tocopherol and trolox block the
early intracellular events (TBARS and calcium rises) elicited 52. Stadtman ER. Protein oxidation in aging and age-related
by oxidized low density lipoproteins in cultured endothelial
cells. Free Radic Biol Med 19: 177–187, 1995.
35. McCord JM and Fridovich I. Superoxide dismutase. An en-
zymic function for erythrocuprein (hemocuprein). J Biol
Chem 244: 6049–6055, 1969.
diseases. Ann N Y Acad Sci 928: 22–38, 2001.
53. Steffen Y, Jung T, Klotz LO, Schewe T, Grune T, and Sies H.
Protein modification elicited by oxidized low-density lipo-
protein (LDL) in endothelial cells: protection by (-)-epica-
techin. Free Radic Biol Med 42: 955–970, 2007.
36. Mira L, Fernandez MT, Santos M, Rocha R, Floreˆncio MH, 54. Steinberg D. Low density lipoprotein oxidation and its pa-
and Jennings KR. Interactions of flavonoids with iron and
copper ions: a mechanism for their antioxidant activity. Free
Radic Res 36: 1199–1208, 2002.
thobiological significance. J Biol Chem 272: 20963–20966,
1997.
55. Steinbrecher UP. Receptors for oxidized low density lipo-
protein. Biochim Biophys Acta 1436: 279–298, 1999.
37. Moghadasian MH, McManus BM, Godin DV, Rodrigues B,
and Frolich JJ. Proatherogenic and antiatherogenic effects of 56. Tedgui A and Mallat Z. Cytokines in atherosclerosis: path-
probucol and phytosterols in apolipoprotein E-deficient
ogenic and regulatory pathways. Physiol Rev 86: 515–581,
mice. Circulation 99: 1733–1739, 1999.
2006.
38. Munzel T, Kurz S, Rajagopalan S, Thoenes M, Berrington 57. Traber MG and Atkinson J. Vitamin E, antioxidant and
WR, Thompson JA, Freeman BA, and Harrison DG. Hy- nothing more. Free Radic Biol Med 43: 4–15, 2007.
dralazine prevents nitroglycerin tolerance by inhibiting ac- 58. Uchida K, Kanematsu M, Morimitsu Y, Morimitsu Y, Osawa
tivation of a membrane-bound NADH oxidase. A new
action for an old drug. J Clin Invest 98: 1465–1470, 1996.
39. Napoli C. Oxidation of LDL, atherogenesis, and apoptosis.
Ann N Y Acad Sci 1010: 698–709, 2003.
T, Noguchi N, and Niki E. Acrolein is a product of lipid
peroxidation reaction. Formation of free acrolein and its
conjugate with lysine residues in oxidized LDLs. J Biol Chem
273: 16058–16066, 1998.
40. Negre-Salvayre A, Coatrieux C, Ingueneau C, and Salvayre 59. Uchida K. Role of reactive aldehyde in cardiovascular dis-
R. Advanced lipid peroxidation end products in oxidative eases. Free Radic Biol Med 28: 1685–1696, 2000.
damage to proteins. Potential role in diseases and thera- 60. Ushio-Fukai M, Tang Y, Fukai T, Dikalov SI, Ma Y, Fujimoto
peutic prospects for the inhibitors. Br J Pharmacol 153: 6–20,
2008.
41. Perrin DD and Armarego WLF. Purification of Laboratory
Chemicals, third edition. Oxford: Pergamon Press, 1988.
M, Quinn MT, Pagano PJ, Johnson C, and Alexander RW.
Novel role of gp91phox-containing NAD(P)H oxidase in
vascular endothelial growth factor–induced signaling and
angiogenesis. Circ Res 91: 1160, 2002.
42. Peyroux J and Sternberg M. Advanced glycation end- 61. Vieira O, Escargueil-Blanc I, Jurgens G, Borner C, Almeida L,
products (AGEs): pharmacological inhibition in diabetes.
Pathol Biol (Paris) 54: 405–419, 2000.
43. Rader DJ and Pure´ E. Lipoproteins, macrophage function,
and atherosclerosis: beyond the foam cell? Cell Metab 1: 223–
230, 2005.
44. Robbesyn F, Salvayre R, and Negre-Salvayre A. Dual role of
oxidized LDL on the NF-kappaB signaling pathway. Free
Radic Res 38: 541–551, 2004.
45. Rollas S and Ku¨c¸u¨kgu¨zel SG. Biological activities of hy-
drazone derivatives. Molecules 12: 1910–1939, 2007.
Salvayre R, and Ne`gre-Salvayre A. Oxidized LDLs alter the
activity of the ubiquitin-proteasome pathway: potential role
in oxidized LDL-induced apoptosis. FASEB J 14: 532–542,
2000.
62. Vindis C, Escargueil-Blanc I, Elbaz M, Marcheix B, Grazide
MH, Uchida K, Salvayre R, and Ne`gre-Salvayre A. De-
sensitization of platelet-derived growth factor receptor-beta
by oxidized lipids in vascular cells and atherosclerotic le-
sions: prevention by aldehyde scavengers. Circ Res 98: 785–
792, 2006.
46. Rueckschloss U, Duerrschmidt N, and Morawietz H. 63. Willcox BJ, Curb JD, and Rodriguez BL. Antioxidants in
NADPH oxidase in endothelial cells: impact on atheroscle-
cardiovascular health and disease: key lessons from epide-
rosis. Antioxid Redox Signal 5: 171–180, 2003.
miologic studies. Am J Cardiol 101: 75D–86D, 2008.
47. Salvayre R, Auge N, Benoist H, and Negre-Salvayre A. 64. Witztum JL and Steinberg D. Role of oxidized LDL in ath-
Oxidized low-density lipoprotein-induced apoptosis. Bio-
chim Biophys Acta 1585: 213–221, 2002.
48. Sanson M, Auge´ N, Vindis C, Muller C, Bando Y, Thiers JC,
Marachet MA, Zarkovic K, Sawa Y, Salvayre R, and Ne`gre-
Salvayre A. Oxidized low-density lipoproteins trigger en-
erogenesis. J Clin Invest 88: 1785–1792, 1991.
65. Zadelaar S, Kleemann R, Verschuren L, de Vries-Van der
Weij J, van der Hoorn J, Princen HM, and Kooistra T. Mouse
models for atherosclerosis and pharmaceutical modifiers.
Arterioscler Thromb Vasc Biol 27: 1706–1721, 2007.

2106
BOUGUERNE ET AL.
Address correspondence to:
Dr. Anne Ne`gre-Salvayre
Biochimie
Abbreviations Used
4-HNE ¼ 4-hydroxynonenal
BV ¼ bisvanilline
INSERM U858
CHU Rangueil
1 Avenue Jean Poulhe`s
BP84225
31432 Toulouse Cedex 4
France
BVH ¼ bisvanillyl-hydralazone
DHE ¼ dihydroethidium
H2DCFDA-AM ¼ 6-carboxy-70 dichlorodihydrofluorescein
diacetate, di(acetoxymethyl ester)
HMEC-1 ¼ human microvascular endothelial cell-1
IR ¼ infrared
LDL ¼ low-density lipoprotein
mp ¼ melting points
E-mail: anne.negre-salvayre@inserm.fr
OxLDL ¼ oxidized LDL
PBS ¼ phosphate-buffered saline
ROS ¼ reactive oxygen species
SOD ¼ superoxide dismutase
TBARS ¼ thiobarbituric reactive substances
Date of first submission to ARS Central, May 21, 2010; date of
final revised submission, October 19, 2010; date of acceptance,
November 2, 2010.