Inhibition of 6-hydroxydopamine-induced oxidative damage by 4,5-dihydro-3H-2-benzazepine N-oxides

Article abstract of DOI:10.1016/j.bcp.2007.12.010

A number of new analogs of 3,3-dimethyl-4,5-dihydro-3H-2-benzazepine 2-oxide, structurally related to the nitrone spin trap α-phenyl-N-tert-butylnitrone (PBN), were synthesized and evaluated for their activity in vitro as protectants against oxidative str

Full text of DOI:10.1016/j.bcp.2007.12.010

b i o c h e m i c a l p h a r m a c o l o g y 7 5 ( 2 0 0 8 ) 1 5 2 6 – 1 5 3 7
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/biochempharm
Inhibition of 6-hydroxydopamine-induced oxidative
damage by 4,5-dihydro-3H-2-benzazepine N-oxides
a
a
a
´
´
´
´
´
´
Ramon Soto-Otero , Estefanıa Mendez-Alvarez , Sofıa Sanchez-Iglesias ,
Fedor I. Zubkov b, Leonid G. Voskressensky b, Alexey V. Varlamov b,
Modesto de Candia c, Cosimo Altomare c,
*
a Departamento de Bioqu´ımica y Biolog´ıa Molecular, Facultad de Medicina, Universidad de Santiago de Compostela,
Santiago de Compostela, Spain
b Organic Chemistry Department of the Russian Peoples Friendship University, Moscow, Russia
c Dipartimento Farmaco-chimico, Facolta` di Farmacia, Universita` degli Studi di Bari, Via E. Orabona 4, I-70125 Bari, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
A number of new analogs of 3,3-dimethyl-4,5-dihydro-3H-2-benzazepine 2-oxide, structu-
rally related to the nitrone spin trap a-phenyl-N-tert-butylnitrone (PBN), were synthesized
and evaluated for their activity in vitro as protectants against oxidative stress induced in rat
brain mitochondria by 6-hydroxydopamine (6-OHDA), a neurotoxin producing experimental
model of Parkinson’s disease (PD). As assessed by a fluorimetric assay, all 2-benzazepine-
based nitrones were shown to decrease hydroxyl radicals (ꢀOH) generated during 6-OHDA
autoxidation. The inhibition effects on the ꢀOH formation shown by the 5-gem-dimethyl
derivatives, 2–4 times higher than those of the corresponding 5-methyl derivatives, were
attributed to the flattening effect of the 5-gem-dimethyl group on the azepine ring, which
should enhance nitrone reactivity and/or increase stability of the radical adducts. In
contrast, owing to steric hindrance, a methyl group to C-1 diminishes the ꢀOH-scavenging
activity of the nitrone group. All the assayed compounds were more potent than PBN as
inhibitors of 6-OHDA-induced lipid peroxidation (LPO) and protein carbonylation (PCO),
taken as an indicator of mitochondrial protein oxidative damage. The most promising
antioxidant (compound 11), bearing 5-gem-dimethyl and spiro C-3 cyclohexyl groups, high-
lighted in this study as the best features, inhibited LPO and PCO with IC50 values of 20 and
48 mM, respectively, showing a potency improvement over PBN of two order magnitude.
Both LPO and PCO inhibition potency data were found primarily related to the ꢀOH-
scavenging activities, whereas lipophilicity plays a role in improving the LPO (but not
PCO) inhibition, as a statistically valuable two-parameter equation proved.
Received 10 October 2007
Accepted 31 December 2007
Keywords:
2-Benzazepine nitrones
Neuroprotection
Radical scavengers
6-Hydroxydopamine
Lipid peroxidation
Protein carbonylation
# 2008 Elsevier Inc. All rights reserved.
* Corresponding author. Tel.: +39 080 5442781; fax: +39 080 5442230.
E-mail address: altomare@farmchim.uniba.it (C. Altomare).
Abbreviations: PD, Parkinson’s disease; AD, Alzheimer’s disease; ALS, amyotrophic lateral sclerosis; 6-OHDA, 6-hydroxydopamine;
MPTP, N-methyl-4-phenyl-1,2,3,6-tetrahydropyridine; PBN, a-phenyl-N-tert-butylnitrone; ROS, reactive oxygen species; ꢀOH, hydroxyl
radical; LPO, lipid peroxidation; RCS, reactive carbonyl species; PCO, protein carbonylation; THA, terephthalic acid; TBA, thiobarbituric
acid; TBARS, thiobarbituric acid reactive substances; TCA, trichloroacetic acid; MDA, malondialdehyde; MAO, monoamine oxidase; BSA,
bovine serum albumin; SDS, sodium dodecylsulfate.
0006-2952/$ – see front matter # 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.bcp.2007.12.010

b i o c h e m i c a l p h a r m a c o l o g y 7 5 ( 2 0 0 8 ) 1 5 2 6 – 1 5 3 7
1527
1.
Introduction
structure. Protein carbonylation (PCO) is an irreversible
oxidative damage, that often leads to formation of high-
molecular-weight aggregates, which are resistant to degra-
dation and accumulated as damaged or unfolded proteins
[21]. PCO can take place through different oxidative path-
ways. ROS can react (i) directly with the protein, or (ii) with
molecules, such as sugars and lipids, generating in turn
reactive carbonyl species (RCS), which then react with
protein. RCS, generated by peroxidation of polyunsaturated
fatty acids, induce PCO, playing a major role in the etiology
and/or progression of several human illnesses, including
neurodegenerative disorders associated with the deposition
of protein aggregates in tissues [22,23]. Some proteins are
more susceptible than others to oxidative stress. For example,
human brain copper–zinc superoxide dismutase (SOD1) has
been proven to be a major target of oxidative damage in PD
and AD [24].
Free radical-induced oxidation of cell macromolecules (lipids,
proteins, DNA, etc.) is implicated in several human pathogenic
processes, including cardiovascular diseases (e.g., stroke,
atherosclerosis) [1] and inflammatory (e.g., rheumatoid
arthritis) [2]. Over the last two decades, significant accumulat-
ing evidence has also shown that generation of reactive
oxygen species (ROS) and oxidative damage in the central
nervous system (CNS) are major events occurring in Parkin-
son’s disease (PD) [3] and other neurodegenerative disorders,
such as Alzheimer’s disease (AD) and amyotrophic lateral
sclerosis (ALS), as well as in aging processes [4]. Despite a large
body of biochemical data gathered from human brain autopsy
studies, the cause of neurodegeneration in PD has not yet been
completely established. Nevertheless, animal models, based
on the toxic damage induced by the neurotoxins 6-hydro-
xydopamine (6-OHDA) [5], N-methyl-4-phenyl-1,2,3,6-tetra-
hydropyridine (MPTP) [6] and related compounds [7,8], have
shown that iron-dependent oxidative stress, increased levels
of iron and monoamine oxidase (MAO)-B activity and deple-
tion of antioxidants (e.g., glutathione, GSH) play major roles in
PD. Indeed, antioxidants, monoamine oxidase (MAO)-B inhi-
bitors and iron chelators, have shown neuroprotective effects
in animal models either in vitro and in vivo [9–13], but
unfortunately they failed in the PD human therapy [3] when
administered as single drugs.
In the last decade, nitrone spin traps, a-phenyl-N-tert-
butylnitrone (PBN) and structurally related compounds (Fig. 1),
which react covalently with short-lived free radicals, such as
ꢀOH, have shown efficacy in a variety of animal models of CNS
injury [25–27], proving to be effective in ischemic stroke
models (especially the PBN disodium 1,3-disulfonate salt
derivative, NXY-059) [28] and aging [29,30]. Incorporating the
nitrone functionality in a ring (cyclic PBN analogs) did strongly
increase potency relative to PBN against oxidative injury and
cell toxicity, as a result of favorable changes in the HOMO/
LUMO energy levels and greater accessibility of the nitrone
double bond [26,31–34]. A number of 3,3-dimethyl-3,4-dihy-
droisoquinoline 2-oxides (1, Fig. 1), 3,3-dimethyl-4,5-dihydro-
3H-2-benzazepine 2-oxides (2), and their spiro C-3 cycloalkyl
analogs, have been found out as inhibitors of LPO. Electron
spin resonance (ESR) spectroscopy has clearly demonstrated
that cyclic nitrones 1 and 2 are able to trap radicals, forming
more stable spin adducts than PBN [33,34]. In particular,
chloro-substituted 2-benzazepine nitrones (2, X = 8-Cl or 7,9-
(Cl)2) proved to be potent radical scavengers in both lipid and
apoprotein fractions of low density lipoproteins [33]. More-
over, structure–activity relationship (SAR) studies have high-
lighted the following main factors increasing their in vitro
inhibitory potency: (i) the presence of a second antioxidant
functionality, such as o,o0-dimethylphenol (e.g., X = 6,8-(CH3)2-
7-OH in 2); (ii) lipophilicity increments, as resulted from
expansion of the nitrone-containing ring (2-benzazepine
versus isoquinoline), replacement of the gem-dimethyl group
6-OHDA, used in this and previous studies [14] for inducing
oxidative stress associated with the PD-related dopaminergic
neuron loss, is readily autoxidized and oxidatively deami-
nated by MAO, yielding hydrogen peroxide (H2O2) and the
corresponding p-quinone. H2O2 may generate the most
reactive ROS, that is hydroxyl radicals (ꢀOH) [15], and,
ultimately, lipid-derived carbon- and oxygen-centered radi-
cals, as lipid peroxidation (LPO) products. In turn, 6-OHDA
quinone triggers a cascade of oxidative reactions finally
resulting in the formation of an insoluble polymeric pigment
related to neuromelanin [16,17]. It has also been reported that
6-OHDA inhibits complexes I and IV of the mitochondrial
respiratory chain [18,19]. The oxidative stress caused by ꢀOH
generated during 6-OHDA autoxidation is suggested as a major
causal factor of its neurotoxicity [20].
Proteins are major targets for ROS and secondary by-
products of oxidative stress, and ROS-induced protein
modifications can lead to unfolding or alteration of protein
Fig. 1 – Structures of PBN spin trap and structurally related cyclic nitrones. The structures of 3,3-dimethyl-3,4-
dihydroisoquinoline 2-oxides (1), 3,3-dimethyl-4,5-dihydro-3H-2-benzazepine 2-oxides (2), and their spiro C-3 cycloalkyl
analogs, are shown, along with benzene ring X-substitution which proved to achieve significant improvements in activity
over the parent compounds (X = H).

1528
b i o c h e m i c a l p h a r m a c o l o g y 7 5 ( 2 0 0 8 ) 1 5 2 6 – 1 5 3 7
Fig. 2 – Scheme of synthesis and structures of the examined 2-benzazepine nitrones.
at the C-3 position with a spirocyclohexyl moiety and/or
hydrophobic X substituents on the fused benzene ring.
Recently, we have prepared a number of 2-benzazepine
nitrone congeners (compounds 3–12) through an efficient
synthesis (Fig. 2) [35] and studied their neuroprotective
properties, as assessed by the ability to scavenge ꢀOH,
generated during 6-OHDA autoxidation, and to inhibit 6-
OHDA-induced lipid peroxide formation (LPO), as assessed by
monitoring the formation of thiobarbituric acid reactive
substances (TBARS), and protein oxidative damage, as
indicated by the increase of protein carbonyl content, in rat
brain mitochondria. Herein we report the results from this
neurobiological investigation, followed by SAR to analyze the
data. Chiral compounds 3, 5, 8–10 and 12 were tested as
racemic mixtures, the investigation of enantioselective effects
being out of the scope of this work.
chemicals used were of analytical grade and were purchased
from Fluka Chemie AG (Buchs, Switzerland).
2.2.
Synthesis of 4,5-dihydro-3H-2-benzazepine N-oxide
derivatives
3,3-Dimethyl-4,5-dihydro-3H-2-benzazepine N-oxide deriva-
tives and their spiro C-3-cycloalkane analogs were prepared
through an efficient synthesis reported by some of us [35].
First, 2,3,4,5-tetrahydro-1H-2-benzazepines were easily
synthesized in two steps from readily available imines. They
were then converted, with moderate-to-high yields, into the 2-
benzazepine nitrones via hydrogen peroxide oxidation in the
presence of catalytic amounts of sodium tungstate, under
mild conditions. 5-Methyl-4,5-dihydrospiro[2-benzazepine-
3,10-cyclopentane] 2-oxide (9), 5-methyl-4,5-dihydrospiro[2-
benzazepine-3,10-cyclohexane] 2-oxide (10), and 1,5-dimethyl-
4,5-dihydrospiro[2-benzazepine-3,10-cyclohexane]
2-oxide
(12) have been reported earlier [35]. Compounds 4, 9 and 12
2.
Materials and methods
were prepared and tested as hydrochlorides.
2.1.
Chemicals
2.2.1. Homoallylamines
PBN, 6-OHDA hydrobromide, terephthalic acid, thiobarbituric
acid, butylated hydroxytoluene crystalline, 2,4-dinitrophenyl-
hydrazine hydrochloride, desferrioxamine, 1,1,3,3-tetraethox-
2-Methyl-N-(phenylmethyl)pent-4-en-2-amine (for 3), 2,4-di-
methyl-N-(phenylmethyl)pent-4-en-2-amine (for 4), 2-methyl-
N-(1-phenylethyl)pent-4-en-2-amine (for 5), 2,4-dimethyl-N-(1-
phenylethyl)pent-4-en-2-amine (for 6), N-(diphenylmethyl)-
2,4-dimethylpent-4-en-2-amine (for 7), 3,5-dimethyl-N-
(phenylmethyl)hex-5-en-3-amine (for 8), and 1-(2-methyl-
prop-2-en-1-yl)-N-(phenylmethyl)cyclohexanamine (for 11)
were obtained according to reported procedures [36–38]
from the corresponding imines and allyl(methallyl)magne-
siumbromide.
ypropane,
5,50-dithiobis-(2-nitrobenzoic
acid),
sodium
dodecylsulfate (SDS), EDTA, and bovine serum albumin
(BSA) were purchased from Sigma Chemical Co. (St. Louis,
MO, USA). Guanidine hydrochloride was purchased from
Aldrich Chemical Co. (Milwaukee, WI, USA). The water used
for the preparations of solutions was of 18.2 MV (Milli-RiOs/Q-
A10 grade, Millipore Corp., Bedford, MA, USA). All remaining

b i o c h e m i c a l p h a r m a c o l o g y 7 5 ( 2 0 0 8 ) 1 5 2 6 – 1 5 3 7
1529
2.2.2. 2,3,4,5-Tetrahydro-1H-2-benzazepine
2.2.3. Oxidation of benz-2-azepines to nitrones (3–8, 11):
Typical procedure
3,3,5-Trimethyl-2,3,4,5-tetrahydro-1H-2-benzazepine (for 3),
3,3,5,5-tetramethyl-2,3,4,5-tetrahydro-1H-2-benzazepine (for
4), 1,3,3,5-tetramethyl-2,3,4,5-tetrahydro-1H-2-benzazepine
(for 5), 1,3,3,5,5-pentamethyl-2,3,4,5-tetrahydro-1H-2-benza-
zepine (for 6), 3,3,5,5-tetramethyl-1-phenyl-2,3,4,5-tetrahy-
dro-1H-2-benzazepine (for 7), 3-ethyl-3,5,5-trimethyl-2,3,4,5-
tetrahydro-1H-2-benzazepine (for 8), and 5,5-dimethyl-1,2,4,5-
tetrahydrospiro[2-benzazepine-3,10-cyclohexane] (for 11)
were obtained according to already described methods
[31,39]. Selected spectral and physical data for 2-benzazepines
are given below.
To a solution of 50 mmol of benz-2-azepine and Na2WO4ꢁ2H2O
(0.83 g, 2.5 mmol) in acetone–water mixture (9:1, v/v, 100 mL),
50% H2O2 (12 mL, 200 mmol) was added dropwise at 0 8C
(0.5 h). The resulting mixture was stirred at room temperature
for 24–96 h (TLC monitoring). After the reaction was com-
pleted, the reaction mixture was poured into water (300 mL)
and extracted with CH2Cl2 (5 ꢂ 30 mL). The combined organic
extracts were dried over MgSO4 and concentrated in vacuo.
Crystallization of the resulting solids from hexane–ethyl
acetate mixture gave white crystals of nitrones (3–8, 11). In
case of compounds 3, 5, 8 the oily residues were purified by
column chromatography on Al2O3 (using hexane-ethyl acet-
ate, 5:1, v/v, as eluent). Slowly crystallizing nitrones (3, 6, 8, 11)
were converted into hydrochlorides by mixing their ether
solutions (1 g of nitrone in 50 mL of dry diethyl ether) with
saturated HCl/Et2O solution until pH 5 was achieved. Resulting
white precipitates were filtered-off, washed with absolute
ether and dried on air (90–95% yields).
2.2.2.1. 3,3,5,5-Tetramethyl-2,3,4,5-tetrahydro-1H-2-benzaze-
pine (for 4). Pale-yellow oil, 55% yield. bp 104–7 8C/2 torr; 1H
NMR (CDCl3, 400 MHz) d 1.22 (s, 6H), 1.41 (s, 6H), 1.43 (br. s, 1H),
1.81 (d, J = 14.9 Hz, 1H), 1.82 (d, J = 14.9 Hz, 1H), 4.03 (s, 2H), 7.01
(dd, J = 7.6, 2.1 Hz, 1H), 7.06 (dt, J = 7.6, 2.6 Hz, 1H), 7.17 (dt,
J = 7.6, 2.1 Hz, 1H), 7.34 (dd, J = 7.6, 2.6 Hz, 1H) ppm; ESI–MS m/z:
204 (MH)+. Anal. Calcd. for C14H21N: C, 82.70; H, 10.41; N, 6.89.
Found: C, 82.76; H, 10.34; N, 6.90.
2.2.3.1. 3,3,5-Trimethyl-4,5-dihydro-3H-2-benzazepine 2-oxide
(3). White crystals, 75% yield. mp 71–73 8C; 1H NMR (CDCl3,
400 MHz) d 1.43 (d, J = 7.2 Hz, 3H), 1.53 (s, 3H), 1.65(s, 3H), 1.99
(dd, J = 1.9, 15.0 Hz, 1H), 2.12 (dd, J = 9.4, 15.0 Hz, 1H), 3.13 (m,
1H), 7.14–7.30 (m, 4H), 7.91 (s, 1H) ppm; EI–MS (70 eV) m/z (rel.
intensity): 203 (32, M+), 144 (16), 132 (100), 131 (26), 130 (28), 129
(16), 128 (19), 115 (29), 104 (37), 77 (22); IR (KBr, n cmꢃ1) 1529
(C N). Anal. Calcd. for C13H17NO: C, 76.81; H, 8.43; N, 6.89.
Found: C, 76.65; H, 8.51; N, 6.81.
2.2.2.2. 1,3,3,5-Tetramethyl-2,3,4,5-tetrahydro-1H-2-benzaze-
pine (for 5). Colourless oil, 42% yield. bp 100–1 8C/3 torr;
isomer ratio 1:1.5, 1H NMR (CDCl3, 400 MHz), major isomer: d
1.06 (s, 3H), 1.26 (dd, J = 13.9, 11.5 Hz, 1H), 1.39 (s, 3H), 1.40 (d,
J = 7.3 Hz, 3H), 1.54 (d, J = 6.8 Hz, 3H), 1.60 (d, J = 13.9 Hz, 1H),
3.37 (dq, J = 11.5, 7.3 Hz, 1H), 7.19–7.34 (m, 4H) ppm; minor
isomer: d 1.15 (s, 3H), 1.17 (s, 3H), 1.41 (d, J = 7.3 Hz, 3H), 1.53 (d,
J = 6.8 Hz, 3H), 1.74 (dd, J = 13.9, 10.3 Hz, 1H), 1.83 (dd, J = 13.9,
3.7 Hz, 1H), 3.43 (m, J = 10.3, 7.3, 3.7 Hz, 1H), 4.41 (q, J = 6.8 Hz,
1H), 7.19-7.34 (m, 4H) ppm; ESI–MS m/z: 204 (MH)+. Anal. Calcd.
for C14H21N: C, 82.70; H, 10.41; N, 6.89. Found: C, 82.66; H, 10.39;
N, 6.85.
2.2.3.2. 3,3,5,5-Tetramethyl-4,5-dihydro-3H-2-benzazepine 2-
oxide (4). White crystals, 55% yield. mp 51–2 8C; 1H NMR
(CDCl3, 400 MHz) d 1.49 (s, 3H), 1.72 (s, 3H), 2.29 (s, 2H), 7.44 (td,
J = 7.5, 1.2 Hz, 1H), 7.52 (bd, J = 7.5 Hz, 1H),7.62–7.71 (m, 2H), 9.09
(s, 1H) ppm; EI–MS (70 eV) m/z (rel. intensity): 217 (35, M+), 161
(22), 146 (39), 145 (29), 144 (55), 143 (16), 130 (14), 129 (28), 128
(100), 115 (31); IR (KBr, n cmꢃ1) 1545 (C N). Anal. Calcd. for
C14H19NO: C, 77.38; H, 8.81; N, 6.45. Found: C, 77.42; H, 8.85; N,
6.66.
2.2.2.3. 1,3,3,5,5-Pentamethyl-2,3,4,5-tetrahydro-1H-2-benza-
zepine (for 6). Pale-yellow oil, 50% yield. bp 103–4 8C/3 torr; 1H
NMR (CDCl3, 400 MHz) d 1.24 (s, 3H), 1.25 (s, 3H), 1.41 (s, 3H), 1.49
(s, 3H), 1.51 (d, J = 6.7 Hz, 3H), 1.60 (d, J = 15.0 Hz, 1H), 2.12 (d,
J = 15.0 Hz, 1H), 4.46 (d, J = 6.7 Hz, 1H), 7.15 (dt, J = 7.3, 1.6 Hz,
1H), 7.19 (dt, J = 7.3, 1.6 Hz, 1H), 7.27 (dd, J = 7.3, 1.6 Hz, 1H), 7.38
(dd, J = 7.3, 1.6 Hz, 1H) ppm; ESI–MS m/z: 218 (MH)+. Anal. Calcd.
for C15H23N: C, 82.89; H, 10.67; N, 6.44. Found: C, 82.95; H, 10.60;
N, 6.45.
2.2.3.3. 1,3,3,5-Tetramethyl-4,5-dihydro-3H-2-benzazepine 2-
oxide (5). White crystals, 66% yield. mp 69–71 8C; 1H NMR
(CDCl3, 400 MHz) d 0.82 (s, 3H), 1.23 (d, J = 6.8 Hz, 3H), 1.30 (s,
3H), 1.86 (dd, J = 12.2, 14.0 Hz, 1H), 2.30 (s, 3H), 2.45(dd, J = 6.8,
14.0 Hz, 1H), 2.98 (m, 1H), 7.27–7.31 (m, 4H) ppm; EI–MS (70 eV)
m/z (rel. intensity): 217 (48, M+), 175 (22), 160 (23), 158 (40), 146
2.2.2.4. 3-Ethyl-3,5,5-trimethyl-2,3,4,5-tetrahydro-1H-2-benza-
zepine (for 8). Pale-yellow oil, 66% yield. bp 131–2 8C/5 torr; 1H
NMR (CDCl3, 400 MHz) d 0.93 (t, J = 7.3 Hz, 3H), 1.16 (s, 3H), 1.42 (s,
3H), 1.46 (s, 3H), 1.54 (m, 2H), 1.74 (d, J = 15.1 Hz, 1H), 1.86 (d,
J = 15.1 Hz, 1H), 4.05 (s, 1H), 7.04 (dd, J = 7.5, 0.9 Hz, 1H), 7.20 (t,
J = 7.5, 1.4 Hz, 1H), 7.37 (dd, J = 7.5, 1.4 Hz, 1H), 7.39 (td, J = 7.5,
0.9 Hz, 1H) ppm;ESI–MSm/z: 218 (MH)+. Anal. Calcd. for C15H23N:
C, 82.89; H, 10.67; N, 6.44. Found: C, 82.92; H, 10.69; N, 6.45.
(100), 145 (54), 144 (56), 143 (31), 128 (26), 91 (29); IR (KBr, n cmꢃ1
)
1542 (C N). Anal. Calcd. for C14H19NO: C, 77.38; H, 8.81; N, 6.45.
Found: C, 77.52; H, 8.75; N, 6.43.
2.2.3.4. 1,3,3,5,5-Pentamethyl-4,5-dihydro-3H-2-benzazepine
2-oxide (6). White crystals (hydrochloride), 70% yield. mp 165–
73 8C; 1H NMR (CDCl3, 400 MHz) d 1.21 (s, 6H), 1.30 (s, 6H), 2.25 (s,
2H), 2.44 (s, 3H), 7.21–7.30 (m, 1H), 7.28–7.33 (m, 2H), 7.35–7.38
(m, 1H) ppm; EI–MS (70 eV) m/z (rel. intensity): 231 (24, M+), 160
(59), 159 (22), 158 (100), 157 (19), 144 (18), 143 (20), 129 (19), 128
(29), 115 (21); IR (KBr, n cmꢃ1) 1555 (C N). Anal. Calcd. for
C15H21NO: C, 77.88; H, 9.15; N, 6.05. Found: C, 77.81; H, 9.30; N,
6.18.
2.2.2.5. 5,5-Dimethyl-1,2,4,5-tetrahydrospiro[2-benzazepine-
3,10-cyclohexane] (for 11). Viscous yellow oil, 85% yield. bp
139–42 8C/2 torr; 1H NMR (CDCl3, 400 MHz) d 1.42 (s, 6H), 1.25–
1.75 (m, 10H), 1.79 (s, 2H), 4.01 (s, 2H), 6.89–7.35 (m, 4H) ppm;
ESI–MS m/z: 244 (MH)+. Anal. Calcd. for C17H25N: C, 83.95; H,
10.29; N, 5.76. Found: C, 84.24; H, 9.95; N, 5.81.

1530
b i o c h e m i c a l p h a r m a c o l o g y 7 5 ( 2 0 0 8 ) 1 5 2 6 – 1 5 3 7
Table 1 – IC50 values for inhibition of ꢀOH production
during 6-OHDA autoxidation
2.2.3.5. 3,3,5,5-Tetramethyl-1-phenyl-4,5-dihydro-3H-2-benza-
zepine 2-oxide (7). White crystals, 73% yield. mp 97–9 8C; 1H
NMR (CDCl3, 400 MHz) d 1.20 (s, 6H), 1.38 (s, 6H), 2.23 (s, 2H),
7.21–7.73 (m, 9H) ppm; EI–MS (70 eV) m/z (rel. intensity): 293
(56, M+), 251 (22), 237 (19), 222 (70), 221 (32), 220 (100), 158 (37),
128 (20), 91 (21), 77 (38), 58 (19); IR (KBr, n cmꢃ1) 1540 (C N).
Anal. Calcd. for C20H23NO: C, 81.87; H, 7.90; N, 4.77. Found: C,
81.97; H, 7.77; N, 4.73.
Compound
IC50 (mM)
PBN
3
782 ꢄ 39
48 ꢄ 3.9*
17 ꢄ 2.7*,§
235 ꢄ 21*,§
63 ꢄ 5.0*
46 ꢄ 5.0*
57 ꢄ 4.2*
7 ꢄ 1.2*,§
37 ꢄ 2.2*
19 ꢄ 2.6*,§
278 ꢄ 14*,§
4
5
6
7
8
2.2.3.6. 3-Ethyl-3,5,5-trimethyl-4,5-dihydro-3H-2-benzazepine
2-oxide (8). White crystals, 67% yield. mp 81–2 8C (162–9 8C for
hydrochloride); 1H NMR (CDCl3, 400 MHz) d 0.81 (t, J = 7.4 Hz,
3H), 1.3 (s, 3H), 1.44 (s, 3H), 1.58 (s, 3H), 1.71 (qd, J = 7.4, 14.4 Hz,
1H), 1.83 (qd, J = 7.4, 14.4 Hz, 1H), 2.04 (d, J = 15.4 Hz, 1H), 2.27 (d,
J = 15.4 Hz, 1H), 7.10 (bd, J = 7.6 Hz, 1H), 7.20 (td, J = 7.6, 1.3 Hz,
1H), 7.26 (td, J = 7.6, 1.3 Hz, 1H), 7.33 (bd, J = 7.6 Hz, 1H), 7.97 (s,
1H) ppm; EI–MS (70 eV) m/z (rel. intensity): 231 (16, M+), 214 (17),
202 (80), 186 (29), 162 (55), 160 (12), 145 (100), 144 (56), 143 (30),
130 (35), 129 (57), 128 (38), 115 (33), 91 (17), 77 (13), 41 (10); IR
(KBr, n cmꢃ1) 1542 (C N). Anal. Calcd. for C15H21NO: C, 77.88; H,
9.15; N, 6.05. Found: C, 77.90; H, 9.21; N, 6.17.
9
10
11
12
The ꢀOH formation was monitored by fluorimetry using terephtha-
lic acid (THA) as a chemical dosimeter. The IC50 values, determined
using the 10-min time point, represent the concentrations of the
test nitrones required to inhibit ꢀOH formation by 50%. Values are
means ꢄ S.E.M. from four independent determinations. Statistical
significance at P < 0.05 (one-way ANOVA and Bonferroni test) was
established in comparison with the corresponding control: *, PBN;
§, cmpd 3.
2.2.3.7. 5,5-Dimethyl-4,5-dihydrospiro[2-benzazepine-3,10-cy-
clohexane] 2-oxide (11). White crystals (hydrochloride), 50%
yield. mp 164–5 8C; 1H NMR (CDCl3, 400 MHz) d 1.36–1.61 (m,
4H), 1.51 (s, 6H), 1.72–1.90 (m, 4H), 2.33–2.43 (m, 2H), 2.41 (s, 2H),
7.43 (td, J = 7.6, 1.3 Hz, 1H), 7.49–7.53 (m, 1H), 7.59–7.66 (m, 2H),
9.13 (s, 1H) ppm; EI–MS (70 eV) m/z (rel. intensity): 257 (11, M+),
240 (100), 198 (26), 145 (17), 144 (16), 143 (12), 130 (15), 129 (28),
128 (33), 115 (16); IR (KBr, n cmꢃ1) 1586 (C N). Anal. Calcd. for
C17H23NO: C, 79.33; H, 9.01; N, 5.44. Found: C, 79.28; H, 9.16; N,
5.38.
The protein concentration of mitochondrial preparations was
determined according to the method of Markwell et al. [41],
using BSA as the standard.
2.4.
Fluorimetric monitoring of hydroxyl radical formation
The hydroxyl radical (ꢀOH) formation was monitored fluor-
imetrically using a previously reported method [14,40,42,43],
in which terephthalic acid (THA) is used as a dosimeter for the
detection of ꢀOH in vitro. A luminescence spectrometer Model
LS50B (Perkin-Elmer, Norwalk, CT, USA) was used. The cuvette
holder was thermostatically maintained at 37 8C and the
sample continuously stirred with a magnetic stirrer.
2.3.
Preparation of brain mitochondria
All procedures were carried out in accordance with the NIH
Guide for Care and Use of Laboratory Animals, and were approved
by the Animal Ethics Committee of the Universidad de
Santiago de Compostela (Santiago de Compostela, Spain).
Male Sprague–Dawley rats weighting 250–300 g were used. The
rats were received from the breeder (Animalario de la
Universidad de Santiago de Compostela) at least 2 days before
A buffer solution (Na2PO4/KH2PO4 isotonized with KCl, pH
7.4) containing 10 mM terephthalic acid was incubated for
5 min to reach the working temperature. Then, the test 2-
benzazepine nitrone (final concentration range from 5 mM to
1 mM) or water was incorporated into the incubation followed
by addition of either 6-OHDA (10 mM) or 1 mM KCl (pH 2.0). All
the reported concentrations are final concentrations in the
ꢀ
sacrifice, and were kept at 22 8C, on
a
12:12 light–dark
incubation mixture. Monitoring of OH formation was imme-
schedule, and with ad libitum access to food and water.
Animals were stunned with carbon dioxide and killed by
decapitation. Brains were immediately removed and washed
in ice-cold isolation medium (pH 7.4, Na2PO4/KH2PO4 iso-
tonized with sucrose). Brain mitochondria were then obtained
by differential centrifugation with minor modifications to a
previously published method [40]. Briefly, after removing
blood vessels and pial membranes, the brains were manually
homogenized with four volumes (w/v) of the isolation
medium. Then, the homogenate was centrifuged in an Avanti
J-25 centrifuge (Beckman Instruments, Palo Alto, USA) at
1000 ꢂ g for 5 min at 4 8C. The supernatant was centrifuged at
12,500 ꢂ g for 15 min. The mitochondrial pellet was then
washed once with isolation medium and recentrifuged under
the same conditions. Finally, the mitochondrial pellet was
reconstituted in a buffer solution (Na2PO4/KH2PO4 isotonized
with KCl, pH 7.4) and stored in aliquots under liquid nitrogen.
diately initiated and maintained for the subsequent 10 min.
The excitation and emission wavelengths used were 312 and
426 nm, respectively. Fluorescence measurements were all
relative to the initial reading and the peak of relative
fluorescence (DFmax) used to express the amount of the
produced ꢀOH.
The test compound concentration that inhibited ꢀOH
formation by 50% (IC50) was calculated using the program
Origin1 v. 6.0 (Microcal Software Inc., Nothampton, MA, USA)
from a concentration-response curve generated with seven
different concentrations of the 2-benzazepine nitrone (Table 1).
2.5.
Lipid peroxidation assay
The inhibitory effects of the test nitrones on the LPO were
assessed by monitoring the formation of TBARS, using a
slightly modified spectrophotometric method [40].

b i o c h e m i c a l p h a r m a c o l o g y 7 5 ( 2 0 0 8 ) 1 5 2 6 – 1 5 3 7
1531
Table 2 – Effects of 2-benzazepine nitrones on 6-OHDA-
induced oxidative stress in brain mitochondria
A brain mitochondrial preparation (1 mg protein/mL), in a
Na2PO4/KH2PO4 buffer (pH 7.4) isotonized with KCl, was
incubated at 37 8C for 5 min to reach the working temperature.
Then, the test nitrone (concentrations ranging from 5 mM to
1 mM) or water was incorporated into the incubation, followed
by addition of 6-OHDA (10 mM) or 1 mM KCl (pH 2.0), and the
mixture incubated for exactly 20 min. Butylated hydroxyto-
luene (20 mM) and desferrioxamine (20 mM) were immediately
added, in order to prevent amplification of the lipid peroxida-
tion during the assay. All the above reported concentrations
are final concentrations in the incubation medium.
Compound IC50 (mM) Lipophilicity
log k0w
b
LPO (TBARS)
PCOs
clog Pa
PBN
3
8500 ꢄ 400
11,300 ꢄ 182
97 ꢄ 9.4*
1.25
2.30
2.84
3.88
4.42
3.97
3.37
2.78
3.34
3.88
4.92
374 ꢄ 27*
2.41
2.72
3.38
3.37
3.43
3.51
3.05
3.55
3.47
3.41
4
176 ꢄ 11*,§
40 ꢄ 3.4*,§
c
c
5
–
–
6
310 ꢄ 11*
137 ꢄ 16*,§
160 ꢄ 8.1*,§
36 ꢄ 6.2*,§
43 ꢄ 2.9*,§
20 ꢄ 3.5*,§
349 ꢄ 14*,§
70 ꢄ 7.8*
7
8
144 ꢄ 9.3*,§
53 ꢄ 3.9*,§
91 ꢄ 5.7*
An aliquot of the sample (200 mL) was treated with SDS
(8.1%, w/v), followed by addition of acetic acid (20%), and the
mixture vortexed for 1 min. Then, thiobarbituric acid (TBA,
0.8%) was added and the resulting mixture incubated at 95 8C
for 60 min. After cooling to room temperature, 3 mL of n-
butanol were added and the mixture shaken vigorously. After
centrifugation at 2500 ꢂ g for 10 min, the absorbance of the
supernatant (organic layer) was measured at 532 nm using an
Ultrospec III spectrophotometer (Pharmacia Biotech, Uppsala,
Sweden).
9
10
11
12
48 ꢄ 4.0*,§
184 ꢄ 14*,§
c
–
In vitro protective effects of nitrones against 6-OHDA autoxida-
tion-dependent lipid peroxidation (LPO), as assessed by TBARS
assay, and protein oxidative damage, as assessed by measuring
the content of protein carbonyls (PCOs), in rat brain mitochondrial
preparations. IC50 values, determined using the 20-min time point,
are means ꢄ S.E.M. from four independent determinations. Statis-
tical significance at P < 0.05 (one-way ANOVA and Bonferroni test)
was established in comparison with the corresponding control: *,
PBN; §, cmpd 3.
For calibration, a standard curve (5–150 nM) was generated
using the malondialdehyde (MDA), prepared via acid-cata-
lyzed hydrolysis (H2SO4; 1.5%, v/v) of 1,1,3,3-tetraethoxypro-
pane. The protein concentration of the sample was
determined according to a previously reported method [41],
using BSA as the standard. The IC50 values are listed in Table 2.
a
Log of n-octanol-water partition coefficient calculated with the
ACDLabs software, release 9.0 (Advanced Chemistry Development,
Inc., Toronto, Canada).
b
Log of RP-HPLC polycratic capacity factor, i.e., capacity factor
extrapolated at 100% aqueous mobile phase.
c
No statistically significant inhibition activity was attained up to
2.6.
Assessment of protein carbonyl content
the maximum tested concentration (1 mM).
The ability of the examined cyclic nitrones to inhibit the
protein oxidative damage caused by the autoxidation of 10 mM
6-OHDA in the rat brain mitochondrial preparation was
assayed by determining the content of protein carbonyls
(PCOs), through a modified published spectrophotometric
method [40]. Brain mitochondria (1 mg protein/mL) were
incubated in Na2PO4/KH2PO4 buffer (pH 7.4, isotonized with
KCl), at 37 8C for 5 min. Then, the test nitrone (final
concentrations ranging from 5 mM to 1 mM) or water was
incorporated into the incubation, followed by 6-OHDA (10 mM)
or 1 mM KCl (pH 2.0), and the mixture incubated for exactly
20 min.
content in the final pellet was estimated on an HCl blank pellet
processed simultaneously, using a BSA standard curve in 6 M
guanidine, and reading the absorbance at 280 nm [44]. The IC50
values are listed in Table 2.
2.7.
Statistical analysis
Data are expressed as mean ꢄ S.E.M. Differences between
means were statistically evaluated using the one-way ANOVA
followed by the Bonferroni test. Normality of populations and
homogeneity of variances were tested before each ANOVA.
The accepted level of significance in all cases was P < 0.05.
An aliquot of the sample (200 mL) was then immediately
submitted to protein precipitation, by adding of trichloroacetic
acid (TCA 20%, w/v), followed by centrifugation at 15,000 ꢂ g
for 5 min. The resulting pellet was reconstituted in 0.5 M NaOH
with sonication (Branson Sonic Corp., Danbury, CT, USA) for
5 s. Then, 10 mM 2,4-dinitrophenylhydrazine in 2 M chloric
acid was added and the mixture incubated at room tempera-
ture for 1 h, in darkness with continuous agitation. After the
addition of TCA (20%, w/v), the mixture was centrifuged at
15,000 ꢂ g for 5 min. The resulting pellet was washed twice
with ethyl acetate:EtOH (1:1, v/v). Then, the washed pellet was
reconstituted with 6 M guanidine in a 20 mM KH2PO4 buffer
(pH 2.3) and centrifuged at 15,000 ꢂ g for 5 min. The absor-
bance of the resulting solution was measured at 370 nm.
The carbonyl content was calculated from the absorbance
data, using as absorption coefficient for dinitrophenylhydra-
zone e = 22,000 Mꢃ1 cmꢃ1, and expressed as nmol carbonyls/
mg protein. Because of the numerous washing steps, protein
2.8.
Determination of lipophilicity parameters by RP-HPLC
The lipophilicity of the 2-benzazepine nitrone derivatives was
determined by reversed-phase HPLC method [45,46].
a
Compounds were dissolved in MeOH at 0.25 mg/mL, injected
onto a Symmetry C18 column (15 cm ꢂ 3.9 mm i.d., 5 mm
particles) from Waters Assoc. (Milford, MA, USA), and
retention data measured at regular increments of the volume
fraction of methanol in pH 5.0 acetate buffer (40 mM). All the
measurements were made at temperature of 25 ꢄ 1 8C, flow-
rate of 1.0 mL/min and at 254 nm wavelength, on a Waters
HPLC 1525 multisolvent delivery system, equipped with a
Waters 2487 variable wavelength UV detector (Waters Assoc.,
Milford, MA, USA).
Capacity factors (k0) of each compound at different mobile
phase compositions (0.05-increments of MeOH volume frac-

1532
b i o c h e m i c a l p h a r m a c o l o g y 7 5 ( 2 0 0 8 ) 1 5 2 6 – 1 5 3 7
tion, ranging between 0.7 and 0.3) were calculated as:
k0 = (tR ꢃ t0)/t0, where tR is the retention time of the solute
and t0 is the column dead time, measured as the elution time
of MeOH. In all the cases, the log k0 values increased with
decreasing MeOH volume fraction. Linear regression analysis
(r2 > 0.96) was performed on at least five data points (the
lowest MeOH concentrations) for each compound and the
linear relationship extrapolated to 100% aqueous mobile
phase to yield log k0w values.
Lipophilicity was also computationally assessed with the
ACDLabs software, release 9.0 (Advanced Chemistry Devel-
opment, Inc., Toronto, Canada). Both, computational and
experimental lipophilicity values are listed in Table 2.
3.
Results
Fig. 4 – The effects of 2-benzazepine nitrones 4, 6, 11 and 12
on the LPO induced by 6-OHDA (10 mM) autoxidation in rat
brain mitochondria. LPO was assessed by monitoring the
TBARS formation after incubation of mitochondrial
preparations (1 mg/mL) at 37 8C for 20 min. Data are
expressed as the means ꢄ S.E.M. (n = 4).
The effects of the examined 2-benzazepine nitrones (Fig. 2) on
ꢀ
the OH generated during 6-OHDA (10 mM) autoxidation were
assessed by measuring the relative fluorescence after 10 min
incubation time, using THA as
a chemical dosimeter
[14,40,42,43], and the maximal relative fluorescence (DFmax
)
was taken as a measure of ꢀOH production. Fig. 3 shows a
typical recording of ꢀOH formation in the absence (control) and
presence of the reference compound PBN, the most simple 2-
benzazepine 2-oxide within the series (3), taken as the positive
control, and the spiro C-3 cyclopentyl analog (9), behaving as
the most effective compound in this assay. The IC50 values
determined for the whole nitrone set are listed in Table 1.
In line with expectations, all the 7-membered ring nitrones
inhibited 6-OHDA-dependent ꢀOH formation more potently
than PBN. The analog 9, (ꢄ)-5-methyl-4,5-dihydrospiro[2-
benzazepine-3,10-cyclopentane] 2-oxide, proved to be the
most active one (IC50 = 7 mM), displaying more than 100-fold
improvement over PBN. A modest, but significant, increase of
the ꢀOH-scavenging activity was observed in compounds
bearing 5-gem-dimethyl group (4, 6 and 11) compared with the
respective mono-methyl derivatives (3, 5 and 10), whereas
adding a methyl group to the nitrone C-1 (5, 6 and 12), and to a
lesser extent a phenyl group (7), decreases the reactivity of the
nitrone group most likely because of its limited steric
accessibility.
6-OHDA-induced oxidative stress, producing significant
increase of LPO product level and protein damage, is believed
to be involved in neuronal cell death related to PD. Using this
model, we studied the effects of the 2-benzazepine nitrones 3–
12 on 6-OHDA-induced oxidative damage in rat brain
mitochondrial membrane homogenates as previously
described [40], assessing LPO, as measured by the formation
of TBARS, and protein oxidative damage, as revealed by the
PCO.
With the exceptions of compounds 5 and 12, which did not
attain 50% LPO inhibition in the 5-1000 mM concentration
range, all the tested benzazepine N-oxides provoked marked
reduction in the level of TBARS, showing inhibitory potencies
higher than that of PBN. Inhibition was concentration
dependent (Fig. 4) and IC50 values varied with different
nitrones (Table 2).
The most simple cyclic nitrone assayed (3) displayed more
than 20-fold increase of LPO inhibition potency over PBN, but
they were once again the spiro C-3 cycloalkyl analogs which
showed the most remarkable effects with IC50 values of 20–
43 mM (9–11). Among them, compound 11, which bears 5-gem-
dimethyl and C-3 spirocylohexyl groups, behaved as the most
potent one. Nitrones 5 and 12 lost LPO inhibition activity,
likely due to the steric effect of CH3 a to N-oxide group, such an
unfavorable effect being partially reversed in compounds 6
(R1 = methyl) and 7 (R1 = phenyl), thanks to the potency-
increasing effects of the gem-dimethyl group at the C-5
position. A plot of pIC50 (ꢃlog IC50) values in TBARS assay
Fig. 3 – Representative recording of ꢀOH formation during
the autoxidation of 6-OHDA in the absence (control) and
presence of PBN and two 2-benzazepine nitrones (3 and 9).
Incubations were performed in a 25 mM phosphate buffer
(pH 7.5) at 37 8C. The concentrations used were: 10 mM for
ꢀ
6-OHDA and 50 mM for the nitrones. The OH formation is
monitored by the fluorescence detected using terephthalic
acid as a chemical dosimeter.

b i o c h e m i c a l p h a r m a c o l o g y 7 5 ( 2 0 0 8 ) 1 5 2 6 – 1 5 3 7
1533
Fig. 5 – (A) Comparison between log (1/IC50) values measuring the ability of the assayed 2-benzazepine nitrones to scavenge
6-OHDA-induced ꢀOH radicals (pIC5O0H) and inhibit lipid peroxidation (pIC5L0PO) in rat brain mitochondria. (B) Plot of measured
LPO
50
(obs) versus calculated (calc), using Eq. (3), pIC
values.
versus ꢀOH-scavenging activity (Fig. 5) shows that the potency
of the 2-benzazepine nitrones in inhibiting LPO is reasonably
correlated with their ꢀOH-scavenging ability. To allow SAR
information to be retained, we used a truncated pICL50PO value of
3.00 for compounds 5 and 12 (IC50 > 1 mM), and obtained the
following linear Eq. (1), which explains more than 70% of the
variance in y-data:
Linear regression between pIC50 value in PCO assay (a
PCO
truncated pIC
value of 3.00 was used, in this case, for
50
compound 5) and ꢀOH-decreasing ability yielded the following
linear Eq. (2), statistically poorer than Eq. (1):
PCO
50
OH
pIC
¼ 0:71ðꢄ0:19ÞpIC þ 0:85ðꢄ0:19Þ
50
n ¼ 10; r2 ¼ 0:640; s ¼ 0:276; F ¼ 14:2
(2)
LPO
50
OH
pIC
¼ 1:02ðꢄ0:23ÞpIC ꢃ 0:61ðꢄ1:00Þ
50
Taking into account the results obtained by others [31–34],
who highlight a major role of lipophilicity in modulating LPO
inhibition by cyclic nitrones, we determined the relative
lipophilicity of the 2-benzazepine nitrones by a well-estab-
lished reversed-phase (RP)–HPLC technique (log k0w) [47], using
previously reported methods [45,46]. A comparison between
the experimental parameters log k0w (Table 2) and octanol–
water partition coefficients calculated by the ACDLabs soft-
ware (clog P) showed that a linear relation between the two
sets of lipophilicity parameters holds up to clog P of ca. 3.5, a
value beyond which log k0w no longer increases. This should be
not surprising, since the ACDLabs calculation system, as well
as other log P fragmental calculation methods, may fail in
lipophilicity prediction either because of the high uncertainty
of the estimated fragmental log P values of the nitrone group
and connected atoms and/or because the highest log P values
are too close or above the upper limit for precise partition
coefficient measurements, which constitute the basis of the
log P calculation fragmental system. The spread of the
measured log k0w parameters is rather limited (1.14 log units),
and most of the examined compounds (5–8, 10–12) varied in
just 0:2 log k0w units, that is a too small range to be a matter of
physicochemical interpretation. Nonetheless, within the
limits of the small set examined, when comparing log kw0
values among themselves, some significant findings can be
highlighted. Thus, the addition of the second CH3 group at C-5
leads to a small (compare 4 with 3) to nil (compare 6 with 5)
increase of lipophilicity. The increment of log k0w by about one
unit caused by 1-methyl group (5 versus 3), and lo a lesser
n ¼ 10; r2 ¼ 0:711; s ¼ 0:336; F ¼ 19:6
(1)
where n represents the number of data points, r2 the squared
correlation coefficient, s the standard deviation of the regres-
sion equation, F the statistical significance of fit; 95% confi-
dence intervals of the regression coefficients are given in
parentheses.
With respect to the effect of 6-OHDA on the oxidative status
of mitochondrial proteins, herein assayed by measuring the
extent of PCO, PBN displayed a very low inhibitory activity
(IC50 = 11.3 mM). In contrast, the addition to the incubation of
the 2-benzazepine nitrones, with the only exception of
compound 5, caused significant reduction in carbonyl content
in rat brain mitochondrial proteins (Table 2). The nitrone
derivative 3 inhibited PCO with an IC50 value of 97 mM, but
compounds 4, 9 and 11 showed a 2-fold increase of the
inhibitory potency over 3. PCO inhibition effects quite similar
to those measured on LPO, albeit not linearly correlated to
them, were observed. As a matter of fact, like in LPO inhibition,
5-gem-dimethyl derivatives were more potent inhibitors than
the corresponding 5-methyl ones (compare 4, 6 and 11 with 3,
5 and 10, respectively), as well as R1-methyl bearing 2-
benzazepine nitrones (5, 6 and 12) loose antioxidant activity
compared to the corresponding R1-desmethyl compounds (3, 4
and 10), but in contrast with the LPO/TBARS results, in the PCO
assay the spiro C-3 cyclohexyl compounds 10 and 11 did not
show any remarkable increase of potency compared with the
corresponding 3-gem-dimethyl analogs 3 and 4, respectively.

1534
b i o c h e m i c a l p h a r m a c o l o g y 7 5 ( 2 0 0 8 ) 1 5 2 6 – 1 5 3 7
extent by 1-phenyl group (7 versus 4), should account for the
decrease in the water-accessible surface area of the polar N+–
Oꢃ group, besides the additive hydrophobic effect of the R1-
substituents.
studies on strictly related analogs [33]. Our data from a
ꢀ
chemical fluorimetric assay using THA as OH dosimeter [42]
indicate that compounds 3–12 reduce, more efficiently than
does PBN, 6-OHDA-induced ꢀOH production (Table 1). They
should act as ꢀOH scavengers, since in our assay conditions we
can rule out iron chelation and, thus, prevention of the Fenton
reaction, as instead demonstrated using other models of
oxidative stress [26]. Relating structure of the 2-benzazepine
nitrones examined with their activity against ꢀOH production
highlighted a 2–4-fold increase effect of the gem-dimethyl
group at C-5 (4, 6 and 11), compared with the respective mono-
methyl derivatives (3, 5 and 10). Such a higher activity of 5,5-
dimethyl-substituted 2-benzazepine nitrones could most
likely be attributed to their structural peculiarity, inasmuch
the presence of the 5-gem-dimethyl group should flatten the
azepine ring, thus enhancing reactivity (radical attack onto
nitrone fragment taking place from both sides of the azepine
pseudo plane) and/or increasing stability of the resulting spin
adducts [33]. The 1H NMR data of the 5-gem-dimethyl
derivatives, in agreement with geometry calculations on
optimized structures, are reasonably consistent with this
assumption. Indeed, the 4-CH2 protons, which lose their axial
and equatorial characters, resonate as singlet, showing their
equivalence in the 5-gem-dimethyl-bearing derivatives 4, 6
and 11. In contrast, in 5-methyl derivatives 3, 5 and 10 the C-4
protons’ signals appear as two distinct doublets of doublets,
with characteristic coupling constants for the coupling with
the vicinal H-5, indicating that, in solution, the azepine ring
preferably adopts a chair conformation with the methyl group
at C-5 disposed equatorially [35]. As an additional SAR
information, a methyl group to the nitrone C-1 (5, 6 and 12),
and to a lesser extent a phenyl group (7), diminish the ꢀOH-
scavenging activity of the nitrone group, because of its limited
steric accessibility.
Investigating the role of lipophilicity in inhibiting TBARS
formation in brain mitochondria, a trend was observed which
LPO
50
suggested a bilinear relationship [48] between pIC
and
log k0w. LPO inhibitory potency of 2-benzazepine nitrones
increases as the log k0w increases from 2.41 for 3 to 3.05 for
9; when log k0w surpasses the value of 3.37, the potency no
longer increases but steeply declines with further increase of
0
0
LPO
50
the log kw. Nonlinear regression between pIC
and log kw did
not afford any statistically valid quantitative SAR, whereas
including in Eq. (1) the lipophilicity parameter as a second
independent variable, the following two-parameter equation
was obtained, which explains 90% of the variance in the LPO
inhibition data:
0
LPO
50
OH
pIC
¼ 1:18ðꢄ0:23ÞpIC þ 0:70ðꢄ0:19Þlog kw ꢃ 3:61ðꢄ1:04Þ
50
n ¼ 10; r2 ¼ 0:900; s ¼ 0:211; F ¼ 31:5
(3)
LPO
50
A plot of the measured versus calculated pIC
values is
illustrated in Fig. 5. Eq. (3) shows that the inhibition potency of
6-OHDA-induced TBARS formation in rat brain mitochondria
primarily depends upon the ꢀOH-scavenging activity, but
significantly increases with increasing lipophilicity, as
accounted for by the log k0w parameter. The introduction of
clog P instead of log k0w proved to be not statistically significant.
Addition of log k0w (and clog P as well) to Eq. (2), which describes
the inhibition potency against PCO (pICP50CO), did not result in
any statistical improvement, suggesting that lipophilicity does
not play a role in inhibiting PCO as important as in inhibiting
formation of lipid peroxide products.
Mitochondria-specific effects, potentially contributing to
neuroprotective properties, were determined as the ability of
the examined 2-benzazepine nitrones to inhibit LPO and PCO
(Table 2), induced in rat brain mitochondria by autoxidation of
6-OHDA. Our data showed that the investigated cyclic nitrones
are one-two orders of magnitude more potent than PBN in
protecting rat brain mitochondria against the oxidative
insults, as measured by inhibition of TBARS formation (LPO)
and PCO increase. The IC50 values determined in both assays
are somehow correlated with the ability of the 2-benzazepine
nitrones to scavenge ꢀOH radicals (Eqs. (1) and (2)), suggesting
that their neuroprotective activity may stem primarily, but not
uniquely, from trapping these damaging radicals. Formation
of nitrosyl radical species, which act as chain-breaking
antioxidants [55], preventing propagation of both LPO and
PCO, may be supposed as an additional mechanism account-
ing for the observed protective properties.
4.
Discussion
There is growing evidence indicating that oxidative stress,
which results in significant increase of lipid peroxidation
products and protein oxidative damage, is a key contributor to
the pathogenesis of several neurodegenerative diseases,
including PD [14,49–51]. In this study, we used 6-OHDA, a
neurotoxin producing experimental model of PD [52], to
generate in vitro ꢀOH and to cause LPO and PCO [51], and
investigated a series of new seven-membered cyclic analogs of
the spin trap PBN, namely 3,3-dimethyl-4,5-dihydro-3H-2-
benzazepine nitrones and their spiro C-3 cycloalkyl analogs,
for their ability to protect rat brain mitochondria against
oxidative injury.
Autoxidation of 6-OHDA is an important source of
intracellular ROS. Under physiological conditions, it is rapidly
oxidized affording formation of H2O2, ꢀOH, and the corre-
sponding p-quinone [53]. The formation of the hazardous ꢀOH
during the autoxidation of 6-OHDA has been proven to occur
without involvement of ferrous ion, or any other transition
metal ion, in Fenton-type reactions [54]. The inhibition of the
6-OHDA-induced oxidative damage by our 2-benzazepine
nitrones should primarily stem from their ability to trap
radical species, as previously shown by others through ESR
From an SAR point of view, R1 substitution with a methyl
group appears as unfavorable, whereas the presence of the 5-
gem-dimethyl group, more than the 5-methyl one, as well as
the replacement of the 3-gem-dimethyl group with a spir-
ocyclopentyl or spirocyclohexyl moiety, increases the inhibi-
tion potency against both LPO and PCO.
A statistically
significant equation (Eq. (3)) was derived, that correlates the
in vitro LPO (but not PCO) inhibition data with ꢀOH-scavenging
reactivity and lipophilicity of the 2-benzazepine nitrones, as

b i o c h e m i c a l p h a r m a c o l o g y 7 5 ( 2 0 0 8 ) 1 5 2 6 – 1 5 3 7
1535
described by the experimental log k0w parameter (but not by the
calculated log P). According to other reported quantitative
SARs [31], Eq. (3) indicates that, their ꢀOH-scavenging potency
being equal, 2-benzazepine nitrones having higher affinity for
and greater residence time in the lipid phase are more
effective in protecting the lipids from peroxidation. Among the
most potent nitrones examined so far, compound 11 is
noteworthy and is expected to be suitable for replacing PBN,
whose oxygen-centered radical spin adducts are very short-
lived [56], in the investigation of radicals in lipid membranes.
Investigating some systematic variations in the structure of
3,3-dimethyl-4,5-dihydro-3H-2-benzazepine N-oxide, some
new cyclic nitrones have been singled out for their promising
activity in protecting rat brain mitochondria from oxidative
injury triggered by the PD-inducing neurotoxin 6-OHDA. Some
of the examined derivatives, especially compound 11, which
incorporates into its structure the best features identified in
this study (5-gem-dimethyl group and spiro C-3 cyclohexyl
groups, and optimal lipophilicity), proved to be orders of
magnitude more effective than PBN in vitro against 6-OHDA-
induced mitochondrial lipid peroxide formation and protein
oxidative damage.
[2] Geronikaki AA, Gavalas AM. Antioxidants and
inflammatory disease: synthetic and natural antioxidants
with anti-inflammatory activity. Comb Chem High
Throughput Screen 2006;9:425–42.
[3] Mandel S, Grunblatt E, Riederer P, Gerlach M, Levites Y,
¨
Youdim MBH. Neuroprotective strategies in Parkinson’s
disease: an update on progress. CNS drugs 2003;17:729–62.
[4] Stadtman ER. Protein oxidation and aging. Free Radic Res
2006;40:1250–8.
[5] Simola N, Morelli M, Carta AR. The 6-hydroxydopamine
model of Parkinson’s disease. Neurotox Res 2007;11:151–67.
[6] Smeyne RJ, Jackson-Lewis V. The MPTP model of
Parkinson’s disease. Mol Brain Res 2005;134:57–66.
[7] Lorenc-Koci E, Rommelspacher H, Schulze G, Wernicke C,
Kuter K, Smiałowska M, et al. Parkinson’s disease-like
syndrome in rats induced by 2,9-dimethyl-beta-
carbolinium ion, a beta-carboline occurring in the human
brain. Behav Pharmacol 2006;17:463–73.
[8] McNaught KStP, Carrupt P-A, Altomare C, Cellamare S,
Carotti A, Testa B, et al. Isoquinoline derivatives as
endogenous neurotoxins in the aetiology of Parkinson’s
disease. Biochem Pharmacol 1998;56:921–33.
[9] Grunblatt E, Mandel S, Youdim MB. Neuroprotective
¨
strategies in Parkinson’s disease using the models of 6-
hydroxydopamine and MPTP. Ann NY Acad Sci
2000;899:262–73.
The SAR information and lipophilicity data reported herein
are important for drug design of potential therapeutics useful
for treatment of neurodegenerative conditions in which
oxidative stress plays a role. Nevertheless, other properties
must be taken into account for improving their therapeutic
index. Indeed, while lipophilicity is generally recognized as a
major property modulating antioxidant activity of cyclic
nitrones in both LPO assay and maintenance of cell viability
in models of oxidative injury in vitro [26,57], brain penetration
and toxicity to neurons (ranging from sedation at moderate
doses to death at very high doses), which in turn did not
appear to linearly depend on lipophilicity and nitrone brain
level following ip injection in rats, respectively, strongly
contribute to the therapeutic utility of cyclic nitrones in
neurodegenerative disorders [57]. Results from our current
investigation on selected 2-benzazepine nitrones in cell
cultures and in vivo models of oxidative stress are expected
to shed light on physicochemical properties of nitrone
derivatives which could readily penetrate the brain while
retaining a desirable side effect profile.
[10] McDowell I. Alzheimer’s disease: insights from
epidemiology. Aging 2001;13:143–62.
[11] Cadet JL, Katz M, Jackson-Lewis V, Fahn S. Vitamin E
attenuates the toxic effects of intrastriatal injection of 6-
hydroxydopamine (6-OHDA) in rats: behavioral and
biochemical evidence. Brain Res 1989;476:10–5.
[12] Kaur D, Yantiri F, Rajagopalan S, Kumar J, Mo JQ,
Boonplueang R, et al. Genetic or pharmacological iron
chelation prevents MPTP-induced neurotoxicity in vivo: a
novel therapy for Parkinson’s disease. Neuron 2003;37:899–
909.
[13] Zecca L, Youdim MB, Riederer P, Condor JR, Crichton RR.
Iron, brain ageing and neurodegenerative disorders. Nat
Rev Neurosci 2004;5:863–73.
´
´
[14] Soto-Otero R, Me´ndez-Alvarez E, Hermida-Ameijeiras A,
Lo´pez-Real AM, Labandeira-Garcı´a JL. Effects of (ꢃ)-nicotine
and (ꢃ)-cotinine on 6-hydroxydopamine-induced oxidative
stress and neurotoxicity: relevance for Parkinson’s disease.
Biochem Pharmacol 2002;64:125–35.
[15] Pezzella A, d’Ischia M, Napolitano A, Misuraca G, Prota G.
Iron-mediated generation of the neurotoxin 6-
hydroxydopamine quinone by reaction of fatty acid
hydoperoxides with dopamine: a possible contributory
mechanism for neuronal degeneration in Parkinson’s
disease. J Med Chem 1997;40:2211–6.
[16] Graham DG, Tiffany SM, Bell Jr WR, Gutknecht WF.
Autoxidation versus covalent binding of quinones as the
mechanism of toxicity of dopamine, 6-hydroxydopamine,
and related compounds toward C1300 neuroblastoma cells
in vitro. Mol Pharmacol 1978;14:644–53.
Acknowledgements
The Spanish authors thank the Ministerio de Educacio´n y
Ciencia and the Europe Regional Development Fund (Madrid,
Spain, Grants BFI2003-00493 and SAF2007-66114) for financial
support. The Italian authors thank the Italian Ministry for
Education Universities and Research (MIUR, Rome, Italy; PRIN
2004, Grant No. 2004037521_006) for financial support.
[17] Gee P, Davison AJ. Intermediates in the aerobic
autoxidation of 6-hydroxydopamine: relative importance
under different reaction conditions. Free Radic Biol Med
1989;6:271–84.
[18] Glinka YY, Youdim MBH. Inhibition of mitochondrial
complexes I and IV by 6-hydroxydopamine. Eur J
Pharmacol 1995;292:329–32.
[19] Glinka YY, Tipton KF, Youdim MBH. Nature of inhibition of
mitochondrial respiratory complex I by 6-
r e f e r e n c e s
hydroxydopamine. J Neurochem 1996;66:2004–10.
[20] Kumar R, Agarwal AK, Seth PK. Free radical-generated
neurotoxicity of 6-hydroxydopamine. J Neurochem
1995;64:1703–7.
[1] Yung LM, Leung FP, Yao X, Chen ZY, Huang Y. Reactive
oxygen species in vascular wall. Cardiovasc Hematol Disord
Drug Targets 2006;6:1–19.

1536
b i o c h e m i c a l p h a r m a c o l o g y 7 5 ( 2 0 0 8 ) 1 5 2 6 – 1 5 3 7
[21] Dalle-Donne I, Aldini G, Carini M, Colombo R, Rossi R,
Milani A. Protein carbonylation, cellular dysfunction, and
disease progression. J Cell Mol Med 2006;10:389–406.
[22] Dobson CM. Protein folding and misfolding. Nature
2003;426:884–90.
[23] Aldini G, Dalle-Donne I, Facino RM, Milani A, Carini M.
Intervention strategies to inhibit carbonylation by
lipoxidation-derived reactive carbonyls. Med Res Rev
2006;27:817–68.
[24] Choi J, Rees HD, Weintraub ST, Levey AI, Chin LS, Li L.
Oxidative modifications and aggregation of Cu/Zn
superoxide dismutase associated with Alzheimer’s and
Parkison’s diseases. J Biol Chem 2005;280:11648–55.
[25] Thomas CE, Huber EW, Ohlweiler DF. Hydroxyl and peroxyl
radical trapping by the monoamine oxidase-B inhibitors
deprenyl and MDL 72,974A: implications for protection of
biological substrates. Free Radic Biol Med 1997;22:733–7.
[26] Thomas CE, Ohlweiler DF, Taylor VL, Schmidt CJ. Radical
trapping and inhibition of iron-dependent CNS damage by
cyclic nitrone spin traps. J Neurochem 1997;68:1173–82.
[27] Maples KR, Ma F, Zhang YK. Comparison of the radical
trapping ability of PBN, S-PPBN and NXY-059. Free Radic
Res 2001;34:417–26.
[39] Kouznetsov VV, Palma AR, Salas S, Vargas LY, Zubkov FI,
Varlamov AV, et al. Chemistry of functionalized
benzazepines. 5. Synthesis and chemical transformation
of the 1,2,4,5-tetrahydrospiro[3H-2-benzazepin-3,10-
cycloalkanes]. J Heterocyclic Chem 1997;34:
1591–5.
´
´
[40] Hermida-Ameijeiras A, Me´ndez-Alvarez E, Sa´nchez-Iglesias
S, Sanmartı´n-Sua´rez C, Soto-Otero R. Autoxidation of MAO-
mediated metabolism of dopamine as a potential cause of
oxidative stress: sole of ferrous and ferric ions. Neurochem
Int 2004;45:103–16.
[41] Markwell MAK, Haas SM, Bieber LL, Tolbert NE. A
modification of the Lowry procedure to simplify protein
determination in membrane and lipoprotein samples. Anal
Biochem 1978;87:206–10.
[42] Barreto JC, Smith GS, Strobel NH, McQuillin PA,
Miller TA. Terephthalic acid: a dosimeter for the
detection of hydroxyl radicals in vitro. Life Sci
1995;56:PL89–96.
[43] Soto-Otero R, Sanmartı´n-Sua´rez C, Sa´nchez-Iglesias S,
Hermida-Ameijeiras A, Sa´nchez-Sellero I, Me´ndez-Alvarez
E. Study on the ability of 1,2,3,4-tetrahydropapaveroline to
cause oxidative stress: mechanisms and potential
implications in relation to Parkinson’s disease. J Biochem
Mol Toxicol 2006;20:209–20.
[28] Wang CX, Shuaib A. NXY-059: a neuroprotective agent in
acute stroke. Int J Clin Pract 2004;58:964–9.
[29] Carney JM, Starke-Reed PE, Oliver CN, Landum RW, Cheng
MS, Wu JF, et al. Reversal of age-related increase in brain
protein oxidation, decrease in enzyme activity, and loss in
temporal and spatial memory by chronic administration of
the spin trapping compound N-tert-butyl-a-phenylnitrone.
Proc Natl Acad Sci USA 1991;88:3633–6.
[30] Chen Q, Fischer A, Reagan JD, Yan L-J, Ames BN. Oxidative
DNA damage and senescenze of human diploid fibroblast
cells. Proc Natl Acad Sci USA 1995;92:4337–41.
[31] Fevig TL, Bowen SM, Janowick DA, Jones BK, Munson HR,
Ohlweiler DF, et al. Design, synthesis and in vitro
evaluation of cyclic nitrones as free radical traps for the
treatment of stroke. J Med Chem 1996;39:4988–96.
[32] Bernotas RC, Thomas CE, Carr AA, Nieduzak TR, Adams GA,
Ohlweiler DF, et al. Synthesis and radical scavenging
activity of 3,3-dialkyl-3,4-dihydroisoquinoline 2-oxides.
Bioorg Med Chem Lett 1996;6:1105–10.
[44] Reznick AZ, Packer L. Oxidative damage to proteins:
spectrophotometric method for carbonyl assay. Methods
Enzymol 1994;233:357–63.
[45] de Candia M, Fossa P, Cellamare S, Mosti L, Carotti A,
Altomare C. Insights into structure–activity relationships
from lipophilicity profiles of pyridin-2(1H)-one analogs of
the cardiotonic agent milrinone. Eur J Pharm Sci
2005;26:78–86.
[46] De Marco A, de Candia M, Carotti A, Cellamare S, De Candia
E, Altomare C. Lipophilicity-related inhibition of platelet
aggregation by nipecotic acid anilides. Eur J Pharm Sci
2004;22:153–64.
[47] Braumann T. Determination of hydrophobic parameters by
reversed-phase liquid chromatography: theory,
experimental technique, and application in studies on
quantitative structure–activity relationships. J Chromatogr
1986;373:191–225.
[33] Thomas CE, Ohlweiler DF, Kalyanaraman B. Multiple
mechanisms for inhibition of low density lipoprotein
oxidation by novel cyclic nitrone spin traps. J Biol Chem
1994;269:28055–61.
[48] Kubinyi H. Quantitative models. In: Mannhold R,
Krogsgaard-Larsen P, Timmerman H, editors. QSAR:
Hansch analysis and related approaches. Weinheim: VCH
Verlagsgesellschaft mbH; 1993. p. 68–77.
[34] Thomas CE, Ohlweiler DF, Carr AA, Nieduzak TR, Hay DA,
Adams G, et al. Characterization of the radical trapping
activity of a novel series of cyclic nitrones spin traps. J Biol
Chem 1996;271:3097–104.
[49] Alam ZI, Daniel SE, Lees AJ, Marsden DC, Jenner P, Halliwell
P. A generalized increase in protein carbonyl in the brain in
Parkinson’s disease but not incidental Lewy body disease. J
Neurochem 1997;69:1326–9.
[35] Varlvamov A, Kouznetsov V, Zubkov F, Chernyshev A,
Alexandrov G, Palma A, et al. Efficient synthesis and
spectroscopic analysis of 8-nitro spiro C-3-annulated 2-
benzazepines and their N-oxides. Synthesis 2001;6:
849–54.
[50] Dexter DT, Holley AE, Flitter WE, Slater TF, Wells FR, Daniel
SE, et al. Increased levels of lipid hydroperoxides in the
parkinsonian substantia nigra: an HPLC and ESR study. Mov
Disord 1994;9:92–7.
´
[51] Sa´nchez-Iglesias S, Rey P, Me´ndez-Alvarez E, Labandeira-
[36] Vargas Me´ndez LY, Kouznetsov V. 4-N-Aryl(benzyl)amino-
4-heteroaryl-1-butenes as building blocks in heterocyclic
synthesis. 3. A simple synthesis of fluorinated 4-methyl-2-
(3-pyridyl)-1,2,3,4-tetrahydroquinolines and their
respective quinolines. Heterocycl Commun 2002;8:
583–6.
Garcı´a JL, Soto-Otero R. Time-course of brain oxidative
damage caused by intrastriatal administration of 6-
hydroxydopamine in a rat model of Parkinson’s disease.
Neurochem Res 2007;32:99–105.
[52] Dauer W, Przedborski S. Parkinson’s disease: mechanisms
and models. Neuron 2003;39:889–909.
´
[37] Urbina JMG, Corte´s JC, Palma A, Lo´pez SN, Zacchino SA,
Enriz DR, et al. Inhibitors of the fungal cell wall. Synthesis
of 4-aryl-4-N-arylamine-1-butenes and related compounds
with inhibitory activities on b (1-3) glucan and chitin
synthases. Bioorg Med Chem 2000;8:691–8.
[38] Puentes CO, Kouznetsov V. Recent advancements in the
homoallylamine chemistry. J Heterocyclic Chem
2002;39:595–614.
[53] Soto-Otero R, Me´ndez-Alvarez E, Hermida-Ameijeiras A,
Munoz-Patino A, Labandeira-Garcı´a JL. Autoxidation and
˜
˜
neurotoxicity of 6-hydroxydopamine in the presence of
some antioxidants: potential implication in relation to the
pathogenesis of Parkinson’s disease. J Neurochem
2000;74:1605–12.
´
[54] Me´ndez-Alvarez E, Soto-Otero R, Hermida-Ameijeiras A,
Lo´pez-Martı´n ME, Labandeira-Garcı´a JL. Effect of iron and

b i o c h e m i c a l p h a r m a c o l o g y 7 5 ( 2 0 0 8 ) 1 5 2 6 – 1 5 3 7
1537
manganese on hydroxyl radical production by 6-
[56] Janzen EG, Kotake Y, Hinton RD. Stabilities of hydroxyl
hydroxydopamine: mediation of antioxidants. Free Radic
Biol Med 2001;31:986–98.
radical spin adducts of PBN-type spin traps. Free Radic Biol
Med 1992;12:169–73.
[55] Skolimowski J, Kochman A, Ge˛bicka L, Metodiewa E.
Synthesis and antioxidant activity evaluation of novel
antiparkinsonian agents, aminoadamantane
derivatives of nitroxyl free radical. Bioorg Med Chem
2003;11:3529–39.
[57] Thomas CE, Bernardelli P, Bowen SM, Chaney SF, Friedrich
D, Janowick DA, et al. Cyclic nitrone free radical traps:
Isolation, Identification, and synthesis of 3,3-dimethyl-3,4-
dihydroisoquinolin-4-ol N-oxide, a metabolite with reduced
side effects. J Med Chem 1996;39:4997–5004.