Detoxification of 4-hydroxynonenal (HNE) in keratinocytes: Characterization of conjugated metabolites by liquid chromatography/electrospray ionization tandem mass spectrometry

Article abstract of DOI:10.1002/jms.533

Keratinocytes are potential targets of lipid peroxidation products (α,β-unsaturated aldehydes) generated in the skin following UV exposure, among which the most abundant and toxic product is 4-hydroxy-trans-2,3-nonenal (HNE). The aim of this study was to investigate the ability of keratinocytes (NCTC2544 cell lines) to detoxify HNE, through characterization of metabolites, until now never demonstrated, using a combined analytical approach (liquid chromatography (LC) and liquid chromatography/mass spectrometry (LC/MS)). Incubation of cells with HNE (up to 200 μM) was performed in order to evaluate the ability of the cells to detoxify this toxic aldehyde, and indicated that the cell viability was maintained under these conditions. LC analysis of the extracellular media from keratinocytes incubated with 100 μM HNE shows a time-dependent decrease of HNE, disappearance from the medium within 2 h and concomitant formation of two unconjugated (phase I) metabolites, 4-hydroxy-2-nonenoic acid (HNA) and 1,4-dihydroxy-2-nonene (DHN), which were both identified and quantified by LC and accounted for 48.8 ± 4.6% of the HNE dose. Four additional metabolites were identified in the extracellular medium by reversed-phase LC coupled with electrospray ionization tandem mass spectrometry (LC/ESI-MS/MS) with positive and negative ion detection as Michael adducts (phase II metabolites), arising by the addition of the nucleophilic sulfur of glutathione (GSH) to the electrophilic C-3 of HNE, followed by oxidation-reduction enzymatic processes. The GSH-HNE conjugates were (a) S-(4-hydroxynonanal-3-yl)glutathione, (b) S-(1,4-dihydroxy-nonane-3-yl)glutathione, (c) S-(4-oxononanal-3-yl)glutathione and (d) S-(4-oxo-nonan-1-ol-3-yl)glutathione, and accounted for 52.3 ± 5.8% of the HNE dose (35 nmol mg^-1 protein), as estimated indirectly by measuring the extent of cellular GSH consumption (18.7 ± 1.8 nmol mg^-1 protein). The time course of HNE biotransformation was then determined by monitoring the formation of metabolites inside and outside the cell at different times after HNE addition (5-120 min). A time-dependent and almost linear formation inside the cell was observed for all the metabolites (plateau after 15 min of incubation), followed by a rapid decay and a concomitant increase in the extracellular medium (plateau of formation after 60 min). This confirms that HNE diffuses into the cell where is totally metabolized through phase I and phase II reactions to unreactive products, which are then exported outside the cell. This is the first demonstration that skin epidermal cells are able to detoxify the cytotoxic products of lipid peroxidation. Copyright

Full text of DOI:10.1002/jms.533

JOURNAL OF MASS SPECTROMETRY
J. Mass Spectrom. 2003; 38: 1160–1168
Published online 29 October 2003 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jms.533
Detoxification of 4-hydroxynonenal (HNE) in
keratinocytes: characterization of conjugated
metabolites by liquid chromatography/electrospray
ionization tandem mass spectrometry
Giancarlo Aldini,1 Paola Granata, Marica Orioli, Enzo Santaniello and Marina Carini
∗
1
1
2
1
1
Istituto Chimico Farmaceutico Tossicologico, University of Milan, Viale Abruzzi 42, 20131 Milan, Italy
Dipartimento di Scienze Precliniche LITA Vialba, University of Milan, Via G. B. Grassi 74, 20157 Milan, Italy
2
Received 21 May 2003; Accepted 27 August 2003
Keratinocytes are potential targets of lipid peroxidation products (a,b-unsaturated aldehydes) generated
in the skin following UV exposure, among which the most abundant and toxic product is 4-hydroxy-trans-
2,3-nonenal (HNE). The aim of this study was to investigate the ability of keratinocytes (NCTC2544 cell
lines) to detoxify HNE, through characterization of metabolites, until now never demonstrated, using a
combined analytical approach (liquid chromatography (LC) and liquid chromatography/mass spectrometry
(LC/MS)). Incubation of cells with HNE (up to 200 µM) was performed in order to evaluate the ability of
the cells to detoxify this toxic aldehyde, and indicated that the cell viability was maintained under these
conditions. LC analysis of the extracellular media from keratinocytes incubated with 100 µM HNE shows a
time-dependent decrease of HNE, disappearance from the medium within 2 h and concomitant formation
of two unconjugated (phase I) metabolites, 4-hydroxy-2-nonenoic acid (HNA) and 1,4-dihydroxy-2-nonene
(
DHN), which were both identified and quantified by LC and accounted for 48.8 ± 4.6% of the HNE dose.
Four additional metabolites were identified in the extracellular medium by reversed-phase LC coupled
with electrospray ionization tandem mass spectrometry (LC/ESI-MS/MS) with positive and negative ion
detection as Michael adducts (phase II metabolites), arising by the addition of the nucleophilic sulfur of
glutathione (GSH) to the electrophilic C-3 of HNE, followed by oxidation–reduction enzymatic processes.
The GSH–HNE conjugates were (a) S-(4-hydroxynonanal-3-yl)glutathione, (b) S-(1,4-dihydroxy-nonane-
3
-yl)glutathione, (c) S-(4-oxononanal-3-yl)glutathione and (d) S-(4-oxo-nonan-1-ol-3-yl)glutathione, and
−1
accounted for 52.3 ± 5.8% of the HNE dose (35 nmol mg protein), as estimated indirectly by measuring
−1
the extent of cellular GSH consumption (18.7 ± 1.8 nmol mg
protein). The time course of HNE
biotransformation was then determined by monitoring the formation of metabolites inside and outside
the cell at different times after HNE addition (5–120 min). A time-dependent and almost linear formation
inside the cell was observed for all the metabolites (plateau after 15 min of incubation), followed by a
rapid decay and a concomitant increase in the extracellular medium (plateau of formation after 60 min).
This confirms that HNE diffuses into the cell where is totally metabolized through phase I and phase II
reactions to unreactive products, which are then exported outside the cell. This is the first demonstration
that skin epidermal cells are able to detoxify the cytotoxic products of lipid peroxidation. Copyright 
2003 John Wiley & Sons, Ltd.
KEYWORDS: keratinocytes; 4-hydroxynonenal; glutathione-conjugated adducts; characterization; liquid
chromatography/electrospray ionization tandem mass spectrometry
INTRODUCTION
non-enzymatic antioxidant impairment, has recently been
recognized as a key factor of intrinsic and photo-induced
skin aging and of several skin disorders such as actinic
Oxidative damage of the skin, induced by several exogenous
and endogenous inducers such as UV radiation, transition
metal ions, pollutants (NO and ozone) and enzymatic and
2
1
elastosis and cancer. The peroxidation of epidermis and
dermis lipid substrates, leading to the formation of ˛,ˇ-
unsaturated aldehydes, has recently been proposed as a key
factor in oxidative-dependent skin damage. In particular,
Ł
Correspondence to: Giancarlo Aldini, Istituto Chimico
Farmaceutico Tossicologico, University of Milan, Viale Abruzzi 42,
2
˛,ˇ-unsaturated aldehydes, owing to their electrophilic
0131 Milan, Italy. E-mail: giancarlo.aldini@unimi.it
nature, are highly reactive with cellular nucleophiles such
as glutathione (GSH), protein side-chain of cysteine, lysine
Contract/grant sponsor: Fondo Interno per la Ricerca Scientifica e
Tecnologica; Contract/grant number: FIRST 2002.
Copyright  2003 John Wiley & Sons, Ltd.

LC/ESI-MS/MS characterization of HNE–GSH adducts in keratinocytes 1161
and histidine and with nucleic acids, leading to enzyme
inactivation, mutagenesis, protein cross-links and signal
Synthesis of HNE, DHN, HNA and GS-HNE
4-Hydroxy-non-2-enal diethylacetal (HNE-DEA), 1,4-dihy-
droxy-trans-nonene (DHN) and 4-hydroxy-trans-2-nonenoic
acid (HNA) were synthesized according to methods reported
2
,3
trasduction.
-Hydroxy-trans-2,3-nonenal (HNE) is one of the most
4
1
3,14
abundant and toxic ˛,ˇ-unsaturated aldehydes that is gen-
erated through the ˇ-cleavage of hydroperoxides from ω-6
in the literature;
HNE was prepared from HNE-
DEA by 1 mM HCl hydrolysis (1 h at room tempera-
2
polyunsaturated fatty acids (arachidonic and linoleic acids),
ture) and quantitated by UV spectroscopy (ꢁmax 224 nm;
which represent the main fatty acid constituents in the
ꢀ1
ꢀ1
ε 13 750 l mol cm ꢂ. The identity and purity of HNE-DEA
4
membrane phospholipids of epidermal keratinocytes. The
1
were established by H NMR spectroscopy (Varian Mer-
involvement of HNE in photo-induced and age-related
skin disorders is supported by the detection of carbonyl
derivatives (determined as hydrazones) in epidermis and
cury VX 300 instrument) and those of HNA and DHN by
H NMR and ESI-MS (positive and negative ion modes)
1
5
ꢀ
C
analyses: [M ꢀ H] ion at m/z 171 for HNA and [M C H]
dermis and by the immunoreactivity to HNE antibodies
of the elastotic material from sun-damaged skin affected
ion at m/z 159 for DHN. ESI-MS analyses were performed
on a Thermo Finnigan LCQ Advantage (Thermo Finnigan,
San Jose, CA, USA) ion trap mass spectrometer, operat-
ing at capillary temperature 250 C, ionization voltage 5 kV
and capillary voltage 12.5 V. The flow-rate of the nebulizer
6
by actinic elastosis. The metabolism of HNE to less reac-
tive molecules should represent a prerequisite to ensure
skin cell survival and normal functioning. As extensively
studied in different biological systems (erythrocytes, liver
slices, isolated perfused liver, smooth muscle cells or syn-
°
ꢀ1
gas (nitrogen) was 0.5 l min . Samples dissolved in water
were diluted 1 : 5 with water–acetonitrile (70 : 30) and intro-
duced into the mass spectrometer using a Harvard syringe
7
–11
ovial fibroblasts),
the main metabolic pathways of HNE
involve oxidation–reduction of the carbonyl function and
GSH conjugation. Although epidermis and dermis are con-
tinuously exposed to oxidative damage and represent a
potential target of ˛,ˇ-unsaturated aldehydes, the metabolic
fate of HNE in the epidermal cells (keratinocytes) has not yet
been investigated. The aim of the present work was to study
the metabolic pathways of HNE in keratinocytes in order to
establish the mechanism of detoxification of this cytotoxic
aldehyde. This was achieved through direct characterization
of the metabolic products in the crude biological matrix by
using a combined analytical approach (liquid chromatog-
raphy (LC) and liquid chromatography/mass spectrometry
ꢀ1
pump at a flow-rate of 25 µl min . GS–HNE conjugate was
prepared non-enzimatically by incubation of 0.2 mM HNE
with 2 mM GSH in 10 mM phosphate buffer pH 7.4 (PBS)
at 37 °C for 2 h. The reaction was monitored by following
the decrease in unreacted HNE by LC and the identity of
GS–HNE adduct was confirmed by LC/ESI-MS analysis.
The GS—HNE adduct was stored in PBS under nitrogen at
ꢀ
20 °C.
Incubation of keratinocytes with HNE
Keratinocytes (NCTC 2544 human cell line; Flow
(
LC/MS)), until now never employed for this purpose. Char-
Laboratories, Irvine, UK) were cultured at 37 °C in
acterization of HNE metabolites has been generally carried
out with labelled [ H]HNE, after LC or thin-layer chromato-
DMEM supplemented with 10% fetal calf serum, 2 mM
3
ꢀ1
L-glutamine and antibiotics (100 U ml
penicillin and
ꢀ
1
graphic isolation of radioactivity and electrospray ionization
0.1 mg ml
streptomycin) until 80–90% confluence as
n
15
(
ESI) MS or gas chromatography/MS analysis of the iso-
described previously, and then exposed to different
lated radioactive products. LC/ESI-MS has only recently
been employed for the study of HNE conjugation with GSH
in isolated perfused liver and erythrocytes and the study
HNE concentrations (20–200 µM in PBS, equivalent to
ꢀ
1
7–70 nmol mg protein HNE). After incubation for 2 or
h, cells were analyzed for cell viability, using the calcein
8
4
1
6
of the ꢀ-glutamyltranspeptidase(GTT)-dependent metabolic
AM assay and by measuring the calcein formation with a
Victor2 Wallac multi-well fluorescent plate reader (excitation
filter, 485 š 10 nm; emission filter, 530 š 12.5 nm). Protein
determinations were performed by a modified Lowry’s
fate of GS–HNE conjugate in GTT-overexpressing fibroblast
1
2
cell lines.
1
7
method using bovine serum albumin as a standard. GSH
levels were determined by the fluorescence assay described
EXPERIMENTAL
1
5
ꢀ1
Chemicals
previously and expressed as nmol mg protein. In another
set of experiments, keratinocytes were exposed to 100 µM
HNE (dissolved in PBS) and, after incubation for different
time periods (5, 15, 30, 60 and 120 min), the extracellular
media were transferred into cryovials and stored at ꢀ20 °C.
After washing the cell monolayer twice with pre-warmed
PBS, the cells were re-suspended in PBS using a rubber
All chemicals and reagents were of analytical grade and pur-
chased from Sigma-Fluka-Aldrich Chemical (Milan, Italy).
LC-grade and analytical-grade organic solvents were pur-
chased from Merck (Bracco, Milan, Italy). LC-grade water
was prepared with a Milli-Q water purification system
(
(
Millipore, Bedford, MA, USA). Calcein acetoxymethyl ester
Calcein AM) was purchased from Molecular Probes (Space
scraper and sonicated at 4
°
C. The cell homogenate was
then centrifuged at 17 000 rpm (10 min at 4
°
C) and the cell
Import–Export, Milan, Italy). Glutathione (GSH), sodium
borohydride (NaBH ), o-phthaldialdehyde (OPA) and triflu-
oroacetic acid (TFA) were purchased from Sigma-Aldrich
Milan, Italy). Materials and chemicals for cell culture were
supernatants (cell lysates) filtered (0.2 µm) and stored at
ꢀ20 °C. The extracellular media and cell lysates were directly
analysed by LC/UV and LC/ESI-MS methods to characterize
the biotransformation products of HNE.
4
(
purchased from Life Technologies (Milan, Italy).
Copyright  2003 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2003; 38: 1160–1168

1162
G. Aldini et al.
ꢀ
1
LC analysis of HNE, HNA and DHN
auxiliary gas flow-rate, 0.5 l min . The flow-rate of the
ꢀ
1
The extracellular media and cell lysates were adjusted to pH 1
with 1 M HCl and 0.5 ml aliquots were extracted three times
with 0.75 ml of dichloromethane–ethanol (2 : 1, v/v). The
organic phases were collected and concentrated to dryness
under nitrogen and the residue was dissolved in 100 µl of
phosphate buffer (pH 7.4) and filtered through a 0.45 µm filter
and 20 µl aliquots were injected for analysis of HNE, HNA
and DHN. Separations were carried out using an HP1050
series instrument (Hewlett-Packard, Palo Alto, CA, USA)
equipped with a quaternary pump system, an autosampler,
a UV/visible diode-array programmable detector operating
both at fixed wavelengths (223 and 202 nm) and in scan mode
nebulizer gas (nitrogen) was 5 l min . Spectra were acquired
in negative and positive ion modes, with a scan range from
ꢀ
1
m/z 100 to 500 (scan rate 0.5 scans s ꢂ. Collision-induced
dissociation (CID) experiments in negative and positive ion
modes were performed by selecting an ion with an isolation
width of 1 m/z and using helium as collision gas. The ion
collision energy was adjusted by changing a percentage of
the 5 V a.c. voltage which was applied to the end-caps of the
ion trap at the resonance frequency of the selected ion and
was set at 30% (referred to as collision energy level).
3
D
(
200–400 nm), an on-line degasser and a LC ChemStation.
NaBH reduction
4
Analyses were done by reversed-phase elution with a
Extracellular media from keratinocytes incubated for 2 h with
00 µM HNE were treated with 20 mM NaBH
Supelco LC18 column (250 ð 4.6 mm i.d.; particle size 5 µm)
1
2
4
, incubated at
(
(
Sigma-Aldrich). The mobile phase, composed of 60% A
water–acetonitrile–formic acid (90 : 10 : 0.1, v/v/v)) and
5 š 1 °C for different time periods (1, 2, 4, 6, 24 h) and
directly analyzed by LC/ESI-MS as described above.
4
0% B (water–acetonitrile (10 : 90, v/v)), was delivered at
ꢀ
1
a flow-rate of 1 ml min
.
Calibration curves, prepared by spiking cell lysate and
supernatant samples with the analytes to give concentrations
ranging between 1 and 200 µM (HNE, HNA) or 10 and
a)
b)
2
00 µM (DHN) were constructed by weighted (1/x) least-
squares linear regression analysis of the peak areas of
analytes versus concentration. The calibration equations
were used to calculate the concentration of analytes in the
replicated and incubated samples. For each compound the
standard response was linear over the entire calibration
range, with correlation coefficients >0.995. The equations
of the calibration lines were as follows: y D 2351x ꢀ 2894
Figure 1. Phase-contrast microscopic examination of
keratinocytes after 2 h of incubation in the presence of PBS
(
A D controls) and 200 µM HNE in PBS (B). After gently washing
2
2
(
r D 0.9969) for HNE, y D 2169x C 3205 (r D 0.9988) for
with PBS, the cells were analyzed by phase-contrast
2
HNA and y D 233x ꢀ 528 (r D 0.9978) for DHN. The limits
of quantitation (LOQs) were determined as 1.0 µM (HNE,
HNA) and 10 µM (DHN), being the lowest concentrations
that can be determined with an acceptable accuracy of š20%
and a precision below 20%. A recovery test was performed
by comparing the response of extracted samples spiked
with the analytes before extraction with that of extracted
blank samples spiked just before injection. The recoveries,
determined at the two extremes of the calibration ranges
microscopy using a 20ð lens (Eclipse TS 100, Nikon).
120
1a
100
OH
O
8
6
4
0
0
0
1a
(
n D 6 each), were satisfactory, ranging from 85 to 95% for
2a
all the compounds (data not shown).
20
0
LC/ESI-MS analyses
1
1
20
00
80
60
OH
OH
LC/MS experiments were carried out on a Thermoquest
Surveyor system equipped with a quaternary pump, a
Surveyor UV/visible diode-array programmable detector
O
OH
2a
OH 3a
(
6000 LP), a Surveyor AS autosampler, a vacuum degasser,
2
a
Xcalibur software and connected to a Thermo Finnigan LCQ
Advantage ion trap mass spectrometer (ITMS). Separations
were done by reversed-phase elution with a Phenomenex
Synergy Polar-RP column (150 ð 2 mm i.d.; particle size
4
2
0
0
0
3
a
0
1
2
3
4
5
6
7
8
9 10
4
µm) (Chemtek Analytica, Anzola Emilia, Italy) protected by
Time (min)
a MAX-RP guard column (4 ð 2 mm i.d.; 4 µm), with isocratic
elution with TFA–CH CN–H O (0.05 : 4 : 6, v/v/v) at a flow-
3
2
Figure 2. Reversed-phase LC analysis of extracellular media
from keratinocytes incubated with HNE (100 µM) at different
observation times (top, 5 min of incubation; bottom, 2 h of
incubation). Detector: UV-DAD in scan mode (200–400 nm).
ꢀ
1
rate of 0.2 ml min . The ESI-MS source was operated as
follows: capillary temperature, 250 °C; spray voltage, 5 kV;
ꢀ
1
capillary voltage, 5 V; sheath gas flow-rate, 2 l min ; and
Copyright  2003 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2003; 38: 1160–1168

LC/ESI-MS/MS characterization of HNE–GSH adducts in keratinocytes 1163
RESULTS AND DISCUSSION
100
80
HNE detoxification by keratinocytes
Incubation of keratinocytes with HNE (20–200 µM) did not
significantly modify cell viability up to 4 h of incubation
since it was always >95% (96.3 š 3.5% in the presence of
6
0
0
4
2
00 µM HNE) and not significantly different compared with
controls (97.1 š 2.0%; p > 0.05). The lack of HNE-induced
toxicity was confirmed by morphological analyses: when
keratinocytes were challenged with the highest HNE dose
20
0
(
2 h incubation), no significant morphological alterations
0
30
60
90
120
were observed compared with controls, as shown by repre-
sentative phase-contrast microscopic analyses (Fig. 1). Since
HNE can easily diffuse through the cell membrane and was
found to be stable in the incubation medium (LC analy-
sis, data not shown), these preliminary results suggest that
keratinocytes are able to entrap the aldehyde and detox-
ify it through a metabolic process. This was unequivocally
demonstrated by incubating keratinocytes with 100 µM HNE
Time (min)
Figure 3. Kinetics of HNE consumption and formation of HNE
metabolites in extracellular media. Values are the means š
S.D. of five independent experiments (LC analysis). (ꢀ) HNE;
(
♦) HNA; (ꢁ) DHN.
in the next intervals were time-dependently reduced and
were practically negligible in the incubation medium after
120 min, where the formation of 2a is maximum (Fig. 2, bot-
tom) and also the second metabolite 3a, identified as DHN,
the reduced derivative of HNE. Figure 3 reports the kinetics
of HNE consumption and of metabolite formation in the
extracellular medium: LC quantitation of the three species
(
a concentration used for all the subsequent metabolic stud-
ies) and by analysing the extracellular media by LC. The
chromatogram of extracellular media relative to 5 min incu-
bation (Fig. 2, top) contains, besides unchanged HNE (peak
1
a), a small peak (2a) that was identified, by comparison with
the standard, as the oxidation product HNA. HNE levels
100
100
100
100
8
0
8
6
4
2
0
0
0
8
6
4
2
0
0
0
0
0
60
80
60
40
20
0
40
20
0
0
0
0
1
2
3
4
5
6
0
1
2
3
4
5
6
Time (min)
Time (min)
0
1
2
3
4
5
6
0
1
2
3
4
5
6
Time (min)
Time (min)
5
00
95
90
85
80
500
4
4
4
4
495
490
485
480
475
470
465
460
455
450
1b
2b
3b
4b
475
4
4
4
4
4
70
65
60
55
50
1
b
2b
4
b
3
b
2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0
2
.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0
Time (min)
Time (min)
Figure 4. LC/ESI-MS analysis of extracellular media from keratinocytes incubated with HNE (100 µM). Left: positive ion detection.
TIC chromatogram (mass range: m/z 450–500; top) and two-dimensional plot (m/z values vs retention times; bottom). Right,
negative ion detection. TIC chromatogram (mass range: m/z 450–500; top) and two-dimensional plot (m/z values vs retention times;
bottom). Insets: TIC chromatograms (positive and negative ion modes) of blanks (extracellular media from control keratinocytes).
Copyright  2003 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2003; 38: 1160–1168

1164
G. Aldini et al.
Table 1. LC/ESI-MS and MS/MS analyses of HNE-conjugated metabolites in keratinocyte extracellular media
Positive ion modea
Negative ion modea
ESI-MS
[M C H]
(m/z)
ESI-MS
[M ꢀ H]
(m/z)
C
ꢀ
Retention
ESI-MS/MS
ESI-MS/MS
Compound
time (min.)
(m/z)
(m/z)
C
ꢀ
1
b
2.58 š 0.03
464
446 [M C H ꢀ H2O]
462
444 [M ꢀ H ꢀ H2O]
C
ꢀ
3
2
1
1
08 [GSH C H]
306 [GSH ꢀ H]
C
C
33 [Glu-cys C H]
79 [Cys-gly C H]
C
39 [nonenyl C H]
C
ꢀ
2
3
b
b
2.60 š 0.04
3.07 š 0.04
466
464
448 [M C H ꢀ H2O]
464
462
446 [M ꢀ H ꢀ H2O]
C
ꢀ
3
19 [MH C H ꢀ Glu ꢀ H2O]
335 [M ꢀ Glu]
ꢀ
2
2
72 [GSH ꢀ H ꢀ H2S]
ꢀ
54 [GSH ꢀ H ꢀ H2S ꢀ H2O]
C
ꢀ
446 [M C H ꢀ H2O]
444 [M ꢀ H ꢀ H2O]
C
ꢀ
3
2
1
1
08 [GSH + H]
306 [GSH ꢀ H]
C
C
33 [Glu-cys C H]
79 [Cys-gly C H]
C
39 [nonyl C H]
C
C
ꢀ
4
b,c
3.08 š 0.05
462
444 [M C H ꢀ H2O]
460
306 [GSH ꢀ H]
ꢀ
ꢀ
3
3
3
1
87 [M C H ꢀ Gly]
288 [GSH ꢀ H ꢀ H2O]
C
33 [MH C H ꢀ Glu]
272 [GSH ꢀ H ꢀ H2S]
C
ꢀ
15 [MH C H ꢀ Glu ꢀ H2O]
254 [GSH ꢀ H ꢀ H2S ꢀ H2O]
C
79 [Cys-gly C H]
a
Gly D —NHCH2COOH; Glu D —COCH2CH2CH(NH2)COOH; Cys D —SCH2CH(NH2)CO—; nonenyl D HNE ꢀ H2O.
1
00
100
80
60
40
20
0
NL: 1.38E6
TIC MS
NL: 1.21E6
TIC MS
8
6
4
2
0
0
0
0
0
1
1
00
3b
100
80
60
40
20
0
1
b
NL: 5.62E3
m/z = 463.5-464.5
8
6
4
2
0
0
0
0
0
NL: 6.30E4
m/z= 463.5-464.5
00
100
80
60
40
20
0
2b
2
b
NL: 1.17E6
m/z = 465.5-466.5
NL: 3.23E5
m/z = 465.5-466.5
8
6
4
2
0
0
0
0
0
1
00
100
80
60
40
20
0
4
b
NL: 5.09E5
m/z = 461.5-462.5
NL: 1.16E4
m/z = 461.5-462.5
8
6
4
2
0
0
0
0
0
0
1
2
3
4
5
6
0
1
2
3
4
5
6
Time (min)
Time (min)
Figure 5. LC/ESI-MS analysis (positive ion mode) of extracellular media from keratinocytes incubated with HNE (100 µM) before and
after reduction with NaBH4. Left: TIC and MCs (mass chromatograms relative to metabolites 1b–4b at m/z 464, 462, 466) before
NaBH4 reduction. Right: TIC and MCs (mass chromatograms relative to metabolites 1b–4b at m/z 464, 462, 466) after
NaBH4 reduction.
Copyright  2003 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2003; 38: 1160–1168

LC/ESI-MS/MS characterization of HNE–GSH adducts in keratinocytes 1165
indicates that after 120 min of incubation, 48.8 š 4.6% of
the incubated HNE dose is converted to the unconjugated
phase I metabolites HNA and DHN and that the oxida-
tion pathway is predominant (36.1 š 4.9% HNA formation).
Analysis of cell lysates indicates that both metabolites were
detectable only at the first observation time (5 min) and were
not retained inside the cell at 120 min (data not shown).
The extracellular media relative to the 2 h incubation time
were then analysed by LC/MS/MS to characterize those
metabolites that were not detected in the 200–400 nm UV
scan range.
a retro-Michael reaction, while the corresponding ion at
m/z 308 in the positive ion mode is detectable only in
C
the ESI-MS/MS [M C H] product ion spectra of 1b and
3b.
Hence we can assign to the four conjugated metabo-
lites the structures reported in Fig. 6. Compound 1b, S-
(4-hydroxynonanal-3-yl)glutathione (GS–HNE), arises from
the first metabolic step, a Michael addition between the
2
electrophilic C-3 of HNE and the thiol group of GSH,
with formation of an intermediate that can undergo an
intramolecular rearrangement to the five-membered cyclic
hemiacetal 1c. The minor metabolite 3b, isobaric with
LC/MS analysis of conjugated metabolites
In a first step, we analyzed the LC profile and the
fragmentation pattern of the synthesized GS–HNE adduct
OH
OH
O
(
positive ion mode): the total ion current (TIC) chromatogram
OH
shows only one peak at a retention time (R.T.) of 2.58 min,
the identity of which was confirmed by the expected
S
O
S
O
O
NH
O
2
O
NH
O
C
C
NH
2
[
M C H] ion at m/z 464, the corresponding [M C Na]
NH
HN
HO
HO
HN
HO
ion at m/z 486 and the CID product ions at m/z 446
[
HO
C
C
M C H ꢀ H
2
O] , m/z 308 [GSH C H] due to a typical
O
O
C
retro-Michael reaction, m/z 233 [glutamylcysteine C H] ,
1
b
2b
C
m/z 179 [cysteinylglycine C H] and m/z 139 due to the
O
O
nonenyl moiety (data not shown). This is in agreement
OH
O
8
with previous results reported by Boon et al. and Enoiu
1
2
S
S
et al. Figure 4 (left) shows the TIC chromatogram in the
positive ion mode of the 2 h extracellular medium. Two main
peaks with R.T. 2.58 and 3.05 min are clearly detectable, and
two-dimensional analysis (m/z values vs retention times)
O
O
O
NH
O
O
NH
O
2
2
NH
NH
HN
HO
HN
HO
HO
HO
C
reveals that the peak at 2.58 min contains two [M C H]
O
O
3b
4b
species at m/z 464 (1b) and m/z 466 (2b) accompanied by
C
the corresponding cationized species [M C Na] at m/z 486
OH
and 488. Also the peak with R.T. 3.08 min contains two
O
C
[
M C H] species at m/z 464 (3b) and m/z 462 (4b), and
C
shows the corresponding [M C Na] species at m/z 486 and
S
O
C
4
84. Hence compounds 1b and 3b giving rise to [M C H] at
O
HN
NH
2
NH
m/z 464, co-elute with 2b and 4b, respectively. The presence
in the extracellular medium of four metabolic species and
their molecular masses were confirmed by negative ion
detection (Fig. 4, right). The two-dimensional plot indicates
the following ionic species: m/z 462, corresponding to the
HO
O
HO
O
O
1c
OH
O
O
ꢀ
ꢀ
[
M ꢀ H] ion of 1b and 3b, m/z 464, the [M ꢀ H] of 2b,
S
S
ꢀ
O
and m/z 460, the [M ꢀ H] of 4b. The mass chromatograms
O
C
O
NH2
O
NH
2
(
MCs) relative to the [M C H] ions (Fig. 5, left) confirm
NH
NH
HN
HO
that 1b and 3b co-elute with 2b and 4b, respectively.
On the basis of these results, we can exclude for 3b the
structure of a GS–HNE isomer since under our LC conditions
no separation of diastereoisomers occurs (a single peak
working with the synthesized GS–HNE adduct). Structure
assignment was achieved by MS/MS analysis (positive and
HN
HO
HO
O
HO
O
O
O
1
b
4c
OH
O
ꢀ
C
negative ion modes) of the [M ꢀ H] and [M C H] ions
S
OH
O
of the four metabolic species. The fragmentation patterns
O
NH
2
NH
(
Table 1) clearly indicate conjugation with GSH through
HN
HO
HO
O
Michael addition of the nucleophilic sulfur of GSH to the
electrophilic C-3 of HNE, followed by oxidation-reduction
enzymatic processes. In particular, the negative ion ESI-
O
ꢀ
1d
MS/MS [M ꢀ H] product ion spectra contain a dominant
peak at m/z 306 corresponding to deprotonated GSH due
Figure 6. Structures of GSH–HNE conjugated adducts (phase
to the cleavage of the Michael adduct at C-3 of HNE via
II metabolites).
Copyright  2003 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2003; 38: 1160–1168

1166
G. Aldini et al.
1
b, can be tentatively assigned to S-(4-oxononan-1-ol-3-
complete disappearance of the peak with R.T. 3.08 min. In
addition, the MCs relative to the [M C H] ions at m/z
C
yl)glutathione, which is formed by oxidation of the sec-
ondary alcohol at C-4 and concomitant reduction of the
aldehydic function. This HNE metabolite has to our knowl-
edge never been identified in other biological systems.
The most abundant metabolic products, 2b and 4b, arise
from 1b, through a reduction of the aldehydic function
to [S-(1,4-dihydroxynonan-3-yl)glutathione] (GS–DHN) and
oxidation of the secondary alcohol at C-4, with formation
of [S-(4-oxononanal-3-yl)glutathione], respectively. Alterna-
462 and 464 showed no detectable signals, while a sig-
nificant increase of the ion current at m/z 466 (2b) is
observed. This unequivocally indicates total conversion of
keto/aldehydic moieties of metabolites 1b, 3b and 4b to
the 1,4-dihydroxy function. Total conversion was observed
already after 6 h of incubation at room temperature. The pos-
itive response to chemical reduction and the scanty evidence
in the literature for lactone reduction by NaBH
4
in aqueous
1
9,20
tively (Fig 6), the metabolite 4b (M
r
461) could be assigned
solution
seem to exclude the formation of 4c. Studies are
an HNA–lactone–GSH structure (4c). The latter structure
can be formed by two different pathways, namely alco-
hol oxidation of the hemiacetalic form 1c, as proposed by
in progress involving the synthesis of 4-oxo-trans-2-nonenal
and the corresponding GSH adduct for definitive structure
assignment.
1
8
Alary et al. and/or aldehyde oxidation of 1b to give 1d
8
(HNA–GSH), in equilibrium with the lactone structure.
Kinetics of metabolite formation
Since MS/MS fragmentation patterns do not discriminate
between 4b and 4c, and in an attempt to gain further
information and at the same time to confirm the pres-
ence of keto/aldehydic functions in the four metabolites,
the extracellular media from keratinocytes incubated for 2 h
with 100 µM HNE were submitted to chemical reduction
The next step was to explore the potential of the LC/MS
approach to study the transport of the GSH adducts through
the keratinocyte plasma membrane, by monitoring the time
course of their intracellular formation and excretion from
the cell. Analyses were done on both extracellular media
and cell lysates from keratinocytes incubated for different
time periods and the MCs (positive ion mode) relative to the
4
with NaBH and re-analyzed by LC/ESI-MS (positive ion
mode). The TIC chromatogram in Fig. 5 (right) indicates
C
[M C H] ion at m/z 464 (1b, 3b), 466 (2b) and 462 (4b) were
Intra-cellular
Extra-cellular
X104
X104
3
2
1
0
0
0
0
30
20
10
0
5
min
X104
X104
30
3
0
0
0
0
1
3
5 min
0 min
2
1
20
10
0
X104
X104
30
3
0
0
0
0
2
1
20
10
0
X104
X104
30
3
0
0
0
0
60 min
2
1
20
10
0
X104
X104
30
3
0
0
0
0
1
20 min
2
1
20
10
0
0
1
2
3
4
5
6
0
1
2
3
4
5
6
Time (min)
Time (min)
Figure 7. LC/ESI-MS detection of metabolite 2b in cell lysates (left) and in extracellular media (right). MCs of the ion at m/z 466 at
different incubation times.
Copyright  2003 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2003; 38: 1160–1168

LC/ESI-MS/MS characterization of HNE–GSH adducts in keratinocytes 1167
recorded. Figure 7 shows the representative MCs of the ion at
m/z 466 (2b): the formation of this metabolic species within
the cell is rapid, already detectable after 5 min of incubation
and characterized by an apparent accumulation (a steady-
state level) until 30 min, followed by a rapid decay and
a concomitant extrusion in the medium (Fig. 7, right). The
kinetics of formation and extrusion of the main metabolites
X106
5
4
3
2
1
0
2
b and 4b (Fig. 8) seem to indicate that export of GSH
conjugates into the extracellular medium is unidirectional.
These results, although they do not allow one to quantitate
the absolute amount of metabolites (analytical standards
are not available), are consistent with an accumulation of
metabolites against a concentration gradient, and suggest
that the process responsible for export is not a simple
diffusion but most likely involves an active transport
mechanism.
0
30
60
90
120
Time (min)
X106
5
Owing to the substantial involvement of GSH in HNE
metabolization, the intracellular levels of this detoxifying
agent were determined by a fluorescence-based assay.
The addition of HNE to keratinocytes resulted in a time-
dependent consumption of cellular GSH, to ¾26% within 2 h
of incubation, accompained by a concomitant production of
the GSH adducts. The formation of GSH adducts is rapid
4
3
2
1
0
(Fig. 7) and occurs strictly in parallel to the decrease in
GSH, indicating that HNE conjugation with GSH occurs via
a GSH S-transferase-catalyzed reaction. By considering the
basal and final levels of GSH (70.8 š 3.8 and 52.2 š 2.6 nmole
ꢀ
1
0
30
60
90
120
mg protein) and the amount of incubated HNE (100 µM,
ꢀ
1
Time (min)
corresponding to ¾35 nmol mg protein), we assume that
0–55% of the added HNE is metabolized through the
GSH conjugation pathway. The remaining HNE follows the
5
Figure 8. Time course of formation in keratinocytes and
extrusion in the extracellular medium of metabolites 4b (top)
and 2b (bottom). Values, expressed as arbitrary units (a.u) and
reported as means š S.D. of five independent experiments,
were calculated as MC peak areas in (ꢀ) cell lysates and (ꢁ)
extracellular media.
7
–11
metabolic fate observed in other biological systems,
with
formation of HNA and DHN as major and minor metabolites,
respectively. Figure 9 summarizes the metabolic fate of HNE
in keratinocytes.
OH
O
1a
Extra-cellular
Intra-cellular
OH
O
ALDH
OH
ADH
OH
Cell membrane
1
a
GSH
O
OH
GSH-T
O
OH
2a
3a
OH
S
O
O
OH
S
O
O
Glu-Cys-Gly
OH
OH
O
1b
S
S
S
Glu-Cys-Gly
Glu-Cys-Gly
Glu-Cys-Gly
Glu-Cys-Gly
2b
3b
4b
4c
Figure 9. Overall picture of HNE biotransformation (metabolic fate and pathways) in keratinocytes. GSH-T, glutathione
S-transferase; ADH, alcohol dehydrogenases; ALDH, aldehyde dehydrogenases.
Copyright  2003 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2003; 38: 1160–1168

1168
G. Aldini et al.
CONCLUSIONS
4. Terashi H, Izumi K, Rhodes LM, Marcelo CL. Human stratified
squamous epithelia differ in cellular fatty acid composition. J.
Dermatol. Sci. 2000; 24: 14.
The results of this study provide a mechanistic insight, until
now never demonstrated, into the capacity of keratinocytes
to detoxify the cytotoxic unsaturated aldehydes (HNE) gen-
erated endogenously as final products of lipid peroxidation
reactions (following, for example, UV exposure). This pro-
tective mechanism involves the biotransformation of HNE
through an integrated system of reductive, oxidative and
conjugative enzymes to highly hydrophilic and unreactive
free (HNA, DHN) and GSH-conjugated adducts, and their
extrusion from the cell, and explains the maintenance of the
viability of keratinocytes even in the presence of high con-
centrations of this cytotoxic agent. This, if extrapolated to an
in vivo situation, means that the human skin, when exposed
to the attack of electrophilic chemicals, is able to maintain
its integrity through the cooperation of aldehyde dehydro-
genases, alcohol dehydrogenases, GSH and GSH transferase
and suggests, although indirectly, that accumulation of car-
bonyl compounds in the photodamaged skin is caused by
a metabolic impairment of the defense mechanism (GSH
depletion and inactivation of enzyme systems).
5
. Sander CS, Chang H, Salzmann S, M u¨ ller CSL, Ekanayake-
Mudiyanselage S, Elsner P, Thiele JJ. Photoaging is associated
with protein oxidation in human skin in vivo. J. Invest. Dermatol.
2002; 118: 618.
6
. Tanaka N, Tajima S, Ishibashi A, Uchida K, Shigematsu T.
Immunohistochemical detection of lipid peroxidation products,
protein-bound acrolein and 4-hydroxynonenal protein adducts
in actinic elastosis of photodamaged skin. Arch. Dermatol. Res.
2001; 293: 363.
7
8
9
. Srivastava S, Dixit BL, Cai J, Sharma S, Hurst HE, Bhatnagar A,
Srivastava SK. Metabolism of lipid peroxidation product, 4-
hydroxynonenal (HNE) in rat erythrocytes: role of aldose
reductase. Free Rad. Biol. Med. 2000; 29: 642.
. Boon PJM, Marinho HS, Oosting R, Mulder GJ. Glutathione
conjugation of 4-hydroxy-trans-2,3-nonenal in the rat in vivo, the
isolated perfused liver and erythrocytes. Toxicol. Appl. Pharmacol.
1999; 159: 214.
. Laurent A, Perdu-Durand E, Alary J, Debrauwer L, Cravedi JP.
Metabolism of 4-hydroxynonenal, a cytotoxic product of lipid
peroxidation, in rat precision-cut liver slices. Toxicol. Lett. 2000;
114: 203.
1
0. Srivastava S, Conklin DJ, Si-Qi L, Prakash N, Boor PJ,
Srivastava SK, Bhatnagar A. Identification of biochemical
pathways for the metabolism of oxidized low-density
lipoprotein derived aldehyde-4-hydroxy-trans-2-nonenal in
vascular smooth muscle cells. Atherosclerosis 2001; 158: 339.
1. Ullrich O, Huser H, Ehrlich W, Grune T. Intracellular
metabolism of 4-hydroxynonenal in primary cultures of rabbit
synovial fibroblasts. Free Rad. Biol. Med. 1997; 22: 1153.
2. Enoiu M, Herber R, Wennig R, Marson R, Bodaud H, Leroy P,
Mitra N, Siest G, Wellman M. ꢀ-Glutamyltranspeptidase-
dependent metabolism of 4-hydroxynonenal–glutathione
conjugate. Arch. Biochem. Biophys. 2002; 397: 18.
3. Rees MS, van Kuijk FJGM, Siakotos AN, Mundy BP. Improved
synthesis of various isotope labelled 4-hydroxyalkenals and
peroxidation intermediates. Synth. Commun. 1995; 25: 3225.
4. Alary J, Bravais F, Cravedi JP, Debrauwer L, Rao D, Bories G.
Mercapturic acid conjugates as urinary end metabolites of the
lipid peroxidation product 4-hydroxy-2-nonenal in the rat. Chem.
Res. Toxicol. 1995; 8: 34.
5. Carini M, Aldini G, Piccone M, Maffei Facino R. Fluorescent
probes as markers of oxidative stress in keratinocyte cell lines
following UVB exposure. Farmaco 2000; 55: 526.
From an analytical viewpoint, it is worth noting that
profiling of the phase II metabolites in keratinocytes has
been performed by direct analysis of the crude biological
matrix. ESI-MS/MS in positive and negative ion modes
coupled to a microbore LC system (employing a very short
gradient elution) proved to be highly suited not only to
direct characterization of the GSH-derived metabolites of
HNE, but also to rapid monitoring of their formation in
keratinocytes and export outside the cell. In addition, the
structural and kinetic information generated by the proposed
LC/ESI-MS/MS approach has been obtained in a very short
time and by working with micromolar concentrations of the
1
1
1
1
ꢀ
1
toxin (35 nmol mg protein). The LC/MS approach, only
8
,12
recently applied to the detection of HNE metabolites,
has
several advantages over the conventional approach used
in the study of HNE metabolism involving radioactively
1
3
labelled [ H]HNE, since the synthesis, purification and
handling of radioactive material and waste are expensive
and time consumimg, and radiation is a potential health risk.
1
6. De Clerck LS, Bridts CH, Mertens AM, Moens MM, Stevens WJ.
Use of fluorescent dyes in the determination of adherence
of human leucocytes to endothelial cells and the effect of
fluorochromes on cellular function. J. Immunol. Methods 1994;
Acknowledgements
172: 115.
Financial support from FIRST 2002 (Fondo Interno per la Ricerca
Scientifica e Tecnologica) is gratefully acknowledged. We thank Dr
Silvana Casati (Dipartimento di Scienze Precliniche, University of
Milan) for skilfull assistance with the chemical syntheses.
1
1
1
7. Markwell MA, Haas SM, Tolbert NE, Bieber LL. Protein
determination in membrane and lipoprotein samples: manual
and automated procedures. Methods Enzymol. 1981; 72: 296.
8. Alary J, Fernandez Y, Debrauwer L, Perdu E, Gu e´ raud F.
Identification of intermediate pathways of 4-hydroxynonenal
metabolism in the rat. Chem. Rese. Toxicol. 2003; 16: 320.
9. Lundt I, Madsen R. Deoxyiminoalditols from aldonolactones.
I. Preparation of 1,4-dideoxy-1,4-iminohexitols with D- and L-
allo and L-talo configurations: potential glycosidase inhibitors.
Synthesis 1993; 7: 714.
REFERENCES
1. Thiele J, Elsner P (eds). Oxidants and Antioxidants in Cutaneous
Biology. Karger: Basel, 2001.
2
. Esterbauer H, Schaur RJ, Zollner H. Chemistry and biochemistry
of 4-hydroxynonenal, malonaldehyde and related aldehydes.
Free Rad. Biol. Med. 1991; 7: 333.
20. Lundt I, Madsen R. Deoxyiminoalditols from aldonolactones.
II. Preparation of 1,4-dideoxy-1,4-iminohexitols with D- and L-
galacto and D- and L-ido configurations: potential glycosidase
inhibitors. Synthesis 1993; 7: 720.
3. Poli G, Schaur J. 4-Hydroxynonenal in the pathomechanisms of
oxidative stress. IUBMB Life 2000; 50: 1.
Copyright  2003 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2003; 38: 1160–1168