Special issue: review

Received: 15 July 2014, Revised: 2 September 2014, Accepted: 12 September 2014 Published online in Wiley Online Library: 30 October 2014

(wileyonlinelibrary.com) DOI 10.1002/bmc.3361

Quantification of malondialdehyde by HPLC-FL

– application to various biological samples
Ana-Marija Domijana*, Jovica Ralićb, Sandra Radić Brkanacc,
Lada Rumorad and Tihana Žanić-Grubišićd
ABSTRACT: Malondialdehyde (MDA) is stabile product of lipid peroxidation (LPO), and therefore MDA is frequently used as a
biomarker of LPO. To determine MDA level in various biological samples (human plasma, fish liver tissue and cells in culture),
we used an HPLC method with fluorescent detection based on 2-thiobarbituric acid (TBA) assay. The method was validated by
the use of spiked pooled plasma samples. In tested concentration range (0.15–3.0 μmol/L) the method was linear
(R2 = 0.9963), the between-day variability (coefficient of variations, CVs) was between 4.7 and 7.6%, the within-day variability
CVs was between 2.6 and 6.4% and recovery was between 91.2 and 107.6%. The level of MDA in human plasma (healthy male,
non-smokers, 46.3 ± 4.7 years; N = 38) was 2.2 ± 1.4 μmol/L; that in liver tissue of common carp (Cyprinus carpio; N = 12) was
0.02 ± 0.004 μmol/g tissue, and in cultured cells (human laryngeal carcinoma cells; N = 10) it was 0.18 ± 0.02 nmol/mg proteins.
The HPLC-FL method is rapid, accurate and reliable to follow the extent of LPO in various biological samples, particularly in
samples in which a low level of MDA is expected, such as cells in culture. Owing to the rapid analytical process and run time, it
can be used for routine analysis of MDA in clinical laboratory. Copyright © 2014 John Wiley & Sons, Ltd.

Keywords: lipid peroxidation; thiobarbituric acid assay; plasma; cells in culture; liver homogenate

Introduction
Lipid peroxidation (LPO) is chain reaction initiated by the
attack of reactive oxygen species (ROS) and reactive nitrogen
species (RNS) on polyunsaturated fatty acid (PUFA) residues
of phospholipids, leading to cell membrane damage and
consequently to cell death (Halliwell and Chirico, 1993).
Primary products of LPO are unstable hydroperoxides that
decompose to various secondary products, among which are
stabile aldehydes, malondialdehyde (MDA) and 4-hydroxynonenal
(HNE). Aldehydes, MDA and HNE, owing to their high reactivity
cross-link with proteins (Esterbauer and Cheeseman, 1990;
Esterbauer et al., 1991) and DNA (Marnett, 1999; Nair et al.,
2007), and are considered toxic and mutagenic. Higher levels of
LPO are detected in human diseases such as atherosclerosis,
some cancers and neurodegenerative disorders (Rosenblat et al.,
2002; Delimaris et al., 2007; Ou et al., 2002; Hashimoto et al.,
2003), hence there is interest in following the extent of LPO in
body fluids (plasma and urine) or in particular tissue.
Several biomarkers of LPO have been established, such as
8-iso-prostaglandin F2α, MDA, HNE (Syslova et al., 2009) and
exhaled breath ethane (Kazui et al., 1992; Cagini et al., 2011).
Because MDA is the most abundant and stabile individual
aldehydic product of LPO and because of the simplicity of the
methods for its determination, MDA is a widely used biomarker
of LPO (Marnett, 1999).
The level of MDA in biological samples is usually assessed by
2-thiobarbituric acid (TBA) assay. In the reaction MDA from a
biological sample is derivatized with TBA, forming an MDA–TBA2
adduct, a red complex (Fig. 1). The level of MDA–TBA2 adduct can
be easily determined spectrophotometrically or spectrofluorimetrically.
However, the method has been criticized for low sensitivity and selec-
react with TBA (Pilz et al., 2000; del Rio et al., 2005). Also, some suspicion
of artifactual generation of MDA during the assay has been raised
(Steghens et al., 2001; Sim et al., 2003). Despite all the mentioned prob-
lems, TBA assay is a still widely used method to monitor the level of
LPO in biological samples. Therefore, the aim of this study was to
optimize an HPLC method based on TBA assay that is simple yet
accurate and sensitive and can be applied to different biological
samples such as plasma, tissue and cells in culture.
Experimental
Chemicals and standard preparation
Disodium hydrogen phosphate (Na2HPO4), ethylenediaminetetraacetic
acid (EDTA), o-phosphoric acid (o-H3PO4), potassium chloride (KCl),
* Correspondence to: A.-M. Domijan, University of Zagreb, Faculty of Pharmacy
and Biochemistry, Department of Pharmaceutical Botany, Zagreb, Croatia.
Email: adomijan@pharma.hr
a
University of Zagreb, Faculty of Pharmacy and Biochemistry, Department of
Pharmaceutical Botany, Zagreb, Croatia
b
Galapagos Research Center, Zagreb, Croatia
c
University of Zagreb, Faculty of Science, Department of Biology, Zagreb,
Croatia
d
University of Zagreb, Faculty of Pharmacy and Biochemistry, Department of
Medical Biochemistry and Hematology, Zagreb, Croatia
Abbreviations used: BHT, butylated hydroxytoluene; DAN, diaminonaphthalene;
DNPH, 2,4-dinitrophenylhydrazine; HNE, 4-hydroxynonenal; LPO, lipid peroxida-
tion; MDA, malondialdehyde; PUFA, polyunsaturated fatty acid; ROS, reactive
41

tivity since several MDA-unrelated species from biological samples can oxygen species; RNS, reactive nitrogen species; TBA, 2-thiobarbituric acid.

Biomed. Chromatogr. 2015; 29: 41–46 Copyright © 2014 John Wiley & Sons, Ltd.

Figure 1. Chemical reaction of one molecule of malondialdehyde
(MDA) with two molecules of 2-thiobarbituric acid (TBA) and formation
of condensation product (MDA–TBA2 adduct).
potassium dihydrogen phosphate (KH2PO4), potassium hydroxide (KOH),
sodium chloride (NaCl), methanol and TBA were from Kemika, Zagreb,
Croatia. 1,1,3,3-Tetraethoxypropane (malondialdehyde bis-dimethylacetal)
as MDA standard was procured from Sigma (St Louis, MO, USA). Methanol,
used for the mobile phase, was HPLC grade, and all other chemicals were
p.a. grade.
MDA stock solution (3.0 mmol/L) was prepared by diluting MDA standard
in MilliQ water. MDA stock solution was aliquoted and stored at 20°C.
Working standards in MilliQ water (ranging from 0.15 to 3.0 μmol/L) were
prepared fresh daily.
HPLC equipment and chromatographic conditions
The HPLC consisted of gradient pump (model 64, Knauer, Berlin, Germany)
and fluorescent detector (FL; F 1000 Hitachi Merck, Darmstad, Germany).
Separation was performed on an analytical column, 125.0 × 4.0 mm, particle
size 5 μm equipped with a guard column (4.0 × 4.0 mm, 5 μm), both
LiChrospher RP18 (Merck, Darmstad, Germany). A manual injector
(Rheyodine, Rohnert Park, CA, USA) with a 100 μL loop was used. Chromato-
graphic control, data collection and processing were carried out using
Eurochrom 2000 software, basic edition (Knauer, Berlin, Germany).
For mobile phase 50 mmol/L potassium dihydrogen phosphate was
prepared in MilliQ water and pH of potassium dihydrogen phosphate
was adjusted with KOH (5 mol/L) to 6.8. Mobile phase consisted of
50 mmol/L potassium dihydrogen phosphate (pH 6.8) and methanol
(60:40; v/v). Before use, the mobile phase was filtered by vacuum
through 0.22 μm cellulose acetate filter. The flow rate was set at
1.0 mL/min. The FL detector wavelengths were set for excitation (λex) at
527 nm and emission (λem) at 551 nm.
Sample preparation
The biological samples obtained were human plasma (healthy male vol-
unteers, non-smokers; N = 38), fish liver tissue (common carp, Cyprinus
carpio; N = 12) and cells in culture (human laryngeal carcinoma cells,
HEp-2 cells; N = 10). Collected samples were analyzed immediately (cells
in culture) or not later than a month after collection (human plasma and
supernatants of fish liver homogenate). Until analyzed, samples of
human plasma and supernatants of fish liver homogenate were stored
at 80°C. All samples were obtained from on-going studies that were
approved by the local ethical committees. Before proceeding with TBA
assay, each biological sample underwent a specific sample preparation
procedure as listed below.
Plasma samples were diluted 10 times with destilled water.
Immediately after removal, carp liver tissues were frozen in liquid
nitrogen and, following addition of ice-cold potassium phosphate buffer
(50 mmol/L; pH 7.4) and EDTA (3 mmol/L), homogenized (TissueLyser,
Qiagen) for 2 min at 30,000 Hz to obtain 10% tissue homogenates. After-
wards liver tissue homogenates were centrifuged (20,000g for 15 min at
4°C) and supernatants were collected and stored (at 80°C) until ana-
lyzed. Before assessing MDA, tissue supernatants were diluted 10 times
with destilled water.
6
After harvesting, HEp-2 cells (2 × 10 cells per well) were washed in
phosphate-buffered saline (PBS) buffer (137 mmol/L sodium chloride,
2.7 mmol/L potassium chloride, 10 mmol/L disodium hydrogen phos-
42
phate and 1.8 mmol/L potassium dihydrogen phosphate; pH 7.4) and
wileyonlinelibrary.com/journal/bmc Copyright © 2014 John Wiley & Sons, Ltd.
A.-M. Domijan et al.
centrifuged (1000 g, 4°C, 4 min) to obtain cell pellets. Resuspended cell
pellets (cold PBS buffer, 70 μL) were lysed by soniciation (three times
for 3 s; cells were kept on ice), centrifuged again (1000g, 4°C, 4 min)
and supernatants were collected for MDA analysis. In the cell superna-
tants the protein level was determined according to Bradford’s (1976)
assay.
To 50 μL of destilled water (blank), 50 μL of standard, 50 μL of sample
(diluted plasma or diluted tissue supernatant) or 60 μL of cells’ lysate
supernatant in a 1.5 mL Eppendorf tube, 400 μL of o-phosphoric acid
(0.1 %) and 100 μL of TBA (0.6%) were added. The tubes were mixed
and afterwards heated at 90°C for 30 min in a temperature-controlled
heating block. The reaction was stopped by placing samples on ice. Samples
were kept on ice until analysis and 100 μL of samples were injected into the
HPLC column.
Statistical analysis
Results are expressed as mean ± standard deviation. Statistical analysis of
the data was performed using Student’s t-test for independent samples
with statistical program Statistica 8.0 (Stat Soft Ltd, Bedford, UK).
Results and discussion
TBA assay
MDA is the most abundant aldehydic breakdown product of
oxidative degradation of PUFAs and is frequently used as an
indicator of LPO extent (Esterbauer et al., 1991; Seljeskog et al.,
2006). Plasma MDA is even suggested as a marker of overall
body LPO (Nielsen et al., 1997; Drury et al., 1997). A
multilaboratory study in which several biomarkers of oxidative
stress in plasma and urine in a rat model were checked con-
firmed MDA to be a good biomarker of LPO (Kadiiska et al.,
2005). To assess MDA level, apart from TBA, several other
derivatizing agents were tested: 2,4-dinitrophenylhydrazine
(DNPH) and diaminonaphthalene (DAN) (Pilz et al., 2000;
Steghens et al., 2001). However, the literature data clearly dem-
onstrate that methods employing DNPH or DAN are not superior
to TBA assay. TBA assay is quicker and simpler compared with
the mentioned methods and, according to Mendes et al.
(2009), TBA assay is equally sensitive to MDA–DNPH method.
Although TBA assay is a widely used method for quantification
of MDA level, the results of TBA assay often vary and no agreement
on method procedure has been reached. We analyzed each step of
the TBA assay in order to identify potential problems and to set up
a reliable method for quantification of MDA level that is applicable
to various biological samples.
Selectivity is the main shortcoming of TBA assay since TBA,
except with MDA, reacts with several other components of biological
samples (such as certain carbohydrates, amino acids or bilirubin),
thus leading to overestimation of MDA level (Pilz et al., 2000; del
Rio et al., 2005). In the above-mentioned multilaboratory study, it is
suggested that MDA should be assessed with GC-MS since results
obtained with GC-MS are superior to spectrophotometric determina-
tion of MDA (Kadiiska et al., 2005). The selectivity and sensitivity of
the method can also be achieved by HPLC with either UV–vis or fluo-
rescence detection (Agrawal and Chase, 2002).
HPLC conditions
Detector wavelengths were set at λex 527 nm and λem 551 nm
(Agarwal and Chase, 2002), and to separate MDA–TBA2 adduct
reverse-phase column, C18, was used. In the literature several
Biomed. Chromatogr. 2015; 29: 41–46
Quantification of MDA by HPLC-FL

1.2
mobile phases have been suggested and we tested: 50 mmol/L
potassium dihydrogen phosphate (pH 6.8) and methanol (60:40; 1
v/v; Agarwal and Chase, 2002); 10 mmol/L potassium dihydrogen

Peak area (m AU)

phosphate (pH 6.8) and methanol (60:40; v/v; Nielsen et al., 1997); 0.8
and 50% methanol (Grotto et al., 2007). A sharp, symmetrical peak
of MDA–TBA2 adduct was only obtained using 50 mmol/L potas- 0.6
sium dihydrogen phosphate buffer and methanol (60:40; v/v).
0.4
Under HPLC conditions, isocratic elution of 50 mmol/L potassium
dihydrogen phosphate buffer and methanol (60:40; v/v) at a flow- 0.2
rate set at 1 mL/min, the retention time of MDA–TBA2 adduct was
~1.5 min and the whole analysis time, including column washing, 0
0 10 20 30 40 50 60 70 80 90 100
was 5 min, indicating that this conditions will allow large number
of samples to be analyzed daily. incubation period (min)

Figure 2. Derivatization rate of malondialdehyde standard in water

Sample preparation with TBA when 0.1% or 1% o-phosphoric acid was used.

A critical step in TBA assay is sample preparation procedure

(derivatization of MDA with TBA), which involves heating samples Therefore, to avoid possible artifactual formation of product, we
under acidic conditions, referred to as ‘stress conditions’; usually standardized the sample preparation step (derivatization) as incuba-
samples are heated at 100°C for 1 h at low pH (Steghens et al., tion of the sample with TBA together with 0.1% phosphoric acid at
2001; Mendes et al., 2009). This step is unavoidable since the reac- 90°C for 30 min.
tion kinetics depends on the temperature, pH and concentration
of TBA (Nielsen et al., 1997). Additionally, acid used in the sample
Validation of HPLC-FL method
preparation step affect the MDA level detected (Seljeskog et al.,
2006). This is due to the fact that in biological samples MDA is The method was validated by multiple analyses of spiked pooled
mainly bound to macromolecules (proteins or DNA) and only a plasma standards (Table 1). The calibration curve of spiked
small amount of MDA is free (Esterbauer et al., 1991; Pilz et al., plasma samples in the concentration range tested (0.15–3.0 μmol/L)
2000; Agarwal and Chase, 2002; Grotto et al., 2007; Nair et al., was linear (R2 was 0.9963) and there was highly linear relationship
2007). The treatment of biological samples with an acid liberates between peak areas of MDA standards in water and in plasma. The
bound MDA and allows detection of total MDA (free and bound). between-day variability (calculated as coefficient of variations, CVs)
Such treatment of samples with an acid is considered acidic was between 4.7 and 7.6% and the within-day variability CVs were
hydrolysis of sample (Steghens et al., 2001; del Rio et al., 2005). between 2.6 and 6.4%. The recovery was established to be between
In some studies, to liberate MDA, pre-treatment of samples with 91.2 and 107.6%. Before assessing MDA in liver tissue homogenate
alkaline solution (alkaline hydrolysis) is introduced (Pilz et al., and cells in culture, the respective pooled sample was prepared,
2000; Sim et al., 2003). However, alkaline hydrolysis is an addi- spiked with MDA standard and tested for linearity and reproducibility.
tional step in the sample preparation procedure, while in TBA
assay acidic hydrolysis occurs during the derivatization step (del
Quantification of MDA in biological samples
Rio et al., 2005).
During the sample preparation step (heating under acidic Quantification of MDA in plasma samples. Total MDA level
conditions) decomposition of lipid hydroperoxides and de novo was assessed in samples of human plasma collected from
formation of MDA can occur, which in turn leads to overestima- healthy male volunteers (N = 38), who were aged 38–58 years
tion of results (Steghens et al., 2001; Mendes et al., 2009). Some (average 46.3 ± 4.7 years) and non-smokers. Total MDA level in
authors, to stabilize samples during sample preparation step and plasma was from 0.72 to 4.51 μmol/L (2.2 ± 1.4 μmol/L) and only
to prevent further LPO, employ a chain-breaking antioxidant, one sample had a total MDA level of 7.61 μmol/L. Reference
butylated hydroxytoluene (BHT). We tested the influence of values for human plasma MDA level have not been set, and in
BHT on the amount of MDA–TBA2 adduct and noticed that the literature different values for healthy subjects are reported
BHT included in the sample preparation procedure did not influ- (Table 2). The result obtained in this study is in agreement with
ence the amount of product detected. Since Pilz et al. (2000) and Pilz et al. (2000), who reported a mean total MDA plasma value
Grotto et al. (2007) observed that addition of BHT to the samples
reduced the amount of product, like Pilz et al. (2000) and Grotto
et al. (2007), we omitted the use of BHT in sample preparation Table 1. Validation data (variability and recovery) of spiked
procedure. pooled plasma sample with known malondialdehyde (MDA)
To minimize ‘stress conditions’ and possible overestimation of concentration
results, we tested the effect of two o-phosphoric acid dilutions
(1%, which is usually used in TBA assay, and 0.1%) on the MDA Between-day Within-day Recovery
amount of product formed. MDA standard (rather than biological concentration coefficient of coefficient of (%; range)
sample) was treated with TBA and with 0.1% or 1% o-phosphoric (μmol/L) variation (%) variation (%)
acid, heated at 90°C, and the amount of product was monitored
0.15 7.6 6.4 91.2–104.8
over the period of 90 min. As shown in Fig. 2, the reaction was
0.6 5.9 4.3 95.6–107.6
complete after 30 min at 90°C and longer incubation did not yield
1.5 4.7 2.6 98.3–104.2
a higher product level; in addition, importantly, there was no dif-
3.0 6.3 3.0 96–105.7
43

ference in the amount of product gained with either acid used.

Biomed. Chromatogr. 2015; 29: 41–46 Copyright © 2014 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/bmc

Table 2. Studiesa reporting level of total MDA in human plasma
Specification of the method
(hydrolysis/derivatization/extraction)
Derivatization with TBAb (60 min, 100°C);
extraction to NaOH + methanol (1:12)
Alkaline hydrolysis (30 min, 60°C) + BHTc;
derivatization with TBA (30 min, 95°C);
extraction to n-butanol
Alkaline hydrolysis (30 min, 60°C);
derivatization with DNPHd (10 min, room temp);
extraction to hexane
Derivatization with DANe (180 min, 37°C)
Derivatization with TBA (60 min, 100°C) + BHT;
extraction to n-butanol
Alkaline hydrolysis (60 min, 60°C);
derivatization with DNPH (10 min, room temperature),
extraction to hexane
Alkaline hydrolysis (30 min, 60°C);
derivatization with TBA (45 min, 90°C);
extraction to n-butanol
Derivatization with TBA (30 min, 90°C)
a
Studies that employed HPLC are included.
b
TBA, thiobarbituric acid;
c
BHT, butylated hydroxytoluene;
d
DNPH, 2,4-dinitrophenylhydrazine;
e
DAN, diaminonaphthalene.
in healthy male volunteers (N = 12; aged 18–30 years; non-
smokers) of 2.16 ± 0.29 μmol/L. It is important to emphasize that
Pilz et al. (2000) used a different method for quantification of to-
tal plasma MDA level from that described here. MDA was
derivatized with DNPH; to liberate bound MDA, alkaline hydroly-
sis of samples was employed (1 mol/L NaOH; 60°C; 30 min) and
similarly they omitted BHT in the sample preparation step. This in-
dicates that different methods can obtain similar results.
In several other studies lower values for mean total MDA in
plasma of healthy subjects are reported (Steghens et al., 2001;
Agarwal and Chase, 2002). In the study by Agarwal and Chase
(2002), plasma total MDA level in healthy male volunteers
(N = 10; age 53 ± 14 years) was 0.69 ± 0.13 μmol/L. In that study
MDA was derivatized with TBA, in the sample preparation step
BHT was included, and before injection into HPLC, product
(MDA–TBA2 adduct) was extracted to n-butanol. The result of
Agarwal and Chase (2002) may indicate that additional steps
(such as BHT and extraction to n-butanol) lead to lower MDA
levels. Even lower values for total MDA in plasma (0.138
± 0.028 μmol/L) in healthy young men (N = 19; aged 21–37 years)
have been reported by Steghens et al. (2001), but in that study
MDA was derivatized with DAN.
On the other hand, several studies found higher values for
mean total MDA plasma level (Sim et al., 2003; Grotto et al.,
2007). Grotto et al. (2007) reported in elderly men (N = 20; 70
± 10 years; non-smokers) a mean total MDA plasma level of 4.6
± 0.95 μmol/L. The authors attributed the higher value to alkaline
hydrolysis (3 mol/L NaOH; 60°C; 30 min), claiming that they
achieved complete release of MDA bound to proteins; in that
44
study extraction to n-butanol was also included. Sim et al. (2003)
wileyonlinelibrary.com/journal/bmc Copyright © 2014 John Wiley & Sons, Ltd.
A.-M. Domijan et al.
MDA level (μmol/L) Reference
0.41–1.29 (men; n = 107) Nielsen et al. (1997)
0.33–1.22 (women; n = 106);
0.025–0.957 fractals
1.54 ± 0.72 (n = 16) Hong et al. (2000)
2.16 ± 0.29 (men; n = 12) Pilz et al. (2000)
0.162 ± 0.05 (women; n = 79) Steghens et al. (2001)
0.138 ± 0.03 (men; n = 19)
0.69 ± 0.13 (men; n = 10) Agarwal and Chase (2002)
13.8 ± 1.32 (n = 20) Sim et al. (2003)
4.60 ± 0.95 (men; n = 20) Grotto et al. (2007)
4.45 ± 0.81 (women, n = 45)
2.2 ± 1.4 (men; n = 38) This study
found in healthy volunteers (N = 20; 11 males and nine females;
aged 17–75; non-smokers) a mean plasma MDA value that was
even higher, 13.8 ± 1.32 μmol/L; the authors explained this higher
value as owing to the longer incubation of samples in the alkaline
hydrolysis step (1 mol/L NaOH, 1 h 60°C) and the dilution of
plasma samples. However, both procedures (Sim et al., 2003;
Grotto et al., 2007), owing to excessive sample preparation steps,
could lead to overestimation of MDA level.
In order to minimize the sample preparation step and pos-
sible overestimation of MDA, we omitted alkaline hydrolysis
and extraction to n-butanol. That in turn simplified the
method and no interferences were observed (chromato-
grams of plasma samples showed no interfering peaks). We
believe that our samples underwent complete release of
bound MDA owing to the dilution of plasma samples. Sim
et al. (2003) noted that dilution of plasma samples leads to
better recovery of MDA, that is, higher MDA levels are
measured if sample is more diluted in the reaction mixture.
They suggested that dilution can influence the accessibility of
bound MDA to hydrolysis or that, by diluting plasma, the con-
centrations of some inhibitors of hydrolysis normally present
in plasma are lower. In the procedure described here plasma
samples were diluted 10 times prior to derivatization/acidic
hydrolysis and it could be that dilution of plasma samples
assisted in the release of bound MDA and allowed assessment
of total MDA.
The influence of some other factors (age, gender and
smoking) on total MDA level in plasma should be mentioned.
According to the literature, total MDA level increases with age
since aging is related to more pronounced oxidative stress
Biomed. Chromatogr. 2015; 29: 41–46
Quantification of MDA by HPLC-FL
(Steghens et al., 2001). However, Nielsen et al. (1997) found no
effect of age on plasma MDA level. On the other hand, the same
authors found that men have higher plasma MDA level com-
pared with women and that smokers have higher plasma MDA
level than non-smokers (Nielsen et al., 1997). Additionally, the
impact of anticoagulant used has an effect on results; lower
values were reported when EDTA was used, probably owing to
the chelation of iron by EDTA, which limits further in vitro perox-
idation (Nielsen et al., 1997; Seljeskog et al., 2006). All of this, to-
gether with differences in sample preparation procedures, can
explain the lack of agreement on reference values for MDA in
human plasma.
Quantification of MDA in liver tissue. In supernatant of
10% carp liver homogenate (N = 20), MDA level was
1.54–2.70 μmol/L (2.16 ± 0.37 μmol/L) or 0.015–0.027 μmol/g
tissue (0.02 ± 0.004 μmol/g tissue). According to the literature,
MDA level in fish liver depends on the fish species. Sanz et al.
(2013) reported that different fish species such as Iberian chub
(Squalius pyrenaicus), common carp, gold fish (Carassius
auratus), Andalusian barbel (Luciobarbus sclateri) and rainbow
trout (Oncorhynchus mykiss) have different liver MDA levels.
The highest value of MDA in liver tissue (around 175 nmol/g
tissue) was detected in Iberian chub and the lowest in Andalusian
barbel (around 55 nmol/g tissue; or 3 times lower than in
Iberian chub). Additionally, water temperature and behavioral
activity of fish can, by increasing metabolic rate, affect level
of MDA. The effect of temperature and vitamin C supplemen-
tation on level of MDA in fish liver was studied on thornfish
(Terapon jarbua) cultured at 28, 32 and 36°C (Chien and
Hwang, 2001). The MDA level in liver of fish cultured at 28
and 32°C was around 0.9 μmol/g tissue, while in liver of fish
cultured at 36°C, the MDA level was 2 times higher, around
1.8 μmol/g tissue. The lowest MDA level (under 0.3 μmol/g
tissue) was detected in fish cultured at 28°C and fed a diet
supplemented with vitamin C. Together this indicates that
environmental as well as species-specific physiological
factors influence the level of MDA in fish liver tissue.
Therefore, to obtain reliable results for MDA level in
experimental animal tissues, animals should be kept in
standardized conditions and a well-defined control group
should be included.
Quantification of MDA in cultured cells. The problem with
cells in culture is a low MDA level expected and only a small
sample volume that can be gained. To overcome low MDA
levels in samples, we increased the volume of sample to
60 μL. In the cell lysate (HEp-2 cells; N = 10) MDA level was
from 0.3 to 0.4 μmol/L (0.37 ± 0.06 μmol/L). Since the number
of cells in plate wells can vary and this affects MDA level, the
MDA concentration for cells in culture should be normalized
on protein level in the same sample. Thus, in tested cultured
cells MDA level normalized on concentration of proteins was
0.16–0.20 nmol/mg proteins (0.18 ± 0.02 nmol/mg proteins).
These results indicate that the described method can be
used to follow the extent of LPO in experiments on cells
in culture.
Conclusion
Here we described an HPLC-FL method based on TBA assay that
was successfully applied to establish MDA level in plasma, tissue
and cells in culture. By omitting an extensive sample preparation
Biomed. Chromatogr. 2015; 29: 41–46 Copyright © 2014 John Wiley & Sons, Ltd.
procedure the method is accelerated and the problem of artifac-
tual peroxidation of samples is overcome. We showed that the
method is relevant to samples that contain low levels of MDA,
such as cells in culture. Owing to the rapid analytical process
and run time, the method can be use for routine analysis of
MDA in clinical laboratory.
Acknowledgments
We wish to thank the Department of Biochemistry and Molecular
Biology at the Zagreb University Faculty of Pharmacy and Bio-
chemistry for support in the experimental part of this study and
Professor M. Osmak’s group for making available cell samples.
References
Agarwal R and Chase SD. Rapid, fluorimetric–liquid chromatographic de-
termination of malondialdehyde in biological samples. Journal of
Chromatography B 2002; 775: 121–126.
Bradford MM. A rapid and sensitive method for the quantitation of mi-
crogram quantities of protein utilizing the principle of protein-dye
binding. Analytical Biochemistry 1976; 72: 248–254.
Cagini C, Giordanelli A, Fiore T, Giardinieri R, Malici B, De Medio GE, Pelli
MA, De Bellis F and Capodicasa F. Study of ethane level in exhaled
breath in patients with age-related macular degeneration: prelimi-
nary study. Ophthalmic Research 2011; 46: 141–144.
Chien L-T and Hwang D-F. Effects of thermal stress and vitamin C on lipid
peroxidation and fatty acid composition in the liver of thornfish
Terapon jarbua. Comparative Biochemistry and Physiology Part B
2001; 128: 91–97.
Delimaris I, Faviou E, Antonakos G, Stathopoulou E, Zachari A and
Dionyssiou-Asteriou A. Oxidized LDL, serum oxidizability and serum
lipid levels in patients with breast or ovarian cancer. Clinical Biochem-
istry 2007; 40: 1129–1134.
Del Rio D, Stewart AJ and Pellegrini N. A review of recent studies on
malondialdehyde as toxic molecule and biological marker of oxida-
tive stress. Nutrition, Metabolism, and Cardiovascular Diseases 2005;
15: 316–328.
Drury JA, Nycyk JA and Cooke RWI. Comparison of urinary and plasma
malondialdehyde in preterm infants. Clinica Chimica Acta 1997; 263:
177–185.
Esterbauer H and Cheeseman KH. Determination of aldehydic lipid-
peroxidation products: malondialdehyde and 4-hydroxynonenal.
Methods in Enzymology 1990; 186: 407–421.
Esterbauer H, Schaur RJ and Zollner H. Chemistry and biochemistry of 4-
hydroxynonenal, malonaldehyde and related aldehydes. Free Radical
Biology and Medicine 1991; 11: 81–128.
Grotto D, Santa Maria LD, Boeira S, Valentini J, Charao MF, Moro AM,
Nascimento PC, Pomblum VJ and Garcia SC. Rapid quantification of
malondialdehyde in plasma by high performance liquid
chromatography–visible detection. Journal of Pharmaceutical and
Biomedical Analysis 2007; 43: 619–624.
Halliwell B and Chirico S. Lipid peroxidation: its mechanism, measure-
ment, and significance. American Journal of Clinical Nutrition 1993;
57: 715S–725S.
Hashimoto M, Rockenstein E, Crews L and Masliah E. Role of protein ag-
gregation in mitochondrial dysfunction and neurodegeneration in
Alzheimer’s and Parkinson’s diseases. NeuroMolecular Medicine 2003;
4: 21–35.
Hong Y-L, Yeh S-L, Chang C-Y and Hu M-L. Total plasma malondialdehyde
levels in 16 Taiwanese College students determined by various thio-
barbituric acid tests and an improved high-performance liquid
chromatography-based method. Clinical Biochemistry 2000; 33: 619–625.
Kadiiska MB, Gladen BC, Baird DD, Germolec D, Graham LB, Parker CE,
et al. Biomarkers of oxidative stress study II. Are oxidative products
of lipids, proteins and DNA markers of CCl4 poisoning? Free Radical
Biology and Medicine 2005; 38: 698–710.
Kazui M, Andreoni KA, Norris EJ, Klein AS, Burdick JF, Beattie C, Sehnert
SS, Bell WR, Bulkley GB and Risby TH. Breath ethane: a specific indica-
tor of free-radical-mediated lipid peroxidation following reperfusion
of the ischemic liver. Free Radical Biology and Medicine 1992; 13:
45
509–515.
wileyonlinelibrary.com/journal/bmc

Marnett LJ. Lipid peroxidation – DNA damage by malondialdehyde. Mu-
tation Research 1999; 424: 83–95.
Mendes R, Cardoso C and Pestana C. Measurement of malondialdehyde
in fish: a comparison study between HPLC methods and the
traditional spectrophotometric test. Food Chemistry 2009; 112:
1038–1045.
Nair U, Bartsch H and Nair J. Lipid peroxidation-induced DNA damage in
cancer-prone inflammatory diseases: a review of published adducts
types and levels in humans. Free Radical Biology and Medicine 2007;
43: 1109–1120.
Nielsen F, Mikkelsen BB, Nielsen JB, Andersen HR and Grandjean P.
Plasma malondialdehyde as biomarker for oxidative stress: reference
interval and effects of life-style factors. Clinical Chemistry 1997; 43:
1209–1214.
Ou JJ, Zhang Y and Montine TJ. In vivo assessment of lipid peroxidation
products associated with age-related neurodegenerative diseases.
Experimental Neurology 2002; 175: 363–369.
Pilz J, Meineke I and Gleiter CH. Measurement of free and bound
malondialdehyde in plasma by high-performance liquid chromatog-
raphy as the 2,4-dinitrophenylhydrazine derivative. Journal of Chro-
matography B 2000; 742: 315–325.
Rosenblat M, Coleman R and Aviram M. Increased macrophage
glutathione content reduces cell-mediated oxidation of LDL and
46
wileyonlinelibrary.com/journal/bmc Copyright © 2014 John Wiley & Sons, Ltd.
A.-M. Domijan et al.
atherosclerosis in apolipoprotein E-deficient mice. Atherosclerosis
2002; 163: 17–28.
Sanz A, Trenzado CE, Botello Castro H, López-Rodríguez MJ and Tierno de
Figueroa JM. Relationship between brain and liver oxidative state and
maximum lifespan potential of different fish species. Comparative
Biochemistry and Physiology Part A 2013; 165: 358–364.
Seljeskog E, Hervig T and Mansoor MA. A novel HPLC method for mea-
surement of thiobarbituric acid reactive substances (TBARS). A com-
parison with a commercially available kit. Clinical Biochemistry 2006;
39: 947–957.
Sim AS, Salonikas C, Naidoo D and Wilcken DEL. Improved method for
plasma malondialdehyde measurement by high-performance liquid
chromatography using methyl malondialdehyde as an internal stan-
dard. Journal of Chromatography B 2003; 785: 337–344.
Steghens J-P, Van Kappel AL, Denis I and Collombel C. Diaminonaphthalene,
a new highly specific reagent for HPLC-UV measurement of total and
free malondialdehyde in human plasma or serum. Free Radical Biology
and Medicine 2001; 31: 242–249.
Syslova K, Kačer P, Kuzma M, Najmanova V, Fenclova Z, Vlačkova Š,
Lebedova J and Pelclova D. Rapid and easy method for monitoring
oxidative stress markers in body fluids of patients with asbestos or
silica-induced lung disease. Journal of Chromatography B 2009; 877:
2477–2486.
Biomed. Chromatogr. 2015; 29: 41–46