Glycobiology vol. 24 no. 3 pp. 281–291, 2014
doi:10.1093/glycob/cwt109
Advance Access publication on December 16, 2013
Glycoxidative damage to human DNA: Neo-antigenic
epitopes on DNA molecule could be a possible reason
for autoimmune response in type 1 diabetes
Saheem Ahmad2,4, Moinuddin1,2, Uzma Shahab2,
Safia Habib2, M Salman Khan3, Khursheed Alam2,
and Asif Ali2
2
Department of Biochemistry, J.N. Medical College, Faculty of Medicine,
Aligarh Muslim University, Aligarh, India and 3Department of Biotechnology,
Integral University, Lucknow, India
Received on December 24, 2012; revised on September 3, 2013; accepted on
September 3, 2013
Advanced glycation end-products (AGEs) are known to be
mutagenic, diabetogenic and vascular disease risk factors.
Methylglyoxal (MG) is a dicarbonyl species that reacts with
biological macromolecule (proteins, DNA and lipids) to give
AGEs. Nonenzymatic glycation of MG with lysine (Lys) in
the presence of copper (Cu2+) is reported to generate reactive
oxygen species (ROS) capable of causing DNA damage. We
show that DNA modification in MG–Lys–Cu2+ system results
in the generation of strand breaks, base modification, hyperchromicity and increased fluorescence intensity. Superoxide
generation in the MG–Lys system was found to be significantly higher when compared with that in the MG and Lys
alone. Moreover, D-penicillamine and pyridoxal phosphate
significantly inhibited the formation of glycation products.
The presence of a major DNA glycation adduct, N2carboxyethyl-2′-deoxyguanosine (CEdG), was detected by
high performance liquid chromatography (HPLC) and confirmed by nuclear magnetic resonance (NMR). As reported
earlier, modified DNA (MG–Lys-Cu2+-DNA) was highly immunogenic in experimental animals. Furthermore, induced
anti-MG–Lys–Cu2+–DNA antibodies were effective probe for
detecting glycoxidative lesions in human genomic DNA of
type I diabetes patients. Our results clearly imply that interaction of MG–Lys and Cu2+ leads to the formation of AGEs
and also the production of potent ROS, capable of causing
DNA damage, thereby playing an important role in diabetes
mellitus.
Keywords: advanced glycation end-products / diabetes / DNA /
glycation / methylglyoxal
1
To whom correspondence should be addressed: Fax: +91-571-2702758;
e-mail: [email protected]
4
Present address: Department of Surgery, Cardinal Bernardin Cancer Center,
Loyola University Medical Center, Maywood, IL 60153, USA.
Introduction
Reactive carbonyl species (RCS) are potent mediators of cellular
carbonyl stress originating from endogenous chemical processes
such as glycation and lipid peroxidation. These are heterogeneous group of carbonyls activated by α, β-unsaturation as in
4-hydroxynonenal and acrolein, α-oxo-substitution as in glyoxal
and methylglyoxal (MG) and β-oxo-substitution as in malondialdehyde (Thornalley 1996). Among all of the reactive carbonyls
produced, MG and glyoxal are exclusively responsible for the
glycation reaction. Glycation is the nonenzymatic addition of
free reducing sugars into biological macromolecules such as
DNA, proteins and lipids (Bucala et al. 1985; Sengupta and
Swenson 2005). Initiation of glycation occurs by the formation
of acid-labile Schiff base adducts, which undergoes Amadori
rearrangements into more stable products known as early
glycation products. The early glycation products undergo slow
transformation to yield the irreversible advanced glycation endproducts (AGEs) (Akhter et al. 2013). Glycation results in the
production of free radicals, which play important roles in the
development of cancer, diabetes, arthritis, heart disease, cataract,
atherosclerosis and neurodegenerative disorders (Schleicher et al.
1997; Baynes 2001; Ahmad, Ahmad, et al. 2011; Mustafa et al.
2012; Raheem et al. 2013). Several interrelations have been
shown between oxidative stress and AGEs. Reactive oxygen
species (ROS) generated in oxidative stress can in turn accelerate
the AGE formation (Ahmad, Akhter, et al. 2013).
Among all dicarbonyl intermediates formed in the glycation
reaction, MG is a potent and most reactive AGE precursor,
forming adducts on arginine, Lys, cysteine and deoxyguanosine
(dG) residues. Furthermore, it has been shown that DNA can be
glycated in vitro yielding N2-carboxyethyl-2′-deoxyguanosine
(CEdG) as a major product. In vitro, nucleobases and ds-DNA
react with sugars in a similar way as proteins (Lee and Cerami
1987; Knerr and Severin 1993; Singh et al. 2001). The exocyclic amino group of 2′-deoxyguanosine is particularly prone
to glycation, leading to the formation of N2-carboxyethyl,
N2-carboxymethyl, as well as cyclic dicarbonyl adducts (Ochs
and Severin 1994). The CEdG is a stable reaction product,
formed from a variety of glycating agents, such as glucose, ascorbic acid, glyceraldehyde, dihydroxyacetone (DHA) and MG
(Larisch et al. 1997; Frischmann et al. 2005). Apart from in
vitro DNA glycation, under hyperglycemic condition the glycation of DNA occurs in vivo and is considered to be a pathogenic
factor for diabetes (Levi and Werman 2003; Dutta et al. 2005).
© The Author 2013. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: [email protected]
281
S Ahmad et al.
The present study aimed to glycate human DNA by MG and
Lys in the presence of Cu2+. Glycation induced structural perturbations in the DNA macromolecule were characterized by
agarose gel electrophoresis and various spectroscopic techniques.
The DNA-AGEs generated in the MG–Lys–Cu2+ system was
confirmed by antiglycation study. Formation of CEdG, a major
adduct of DNA glycation, was confirmed by high performance
liquid chromatography (HPLC) and nuclear magnetic resonance
(NMR) studies. We propose one of the possible reason for type I
diabetes autoimmunity by utilizing the induced antibodies against
glycated human DNA in experimental rabbits (Ahmad et al.
2011). These experimentally induced antibodies have been used
as an immunochemical probe to detect glycoxidative damage in
the lymphocyte DNA isolated from type I diabetic patients.
Materials and methods
Human placental DNA, MG, anti-rabbit IgG alkaline phosphatase conjugates, p-nitrophenyl phosphate, Tween-20, Protein
A-agarose (2.5 mL prepacked column), nuclease S1, superoxide dismutase (SOD), ethidium bromide, agarose and dialysis tubing were purchased from Sigma Chemical Company,
Saint Louis, MO. Polystyrene microtiter flat bottom enzyme
linked immunosorbent assay (ELISA) plates (maxisorp) were
from Nunc (Denmark). Pyridoxal phosphate (PLP) was from
Calbiochem, Billerica, MA. DHA was from Merck, Germany.
Lysine was from Sisco Research Laboratory, Pallikaranai,
Tamil Nadu. HbA1C kit was purchased from Bio-Rad,
Hercules, CA. All other reagents/chemicals were of the highest
analytical grade available. All the ethical requirements were
adhered to for this study.
Purification of human placental DNA
Commercially available human placental DNA was purified
free of proteins and single stranded regions with slight modifications (Habib et al. 2009). DNA (2 mg/mL) was dissolved in
0.1 × saline sodium citrate buffer, pH 7.3 and extracted with an
equal volume of chloroform-isoamyl alcohol (24:1) in a stoppered container and shaken for 1 h. The aqueous layer containing
DNA was separated from organic layer and re-extracted with
chloroform-isoamyl alcohol mixture. DNA was precipitated with
two volumes of chilled absolute ethanol and collected on a glass
rod. After drying in air, the DNA was dissolved in acetate buffer
(30 mM sodium acetate containing 30 mM zinc chloride, pH 5.0)
and treated with nuclease S1 (150 units/mg DNA) at 37°C for
30 min to remove single stranded regions. The reaction was
stopped by adding one-tenth volume of 200 mM ethylene diamine
tertra acetic acid (EDTA), pH 8.0. Nuclease S1 treated DNA was
extracted twice with chloroform-isoamyl alcohol and finally precipitated with two volumes of chilled ethanol. The precipitate was
dissolved in the required buffer. Purity of DNA was confirmed by
A260/A280 ratio which was in the range of 1.8–2.0.
Modification of human placental DNA
Human placental DNA was modified by MG and Lys in the
presence and absence of Cu2+ as described previously (Ahmad
et al. 2011). 37.8 μM of human DNA was thoroughly mixed
with MG (40 mM), Lys (40 mM) and Cu2+ (300 μM) in 10 mM
sodium phosphate buffer, pH 7.4 containing 150 mM NaCl and
282
incubated at 37°C for 24 h followed by extensive dialysis
against phosphate buffer saline (PBS) to remove unbound
constituents.
Spectroscopic analysis
The ultraviolet absorption profile of native human placental
DNA and modified DNA was recorded in the wavelength range
of 200–400 nm on Shimadzu UV-1700 spectrophotometer.
Fluorescence emission profiles of native and modified human
placental DNA were recorded on Shimadzu RF-5301 spectrofluorometer. Native and modified DNA was excited at 370 nm
and emission profile was recorded at 450 nm.
Agarose gel electrophoresis
0.8% agarose in tris acetate EDTA buffer (40 mM Tris-acetate,
pH 8.0 containing 2 mM EDTA) was dissolved by heating. The
solution was cooled to about 50°C and then poured into the gel
tray and allowed to solidify at room temperature. Native and
modified human placental DNA were mixed with 1/10 volume
of sample buffer (0.125% bromophenol blue, 30% Ficoll-400,
5 mM EDTA in 10 × electrophoresis buffer) and loaded in
the wells of 0.8% agarose gels (Shahab et al. 2012).
Electrophoresis was done for 2 h at 40 mA and the gels were
stained with ethidium bromide (0.5 μg/mL), and viewed under
UV light.
Nuclease S1 digestibility
Native human placental DNA and its modified analogue were
characterized by nuclease S1 digestibility assay (Matsuo and
Ross 1987). One μg each of native DNA and the modified DNA
in acetate buffer (30 mM sodium acetate, 1 mM zinc sulphate,
100 mM sodium chloride, pH 4.6) were treated with nuclease S1
(20 unit/μg of DNA) for 30 min at 37°C. The reaction was
stopped by adding one-tenth volumes of 200 mM EDTA, pH
8.0. The samples were subjected to agarose gel electrophoresis
and viewed by illumination under UV light after ethidium
bromide (0.5 μg/mL) staining.
Measurement of superoxide anion
The generation of O2˙ ¯ in the reaction mixture was determined by
cytochrome-c reduction assay (Beauchamp and Fridovich 1971).
A reaction mixture contained MG (40 mM) and Lys (40 mM)
and 100 μM cytochrome-c in 10 mM phosphate buffer (pH 7.4).
The reduction rate was determined as the increase in absorbance
at 550 nm for 10 min at room temperature. Absorbance was taken
at the interval of every 1 min.
AGE inhibition study
AGE inhibition was performed using a carbonyl trapping compound, D-penicillamine (1 mM) and PLP (10 mM) on human
placental DNA. The modifications were studied by incubating
DNA (25 μg/mL) in 10 mM sodium phosphate buffer, pH 7.4
containing MG (40 mM), Lys (40 mM) and Cu2+ (300 μM) at
37°C for 24 h. Percent DNA quenching was calculated from
absorbance at A260 nm. AGE inhibition was further confirmed
from absorbance at A330 nm.
Implication of DNA glycation in diabetes mellitus
Synthesis of the CEdG
The synthesis of CEdG was carried out as described by Seidel
and Pischetsrieder (1998), with slight modifications. Briefly,
50 mg of deoxyguanosine suspended in 1 mL of 100 mM phosphate buffer ( pH, 7.4), was incubated with 100 mg of DHA at
70°C in a shaking water bath. It got dissolved at 70°C in the
course of reaction after 24 h. CEdG was isolated by preparative
HPLC using 50 mM ammonium acetate buffer solution and
methanol as eluents.
1
H NMR of the synthesized CEdG
H NMR spectra were recorded on a Bruker DRX-300 MHz FT
NMR spectrometer. All samples of NMR spectroscopy were
first lyophilized and prepared using dimethyl sulphoxide
(DMSO) as a soluent. The chemical shifts in parts per million
( ppm) are expressed with respect to tetra methyl silane as a
reference.
1
HPLC of native and modified human DNA
An Agilent 1100 capillary HPLC System equipped with a Synergi
C18 analytical column was used for HPLC analysis of native and
modified analog of human placental DNA. General chromatographic conditions were as follows: C18 column (2 mm × 150 mm
with 4 μm particle size); eluent A, 5 mM aqueous ammonium
acetate buffer, pH 7; eluent B, acetonitrile (CH3N) gradient solution—the CH3N concentration was raised from 0 to 4.0% in the
first 5 min; from 4.0 to 6.5% over 30 min; held at 6.5% for 5 min,
and then raised to 90% to wash residual material off the column at
a constant flow rate of 500 μL/min.
Isolation of DNA from human lymphocytes
Blood samples were collected in EDTA vials from different
type I diabetes patients and healthy individuals after prior
consent. Lymphocyte DNA was isolated as per the manufacturer’s instructions. To 500 µL of Qiagen protease, 5 mL of
blood was added and then buffer AL (6 mL) was mixed, followed by vigorous shaking. Mixture was incubated at 70°C for
10 min. Then 5 mL of ethanol (98%) was added to the sample
and mixed by inverting the tube, followed by vigorous shaking.
The solution was then carefully transferred onto QIAamp Maxi
column placed in a 50 mL centrifuge tube and was centrifuged
at 1850 × g for 3 min. Filtrate was discarded and 5 mL of
Buffer AW2 was added and again centrifuged at 4500 × g for 1
min. Again 5 mL of AW2 buffer was added and centrifuged at
4500 × g for 15 min discarding the filtrate. Then 600 µL of
buffer AE was poured directly onto the membrane of the
column which was later incubated for 5 min, and centrifuged at
4500 × g for 2 min. Eluent containing DNA was reloaded onto
the membrane of the column, and was again incubated for 5
min, and centrifuged at 4500 × g for 5 min. Eluent containing
DNA was air dried and dissolved in PBS, pH 7.4. Absorbance
of DNA solution was monitored at 260 and 280 nm to ascertain
its purity and concentration.
Direct binding ELISA for the detection of autoantibodies
ELISA was carried out on flat bottom polystyrene plates as
described earlier (Moinuddin et al. 2012; Shahab et al. 2013).
Briefly, microtiter wells were coated with 100 µl of 2.5 μg/mL
of DNA (in tris buffer saline (TBS), pH 7.4) and incubated for
2 h at 37°C and overnight at 4°C. Each sample was coated in
duplicate and half of the plate, devoid of antigen, served as
control. The test-plate wells were emptied and washed thrice
with tris buffer saline- tween 20 (TBS-T) to remove the
unbound antigen. Unoccupied sites were blocked with 150 μL
of 1.5% nonfat dry milk (in TBS, pH 7.4) for 4–5 h at 4°C followed by single wash with TBS-T. In direct binding ELISA,
sera were directly added into antigen-coated wells and incubated
for 2 h at 37°C and overnight at 4°C respectively. The wells
were emptied and extensively washed with TBS-T. Antiimmunoglobulin G alkaline phosphatase conjugate was added to
each well and incubated at 37°C for 2 h and then the plates were
washed thrice with TBS-T followed by a single wash with distilled water. Para-nitrophenyl phosphate was added and the
developed color was read at 410 nm on a microplate reader. The
results were expressed as mean of difference of absorbance
values in test and control wells (Atest − Acontrol ).
Detection of glycoxidative lesions in human DNA
by enzyme-linked immunosorbent assay
ELISA was performed as described elsewhere with slight modifications (Khan et al. 2007). In order to detect glycation induced
lesions in diabetic subjects, DNA was isolated from human
lymphocytes of diabetes patients and experimentally induced
anti-MG–Lys–Cu2+–DNA IgG was used as a probe to detect
glycoxidative DNA lesions. Immune complexes, formed by incubating a fixed amount (40 μg) of anti-MG–Lys–Cu2+–DNA
IgG with increasing concentration (0–20 μg/mL) of lymphocyte
DNA, were coated on microtiter plates, already coated with
MG–Lys–Cu2+ modified human DNA (2.5 μg/mL).
Estimation of HbA1C
Kit purchased from of Bio-Rad was used for the estimation of
HbA1c in diabetes sera using HPLC based D-10 HbA1c
program. In this procedure, sample is automatically diluted and
injected into the analytical cartridge. The D-10 device delivers
buffer gradient of increasing ionic strength to the cartridge, and
the hemoglobin are separated on the basis of their interactions
with the cartridge material. The separated hemoglobin pass
through the flow cell of the photometer and the absorbance is
recorded at 415 nm. The software performs reduction of raw
data collected from each analysis. Two-level calibration is used
for quantitation of HbA1c values. A sample report and chromatogram is generated for each sample. The HbA1c area is calculated using an exponentially modified Gaussian (EMG)
algorithm that excludes the labile A1c and carbamylated peak
area from the A1c peak area. The normal HbA1c reference
range of the kit used in this study was 4.2–6.1%.
Statistical analysis
Results are expressed as mean ± SEM. All statistical calculations were performed using the SPSS software (version 16).
A P-values of <0.05 was considered as statistically significant.
Linear correlation analysis using absorbance difference as dependent variables was also performed.
283
S Ahmad et al.
Results and discussions
Glycation of human placental DNA
Human placental DNA was glycated by MG and Lys in the
presence and absence of Cu2+. Pilot experiments were run to
work out the incubation time and optimum concentration of
MG, Lys and Cu2+ required to modify the DNA. Human DNA
(37.8 μM) was incubated with MG, Lys and Cu2+ along with respective controls for different time intervals (3, 6, 12 and 24 h)
at 37°C. Maximum hyperchromicity at 260 nm was observed at
24 h incubation as further incubation did not result in any
change in the hyperchromicity. Therefore, our studies were performed on human DNA incubated for 24 h with 40 mM each of
MG and Lys in the presence and absence of copper sulphate
(Cu2+, 300 μM) in PBS.
A characteristic peak at 260 nm was observed with native
human DNA in UV-Vis spectroscopy. However, upon modification with MG and Lys, 68% hyperchromicity was observed
in the DNA peak. Moreover, as reported previously (Ahmad
et al. 2011), in MG + Lys + Cu2+ modified DNA, the observed
hyperchromicity in the DNA peak was 75.5% (Figure 1). This
shows extensive damage to the human DNA, possibly leading
to the formation of single strand breaks and base modifications. However, when MG, Lys and Cu2+ were individually
incubated with human DNA under identical conditions, no
significant change in UV-Vis spectrum was observed.
Furthermore, MG–Cu2+ and Lys-Cu2+ systems also did not
show any detectable change in UV-spectral profile. The
change in hyperchromicity of MG–Lys and MG–Lys–Cu2+
system could be attributed to the destabilization of hydrogen
bonds and modification of nitrogenous bases which result in
the destruction of chromophoric groups through attack on the
sugar-phosphate back bone.
Interestingly, a new additional peak was obtained at 330
nm, when human DNA was incubated with MG–Lys and MG–
Lys–Cu2+ systems. This could be attributed to the generation of
DNA-AGEs. MG–Lys–Cu2+ derived DNA-AGEs show the
highest absorbance, suggesting Cu2+ enhanced AGEs formation. UV absorbance between 320 and 370 nm can be used to
judge the onset of AGE formation that are derived from the reaction of DNA with the modifier (MG–Lys–Cu2+). An earlier
study (Schmitt et al. 2005) has also reported an extra peak in
the 320–330 range, due to the formation of glycation induced
AGEs, thus supporting our findings.
DNA-AGEs are formed spontaneously by the reaction of carbonyl compounds such as MG both in vitro and in vivo, after a
complex cascade of dehydration, condensation, fragmentation,
oxidation and cyclization reactions (Schleicher et al. 2001).
The detection can be based on the fluorescent properties of
AGEs (Monnier and Cerami 1981). After excitation at 370 nm,
fluorescence emission wavelength is typically at 440 nm due to
the presence of heterocyclic compounds (Ahmad, Shahab, et al.
2013). Generation of fluorogenic AGEs in glycated-DNA
samples was measured using excitation wavelength of 370 nm
(λex) and emission wavelength of 450 nm (λem), the characteristic excitation and emission wavelength of AGEs fluorophore.
Under identical conditions, native human DNA alone does not
give any fluorescence. Glycation of DNA by MG and Lys in
the presence copper sulphate generated fluorescent DNA-AGEs
characterized by an increase of 76.2% in the fluorescence
284
Fig. 1. Ultraviolet absorption spectra of native human DNA (―); human DNA
modified with 300 μM Cu2+ (----); 40 mM Lys (open triangle); 40 mM
Lys + 300 μM Cu2+ (filled triangle); 40 mM MG (open circle); 40 mM
MG + 300 μM Cu2+ (filled circle); 40 mM MG + 40 mM Lys (open square) and
40 mM MG + 40 mM Lys + 300 μM Cu2+ (filled square).
Fig. 2. Fluorescence emission spectra of native human DNA (─) and modified
human DNAwith 40 mM MG + 40 mM Lys + 300 μM Cu2+ (----).
intensity in the case of glycated DNA when compared with to
the native form (Figure 2).
Agarose gel electrophoresis of native and MG–Lys–Cu2+
glycated human DNA
Appreciable evidence for generation of strand breaks (single
and double) in the human DNA as a consequence of MG–Lys–
Cu2+ modification was gathered by gel electrophoresis on 0.8%
agarose. The gel pattern (Figure 3A) shows an increase in the
mobility of glycated DNA with increasing incubation time
(Lanes 2–5). Maximum mobility was observed at 24 h incubation, and further incubation did not have any consequential
effect on DNA migration pattern. The increase in mobility may
be due to the generation of single stranded breaks by glycation
induced intermediates or free radical generation or both, which
Implication of DNA glycation in diabetes mellitus
Fig. 3. (A) Agarose gel electrophoresis of native and modified human DNA. Lanes 2–5: DNA samples treated with MG–Lys (40 mM each) and Cu2+ (300 μM).
Lane 1: Native human DNA. Lane 2: Modified human DNAwith 3 h incubation. Lane 3: Modified human DNAwith 6 h incubation. Lane 4: Modified human DNA
with 12 h incubation. Lane 5: Modified human DNAwith 24 h incubation. (B) Nuclease S1 digestibility of native and modified DNA. Lane 1 contains native DNA,
while Lane 2 contains native DNA treated with nuclease S1 for 30 min; Lane 3 contains modified DNA and Lane 4 contains modified DNA treated with nuclease S1
for 30 min. Electrophoresis was carried out on 0.8% agarose gel for 2 h at 30 mA.
may result in the formation of small size DNA showing faster
mobility compared with the native DNA of Lane 1.
Earlier studies have demonstrated that the structural alteration
in DNA following damage by various agents result in the generation of single strands in the DNA molecule (Shishido and
Ando 1974; Yamasaki et al. 1997). Therefore, generation of
single strand breaks was analyzed by nuclease S1 digestibility
assay. Native and modified human DNA were digested with nuclease S1 (20 units/μg DNA) for 30 min and electrophoresed
on 0.8% agarose to visualize the generation of single strand
breaks. The controls were native and modified DNA samples,
untreated with nuclease S1. Modified DNA showed decrease in
fluorescence intensity following nuclease S1 digestion. On the
other hand, nuclease S1 treated and untreated native DNA
showed almost identical electrophoretic migration pattern and
fluorescence intensity (Figure 3B). The results show substantial
digestion of modified DNA by nuclease S1, while native DNA
remained unaffected. These observations clearly demonstrate
sufficient distortions (formation of single strand breaks) in
the helical structure of DNA upon MG–Lys–Cu2+ treatment,
rendering it susceptible to digestion by single strand specific
nuclease S1.
Quantitation of superoxide anion
Amino acids (Lys and arginine) treated with MG have been
reported to give electron paramagnetic resonance signals
(Pethig and Szent-Gyorgyi 1977), indicative of the formation
of free radicals (˙OH and O2˙–). Several reports (Bucala et al.
1991; Simpson et al. 1992) have shown that glycated protein
generated superoxide anions, which initiated lipid peroxidation.
During incubation of MG with Lys, the formation of superoxide
anion was gradually increased in a time-dependent manner. It
has also been reported that the incubation of threose with
−1 −1
N-acetyl-Lys produced 16.2 nmol O•–
h (Ortwerth et al.
2 mL
1998). In this study, the incubation of MG with Lys produced
−1 −1
28.2 nmol O•–
h . However, MG and Lys alone gener2 mL
−1 −1
ated only 2.46 and 1.96 nmol O•–
h , respectively. The
2 mL
generation of superoxide anion in terms of reduced
−1 −1
cytochrome-c (28.2 nmol O•–
h ) was found to be statis2 mL
tically significant (P-value < 0.001) against the controls MG
and Lys. During incubation of MG with Lys, the formation of
superoxide anion was gradually increased in time dependent
manner (Figure 4A). The result suggests that DNA strand
breakage by the MG–Lys–Cu2+ system may be dependent on
the formation of superoxide anion in this system. Moreover, the
production of superoxide radical gradually inhibited by increasing concentration of SOD, further confirming the formation of
superoxide radical (Figure 4B). The decrease in the superoxide
formation was statistically highly significant for 500 units of
SOD (P < 0.001).
Antiglycation study
Antiglycating activity of D-penicillamine and PLP was assessed
in vitro via UV-Vis spectroscopic measurements. The extra
peak at 330 nm in UV-Vis spectroscopy, as reported earlier, is
due to the formation of DNA-advance glycation end-products
(Ahmad et al. 2011). Quenching of this peak was exploited to
see their antiglycating effect from absorbance at A330 nm.
D-penicillamine (1 mM) and PLP (10 mM) showed remarkable
inhibition of 55 and 73% respectively in DNA modification as
analyzed at 260 nm; the characteristic wavelength maxima for
DNA. However, at 330 nm D-penicillamine and PLP showed
still higher inhibition of 69 and 85%, respectively (Figure 5).
The decrease in the percent modification from PLP and pyridoxamine at 260 and 330 nm were found statistically significant
(P < 0.001) when compared with original modification of the
glycation reaction. The enhanced inhibition of 330 nm peak
represents inhibition of AGEs formation, clearly indicating the
formation of AGEs in the MG–Lys–Cu2+ modified human
DNA.
285
S Ahmad et al.
Fig. 5. Effect of antiglycating agent D-penicillamine (1 mM) and PLP (10 mM)
on the modification of DNA induced by MG–Lys–Cu2+ system. Percent DNA
modification was calculated at 260 and 330 nm after 24 h incubation. The
decrease in the % modification from PLP and pyridoxamine at 260 and 330 nm
were statistically significant (P < 0.001) when compared with original
modification.
Fig. 4. (A) Generation of the superoxide anion in the glycation reaction of Lys
with MG. Reduction of cytochrome-c was measured by increasing
concentrations of reaction products. The reaction mixture contained 10 μM
cytochrome-c in 10 mM phosphate buffer pH 7.4 and the following: 40 mM
MG (open square); 40 mM Lys (filled square); 40 mM MG and Lys (filled
triangle). Absorbance changes were monitored at 550 nm for 10 min at room
temperature. The increase in reduced Cyt-C were statistically highly significant
than the controls (P < 0.001). (B) Superoxide formation during the glycation
reaction of Lys by MG. MG (40 mM) and Lys (40 mM) was reacted in presence
of various concentrations of SOD, 50 units (1), 100 units (2) and 500 units (3).
The decrease in the superoxide formation were statistically highly significant
for 500 units (P < 0.001); significant for the 100 and 50 units (P < 0.05).
Characterization of CEdG by HPLC
Synthesis of the standard, CEdG was performed as described
earlier (Seidel and Pischetsrieder 1998). After final preparation,
CEdG was isolated by preparative HPLC using 50 mM ammonium acetate buffer solution and methanol as eluents. The
elution of CEdG was obtained at a retention time of 14.399 min
when UV detector was used for the experiment (Figure 6A).
However, deoxyguanosine gave elution at a retention time of
9.1 min.
NMR of CEdG
For structural assignment of the compound, the CEdG peak
was isolated by HPLC and then subjected to 1H NMR analysis.
286
Resonance signals can be identified and characterized in 1H
NMR spectrum of CEdG (Figure 6B), recorded in DMSO-d6.
1
H NMR (400 MHz, DMSO-d6, 20°C) assignment for CEdG:
δ 10.6 (s,1H, N1-H), δ 7.93 (s, 1H, C8-H), δ 6.76 (d, 1H,
C2-NH), δ 6.12 (dd, 1H, C1′-H), δ 5.3 (d, 1H, C3′-OH), δ 4.89
(s, 1H, C5′-OH), δ 4.36 (m, 1H, C2-NH-CH), δ 4.18 (q, 1H,
C2-NH-CH), δ 3.81 (m, 1H,C3′-H), δ 3.5 (ddd, 2H, C5′-H2),
δ 2.64 (ddd, 1H, C2′-H), δ 1.39 (d, 3H, C2-NH-CH-CH3).
The 1H NMR analysis indicated a -CHX-CH3 group bound
to the purine (guanine) ring. The chemical shifts (δ), 4.18 ppm
(q, 1H, C2–NH–CH) and 1.39 ppm (d, 3H, C2–NH–CH–CH3),
respectively, is attributed to the carboxyethyl group of the
standard CEdG.
Characterization of native and MG–Lys–Cu2+
glycated human DNA by HPLC
Figure 7 shows the representative HPLC chromatograms of
acid hydrolyzed samples of native and modified human DNA.
Well-defined peaks at retention time 4.467, 7.332 and 8.727 min
were observed in native DNA. However, in the case of modified
DNA, these peaks shifted to 6.748, 8.455 and 12.399 min,
respectively, suggesting considerable change in the DNA
bases. The extra peak at a retention time of 14.249 min in
the acid hydrolysate of modified DNA is characteristic of
N2-(1-carboxyethyl)-2-deoxyguanosine (CEdG) adduct. This
is in conformity with the standard CEdG results wherein also
when deoxyguanosine was exposed to DHA, a distinct peak at
retention time of 14.399 min was observed (Figure 6A). The
CEdG adduct is a marker of glucose and MG-induced DNA
glycation.
Implication of DNA glycation in diabetes mellitus
Fig. 6. (A) Representative HPLC chromatogram of the reaction of 2′-deoxyguanosine with DHA. (B). 1H NMR spectra of the standard CEdG.
Glycation adducts of DNA may have a potential as biomarkers, since all nucleated cells contain the same DNA content
and should reflect the relative level of MG in the target tissue.
Reaction of double-stranded DNA with MG or glucose in vitro
primarily produces N 2-carboxyethyl-2′-deoxyguanosine (CEdG),
suggesting it to be the likely major adduct formed in vivo
(Papoulis et al. 1995). This implicates that CEdG is a useful
biomarker for monitoring oxoaldehyde-induced stress in response to enhanced glycolytic flux or environmental exposure
to MG. The standard adduct, CEdG, formed by preparative
HPLC was further proved by nuclear magnetic resonance analysis. The HPLC performed for native and glycated DNA
showed remarkable difference in peak retention times. The
extra peak in modified DNA is in accordance with the standard
287
S Ahmad et al.
Fig. 7. (A) Representative HPLC chromatogram of acid hydrolysate of native human DNA. (B) Representative HPLC chromatogram of acid hydrolysate of modified
human DNA.
CEdG results. However, native DNA did not show any peak at
this retention time, suggesting sufficient distortion in the DNA
structure has caused it to form the glycation adduct, CEdG,
which might be used as a biological marker for early detection
of diabetes.
288
Detection of autoantibodies against MG–Lys–Cu2+–DNA
in diabetic subjects
In order to analyze the role of glycated DNA in eliciting
immune response in diabetic patients, the presence of autoantibodies in the diabetic subjects against MG–Lys–Cu2+–DNA
Implication of DNA glycation in diabetes mellitus
was probed. The pilot study was performed to screen out the
positive sera samples from both type I and type II diabetic subjects. Our study comprised 40 serum samples each of type I
and type II diabetes mellitus patients (totally 80 sera samples).
Control serum samples from age- and sex-matched individuals
were obtained from 20 normal healthy subjects. All sera were
diluted to 1:100 in TBS-T and subjected to direct binding
ELISA on solid phase, separately coated with equal amounts of
native human DNA and MG–Lys–Cu2+ glycated human DNA.
Out of 40 sera from type I diabetes, 27 (67.5%) showed higher
binding with the MG–Lys–Cu2+–DNA, whereas in type II diabetes, only 10 (25%) out of 40 showed enhanced binding with
the MG–Lys–Cu2+-DNA when compared with the native form
(Figure 8). The serum samples that showed enhanced binding
(more than double the background) were considered for positive sera having antibodies against glycated DNA. In case of
type I diabetes mellitus patients we observed high statistical
significance (P < 0.001) in the binding of glycated human
DNA when compared with the native analog; however, in type
II diabetes patients, the binding was moderately significant in
statistical terms (P ≤ 0.05) when compared with native human
DNA. Moreover, the 10 sera that showed enhance binding in
type II diabetic sera were having secondary complication of
nephropathy. These results points toward that most type I
Fig. 8. Direct binding ELISA of serum antibodies from diabetes mellitus (DM)
patients to native human DNA (open square) and glycated human DNA (filled
square). Serum from normal human subjects (NHS) served as control. The
microtiter plates were coated with the respective antigens (2.5 μg/mL).
*P < 0.001 vs. native human DNA for type 1 diabetes; *P ≤ 0.05 vs. native
human DNA for type 2 diabetes.
diabetic subjects (67.5%) and some type II (25%) (only secondary complication samples) showed the presence of autoantibodies against MG–Lys–Cu2+–DNA.
Glycoxidative lesion detection
As reported earlier, female rabbits immunized with MG–Lys–
Cu2+ modified human DNA elicited high titer immunogenspecific antibodies (Ahmad et al. 2011). The experimentally
induced anti-MG–Lys–Cu2+–DNA IgG was used as an immunochemical probe to detect the glycoxidative lesions in the
lymphocyte DNA isolated from the blood of type I and II diabetes patients. After the initial screening of the type I and II diabetic sera for the presence of autoantibodies, the detection of
glycoxidative lesion was executed only for the sera displaying
very high absorbance in direct binding ELISA. Therefore, 12
type I and 8 type II sera were screened out from 40 sera
samples of each group. The binding pattern of the DNA isolates
was quite revealing; DNA of type I diabetes patients produced
appreciable inhibition in the binding of experimentally induced
anti-MG–Lys–Cu2+–DNA antibodies. Inhibition of anti-MG–
Lys–Cu2+–DNA-IgG by lymphocyte DNA from type I diabetes
patients was recorded in the range of 58.2–72.4%. However,
type II diabetes subjects exhibited moderate inhibition in the
range of 41.3–54.4%. The inhibition of anti-MG–Lys–Cu2+–
DNA-IgG by lymphocyte DNA from healthy individuals was
quite low (Table I). Moreover, the average inhibition for type I
diabetes was 67.03%, while for type II subjects it was recorded
to be meager 46.91%.
Significantly high recognition of the lymphocyte DNA from
type I diabetes patients by the experimentally induced antibodies against the modified DNA is a clear indicator of epitope
sharing between the human DNA modified in vitro by MG–
Lys–Cu2+ and the lymphocyte DNA of type I diabetes patients.
This leads to the conclusion that MG–Lys–Cu2+ generates
neoepitopes on the lymphocyte DNA molecule that are recognized as “alien” or nonself by the immune system resulting in
the autoantibody generation in diabetes type I patients.
Significantly high recognition of the lymphocyte DNA from
type I diabetes patient by the experimentally induced antibodies
against MG–Lys–Cu2+ modified human DNA forms an evidence toward the involvement of modified bases and single
strand regions in disease pathogenesis.
Correlation between HbA1c and anti-glycated-DNA
autoantibodies
Table II shows the glycemic parameters in diabetic subjects
based on relative absorption to native DNA and DNA-AGEs.
Table I. Detection of glycoxidative lesions in the DNA samples isolated from type I and II diabetes mellitus patients and normal healthy individuals
Sera group
Number of samples tested
Maximum percent inhibition at 20 μg/mL
Mean ± SD
Type I diabetes
Type II diabetes
Normal human
12
8
5
70.2, 60.5, 58.2, 61.7, 69.6, 65.4, 68.7, 68.9, 71.4, 67.3, 70.1, 72.4
41.3, 44.2, 46, 54.4, 43.6, 46.5, 48.3, 51
30.4, 25.6, 20.85, 29.6, 24.3
67.03 ± 5.35%
46.91 ± 4.67%
26.15 ± 3.93%
Binding of anti-MG–Lys–Cu2+ modified DNA IgG to DNA isolated from lymphocytes of type I and II diabetes patients and normal healthy individuals was
ascertained by competitive inhibition ELISA. The microtiter plates were coated with MG–Lys–Cu2+modified human DNA (2.5 μg/mL); DNA isolated from
lymphocytes of patients with type I and II diabetes and healthy individuals were used as inhibitors. DNA isolates from diabetic subjects showed significantly higher
inhibition than normal human DNA (P < 0.001).
289
S Ahmad et al.
Table II. Application of t test on various glycemic parameters and disease
duration
Variable
Fasting blood sugar (mg/dl)
Postprandial blood sugar
(mg/dl)
HbA1c (%)
Disease duration (in years)
Value with
native DNA
Value with
DNA-AGEs
t*
P**
Mean
SD
Mean
SD
118.4
166.6
36.3
48.5
154.5
202.3
56.2
71.3
−2.5
−2.89
<0.015
<0.005
7.6
7.0
1.8
5.0
10.2
6.5
2.8
4.5
−3.3
0.45
<0.001
>0.533
t* is the value of Student’s t test.
P** statistical significance, indicate difference (independent sample t test).
Table III. Linear correlation using absorbance difference as dependent variable
in diabetic subjects
Variable
r
r2
t
P
% HbA1c
Nephropathy
Triglycerides, mg%
HDL, mg%
LDL, mg%
VLDL, mg%
0.486
0.239
0.186
0.065
0.013
0.190
0.236
0.057
0.034
0.004
0.001
0.036
6.55
2.90
2.230
0.768
0.153
2.282
<0.001
<0.003
<0.001
>0.05
>0.05
<0.001
There was significant difference in fasting and postprandial
blood sugar and HbA1c (P < 0.015, P < 0.005, P < 0.001,
respectively) but no significant difference in disease duration.
Moreover, linear correlation analysis was performed using
absorbance difference as dependent variable and HbA1c%,
nephropathy and lipids as independent variables (Table III). It
was found that the absorbance difference was significantly
related to HbA1c (r = 0.486, P < 0.001), nephropathy (r = 0.239,
P < 0.003) and serum triglycerides and very low density lipoprotein-cholestrol (VLDL-cholesterol) (r = 0.186, P < 0.001 and
r = 0.190, P < 0.001, respectively).
with HbA1c of 7.6 ± 1.2% (P < 0.001) indicates that the effects
of DNA glycation are marked in diabetes. The results are in
line with an earlier report (Bucala et al. 1984), who have shown
that like protein, nucleic acids too can suffer nonenzymatic
modification by reducing sugars and persuade structural and organizational changes analogous to those reported for nonenzymatic browning of proteins. The DNA glycation marker,
N2-carboxyethyl-2′-deoxyguanosine, has been reported in
kidney and aorta of diabetic and uremic patients (Li et al.
2006). In view of the immunogenic nature of AGEs, it may not
be out of context to speculate that persistence of comparable
structures in vivo can initiate and/or propagate autoimmune response (Turk et al. 2001). In our study, the presence of
anti-DNA-AGEs autoantibodies in 67.5% of type I diabetic
sera may be a consequence of auto-immune response against
persistent DNA-AGEs. Anti-DNA-AGEs autoantibodies were
also correlated with diabetic nephropathy. Findings from the
present study indicate that DNA-AGEs and autoantibodies
against glycated DNA are correlated with HbA1c and microvascular complication like diabetic nephropathy and may help
as supplementary biomarkers for valuation of chronic glycemia. Nevertheless, a comprehensive, potential and prospective study is required to establish efficacy of DNA-AGEs and its
autoantibodies in the assessment of glycemic control and its relationship with microvascular hitches in diabetic subjects.
Acknowledgements
Equipment facility of the DST-FIST program of the
Department is thankfully acknowledged. Thanks are also to
sophisticated analytical instrumentation facility (SAIF) provided by CDRI, CSIR, Lucknow, India.
Conflict of interest
None declared.
Abbreviations
Conclusion
The products of nonenzymatic glycation and oxidation of
DNA, the DNA-AGEs, accumulate in a wide variety of environments. MG, an offshoot product of carbohydrate and protein
metabolism is involved in the formation of AGEs by reacting
with free amino groups of Lys and arginine, which are present
at the active sites in most of the enzymes. The AGEs was more
pronounced in the presence of Cu2+. Thus, the MG–Lys–Cu2+
system can be very deleterious for macromolecules like DNA,
proteins etc. Our results suggest that the cumulative action of
these species produce AGEs and potent oxidants as well, which
cause damage to the DNA molecule. Moreover, similar pattern
of perturbations was found in DNA isolates from type I diabetes patients, indicating the generation of neo-antigenic epitopes on DNA molecule that could be a possible reason for
autoimmune response in type I diabetes.
In the present study, higher binding of serum antibodies with
MG–Lys–Cu2+glycated-DNA as against native DNA in diabetic patients with HbA1c of 10.2 ± 2.8% compared with those
290
AGEs, Advanced glycation end-product; CEdG, N2-carboxyethyl2′-deoxyguanosine; Cu2+, copper; DHA, dihydroxy acetone;
DHA, dihydroxyacetone; EMG, exponentially modified Gaussian;
Lys, lysine; MG, methylglyoxal; PBS, phosphate buffer saline;
PLP, pyridoxal phosphate; ROS, reactive oxygen species; SOD,
superoxide dismutase.
References
Ahmad MI, Ahmad S, Moinuddin. 2011. Preferential recognition of
methylglyoxal-modified calf thymus DNA by circulating antibodies in
cancer patients. Ind J Biochem Biophys. 48:290–296.
Ahmad S, Moinuddin, Dixit K, Shahab U, Alam K, Ali A. 2011. Genotoxicity
and immunogenicity of DNA-advanced glycation end products formed by
methylglyoxal and lysine in presence of Cu2+. Biochem Biophys Res
Commun. 407:568–574.
Ahmad S, Akhter F, Shahab U, Moinuddin, Khan MS. 2013. Studies on glycation of human low density lipoprotein: A functional insight into physicochemical analysis. Int J Biol Macro. 62:167–71.
Ahmad S, Shahab U, Baig MH, Khan MS, Khan M. Sajid, Saeed M, Srivastava
AK, Moinuddin. 2013. Inhibitory effect of Metformin and Pyridoxamine in
the formation of early, intermediate and advanced glycation end-products.
PLoS ONE. 8(9):e72128.
Implication of DNA glycation in diabetes mellitus
Akhter F, Khan MS, Shahab U, Moinuddin, Ahmad S. 2013. Bio-physical characterization of ribose induced glycation: A mechanistic study on DNA perturbations. Int J Biol Macro. 58:206–210.
Baynes JW. 2001. The role of AGEs in aging: Causation or correlation. Exp
Gerontol. 36:1527–1537.
Beauchamp C, Fridovich I. 1971. Superoxide dismutase: Improved assays and
an assay applicable to acrylamide gels. Anal Biochem. 44:276–287.
Bucala R, Model P, Anthony C. 1984. Modification of DNA by reducing
sugars: A possible mechanism for nucleic acid aging and age-related dysfunction in gene expression. Biochemistry. 81:105–109.
Bucala R, Model P, Russel M, Cerami A. 1985. Modification of DNA by
glucose 6-phosphate induces DNA rearrangements in an Escherichia coli
plasmid. Proc Natl Acad Sci USA. 82:8439–8442.
Bucala R, Tracey KJ, Cerami A. 1991. Advanced glycosylation products
quench nitric oxide and mediate defective endothelium-dependent vasodilatation in experimental diabetes. J Clin Invest. 87:432–438.
Dutta U, Cohenford MA, Dain JA. 2005. Nonenzymatic glycation of DNA
nucleosides with reducing sugars. Anal Biochem. 345:171–180.
Frischmann M, Bidmon C, Angerer J, Pischetsrieder M. 2005. Identification of
DNA adducts of methylglyoxal. Chem Res Toxicol. 18:1586–1592.
Habib S, Moinuddin, Ali A, Ali R. 2009. Preferential recognition of peroxynitrite modified human DNA by circulating autoantibodies in cancer patients.
Cell Immunol. 254:117–123.
Khan WA, Alam K, Moinuddin . 2007. Catechol-estrogen modified DNA: A
better antigen for cancer autoantibody. Arch Biochem Biophys. 465:293–300.
Knerr T, Severin T. 1993. Reaction of glucose with guanosine. Tetrahedron Lett.
34:7389–7390.
Larisch B, Pischetsrieder M, Severin T. 1997. Formation of guanosine adducts
from l-ascorbic acid under oxidative conditions. Bioorg Med Chem Lett.
7:2681–2686.
Lee A, Cerami A. 1987. The formation of reactive intermediate(s) of glucose
6-phosphate and lysine capable of rapidly reacting with DNA. Mutat Res.
179:151–158.
Levi B, Werman MJ. 2003. Fructose and related phosphate derivatives impose
DNA damage and apoptosis in L5178Y mouse lymphoma cells. J Nutr
Biochem. 14:49–60.
Li H, Nakmura S, Miyazaki S, Morita T, Suzuki M, Pischetsrieder M, Niwa T.
2006. Carboxyethyl-2′-deoxyguanosine, a DNA glycation marked, in
kidneys and aortas of diabetic and uremic patients. Kidney Int. 69:388–392.
Matsuo N, Ross PM. 1987. Measurement of interstrand cross-link frequency and
distance between interruptions in DNA exposed to 4,5′,8-trimethylpsoralen
and near-ultraviolet light. Biochemistry. 26:2001–2009.
Moinuddin, Dixit K, Ahmad S, Shahab U, Habib S, Naim M, Alam K, Ali A.
2012. Human DNA damage by the synergistic action of 4-aminobiphenyl
and nitric oxide: An immunochemical study. Environ Toxicol. doi:10.1002/
tox.21782.
Monnier VM, Cerami A. 1981. Nonenzymatic browning in vivo: Possible
process for aging of long-lived proteins. Science. 211:491–493.
Mustafa I, Ahmad S, Dixit K, Moinuddin, Ahmad J, Ali A. 2012. Glycated
human DNA is a preferred antigen for anti-DNA antibodies in diabetic
patients. Diabetes Res Clin Pract. 95:98–104.
Ochs S, Severin T. 1994. Reaction of 2′-deoxyguanosine with glyceraldehyde.
Liebigs Ann Chem. 5:851–853.
Ortwerth BJ, James H, Simpson G, Linetsky M. 1998. The generation of superoxide anions in glycation reactions with sugars, osones, and 3-deoxyosones.
Biochem Biophys Res Commun. 245:161–165.
Papoulis A, al-Abed Y, Bucala R. 1995. Identification of N2-(1-carboxyethyl)
guanine (CEG) as a guanine advanced glycosylation end product.
Biochemistry. 34:648–655.
Pethig R, Szent-Gyorgyi A. 1977. Electronic properties of the caseinmethylglyoxal complex. Proc Natl Acad Sci USA. 74:226–228.
Raheem M, Iram S, Sajid KM, Khan MS, Shukla AR, Srivastava AK, Ahmad
S. 2013. Glycation-assisted synthesized gold nanoparticles inhibit growth
of bone cancer cells. Colloid and Surfaces-B. http://dx.doi.org/10.1016/j.
colsurfb.2013.12.008.
Schleicher ED, Bierhaus A, Haring HU, Nawroth PP, Lehmann R. 2001.
Chemistry and pathobiology of advanced glycation end products. Contrib
Nephrol. 131:1–9.
Schleicher ED, Wagner E, Nerlich AG. 1997. Increased accumulation of the
glycoxidation product N(epsilon)-(carboxymethyl)lysine in human tissues in
diabetes and aging. J Clin Invest. 99:457–468.
Schmitt A, Schmitt J, Munch G, Gasic-Milencovic J. 2005. Characterization of
advanced glycation end products for biochemical studies: Side chain modifications and fluorescence characteristics. Anal Biochem. 338:201–215.
Seidel W, Pischetsrieder M. 1998. DNA-glycation leads to depurination by the
loss of N2-carboxyethylguanine in vitro. Cell Mol Biol. 44:1165–1170.
Sengupta B, Swenson J. 2005. Properties of normal and glycated human hemoglobin in presence and absence of antioxidant. Biochem Biophys Res
Commun. 334:954–959.
Shahab U, Moinuddin, Ahmad S, Dixit K, Abidi SMA, Alam K, Ali A. 2012.
Acquired immunogenicity of human DNA damaged by N-hydroxy-N-acetyl-4aminobiphenyl. IUBMB Life. 64:340–345.
Shahab U, Moinuddin, Ahmad S, Dixit K, Habib S, Alam K, Ali A. 2013.
Genotoxic effect of N-hydroxy-4-acetylaminobiphenyl on human DNA:
Implications in bladder cancer. PL0S One. 8(1):e53205.
Shishido K, Ando T. 1974. Cleavage of ultraviolet light-irradiated DNA by single
strand-specific S1 endonuclease. Biochem Biophys Res Commun. 59:1380–1388.
Simpson JA, Narita S, Gieseg S, Gebicki S, Gebicki JM, Dean RT. 1992.
Long-lived reactive species on free-radical-damaged proteins. Biochem J.
282:621–624.
Singh R, Barden A, Mori T, Beilin L. 2001. Advanced glycation end-products:
A review. Diabetologia. 44:129–146.
Thornalley PJ. 1996. Pharmacology of methylglyoxal: Formation, modification
of proteins and nucleic acids, and enzymatic detoxification—A role in pathogenesis and antiproliferative chemotherapy. Genet Pharmacol. 27:565–573.
Turk Z, Ljubic S, Turk N, Benko B. 2001. Detection of autoantibodies against
advanced glycation end products and AGE immune complexes in serum of
patients with diabetes mellitus. Clin Chim Acta. 303(1–2):105–115.
Yamasaki H, Pulkabrek P, Grurberger D, Weinstein IB. 1997. Differential excision from DNA of the C-8 and N2 guanosine adducts of N-acetyl-2aminofluorene by single strand-specific endonucleases. Cancer Res.
37:3756–3760.
291