© 2000 American Heart Association, Inc.
Atherosclerosis and Lipoproteins |
Glucosylated Glycerophosphoethanolamines are the Major LDL Glycation Products and Increase LDL Susceptibility to Oxidation
Evidence of Their Presence in Atherosclerotic Lesions
From the Department of Laboratory Medicine and Pathobiology (A.R., N.A.S.), and Banting and Best Department of Medical Research (A.R., A.K.), University of Toronto; and Spectral Diagnostics Inc (N.A.S.), Toronto, Canada.
Correspondence to Dr Arnis Kuksis, Banting and Best Department of Medical Research, University of Toronto, 112 College St, Toronto, Ontario M5G 1L6, Canada. E-mail arnis.kuksis{at}utoronto.ca
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Abstract—Glycation of both protein and lipid components is believed to be involved in LDL oxidation. However, the relative importance of lipid and protein glycation in the oxidation process has not been established, and products of lipid glycation have not been isolated. Using glucosylated phosphatidylethanolamine (Glc PtdEtn) prepared synthetically, we have identified glycated diacyl and alkenylacyl species among the ethanolamine phospholipids in LDL. Accumulation of these glycation products in LDL incubated with glucose showed a time- and glucose concentration–dependent increase. LDL specifically enriched with Glc PtdEtn (25 nmol/mg protein) showed increased susceptibility to lipid oxidation when dialyzed against a 5-µmol/L Cu2+ solution. The presence of this glucosylated lipid resulted in a 5-fold increase in production of phospholipid-bound hydroperoxides and 4-fold increase in phospholipid-bound aldehydes. Inclusion of glucosylated phosphatidylethanolamine in the surface lipid monolayer of the LDL resulted in rapid loss of polyunsaturated cholesteryl esters from the interior of the particle during oxidation. Glycated ethanolamine phospholipids were also isolated and identified from atherosclerotic plaques collected from both diabetic and nondiabetic subjects. The present findings provide direct evidence for the previously proposed causative effect of lipid glycation on LDL oxidation.
Key Words: low-density lipoproteins • peroxidation, lipid • glycation, phospholipid • diabetes • atheroma
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Recent work on the pathogenesis of atherosclerosis implicates "oxidized" low density lipoprotein (LDL) as a key factor in the initiation of atherosclerotic lesions.1 Research has suggested2 that nonenzymatic glycation of apoprotein B and possibly aminophospholipids could promote generation of "oxidized" LDL. However, specific contributions of lipid and protein glycation to LDL oxidation have not been determined.
Nonenzymatic glycation is defined as a posttranslational modification of a protein by the covalent attachment of a sugar residue, which results in an amino-carbonyl bonding or a Schiff base linkage. The stable Amadori rearrangement products of these Schiff bases undergo further rearrangement and modification, which results in advanced glycation end products (AGE).3 Glycation has been shown to alter the biological activity of LDL and to result in reduced rate of degradation by fibroblasts4 and increased cholesteryl ester deposition in human aortic cells compared with normal LDL.5 Using anti-AGE antibodies, Bucala et al6 have shown that the lipid moiety of glycated LDL is involved in LDL susceptibility to oxidation, but no products of lipid glycation were identified. Gugliucci et al7 showed that LDL glycation constitutes a predisposing event to its subsequent oxidative modification and that LDL and VLDL fractions from diabetic patients are more susceptible to oxidation. Furthermore, glycation of LDL has been reported8 to accelerate LDL oxidation by copper ions, and glycation and lipid oxidation have been reported to increase the uptake of LDL particles by macrophages.9 In none of these instances were glycated or oxidized lipids isolated or identified nor was a distinction made between effects of protein and lipid glycation. We have recently shown10 11 that plasma phosphatidylethanolamine (PtdEtn), which composes 5% to 6% of total phospholipid, can be readily glucosylated and the glucosylation products isolated. The present study demonstrates that glucosylated glycerophosphoethanolamines are the major LDL lipid glycation products. Furthermore, glucosylated (Glc) PtdEtn added to nonglycated LDL promotes the oxidation of phosphatidylcholine (PtdCho) and cholesteryl esters.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Materials
Egg yolk PtdEtn, bovine brain PtdEtn, 1-palmitoyl-2-linoleoyl glycerophosphoethanolamine (GroPEtn), NaCNBH3, and NaBH4 were obtained from Sigma-Aldrich. A standard mixture of phospholipids made up of equal-weight proportions of PtdCho, PtdEtn, phosphatidylserine (PtdSer), phosphatidylinositol (PtdIns), lysoPtdCho, and sphingomyelin (SM) was also purchased from Sigma-Aldrich. D-Glucose-1-[14C] was obtained from New England Nuclear. All solvents were of high-performance liquid chromatographic (HPLC) grade or higher quality and were obtained from local suppliers.
Synthesis and Isolation of Glc PtdEtn
Glc PtdEtn was prepared and purified as described in detail
previously.10 Briefly, PtdEtn (2 mg) dissolved in 1 mL of
methanol was transferred to a 15-mL test tube and the solvent
evaporated under nitrogen. Four mL of 0.1 mol/L phosphate buffer
containing 0 to 400 mmol/L glucose and 0.1 mmol/L EDTA were
added and sonicated at low power for 5 minutes at room temperature. The
mixture was incubated under nitrogen at 37°C for various periods of
time. Lipids were extracted into chloroform/methanol (2:1, vol/vol) as
described by Folch et al12 and the solvents evaporated
under nitrogen. Some samples were reduced by addition of a methanolic
solution of either NaCNBH3 or
NaBH4 to a final concentration of 70 mmol/L
and incubation at 4°C for 1 hour. Reduction products were washed
with water, reextracted with chloroform/methanol (2:1, vol/vol), and
dried under nitrogen. Samples were redissolved in chloroform/methanol
(2:1, vol/vol) and kept at -20°C until analysis. Glc PtdEtn
(2 mg) was purified by preparative thin-layer
chromatography (TLC; 20x20-cm glass plates) coated
with silica gel H (250-µm-thick layer). Chromatoplates were
developed with chloroform/methanol/30% ammonia (65:35:7 by volume) as
described.10 Phospholipids were identified by
cochromatography with appropriate standards and
visualization of any lipid bands under ultraviolet light after the
plate was sprayed with 0.05% 2,7-dichlorofluorescein in
methanol. Both glucosylated and nonglucosylated lipids were recovered
by scraping the gel from appropriate areas of the plate and extracting
it twice with the developing solvent.
Liposomal Oxidation
Unilamellar liposomes of egg yolk phospholipids or of the
standard phospholipid mix (0.5 mg/mL) were prepared by ethanol
injection method13 in 0.1 mol/L phosphate buffer.
Unilamellar liposomes containing 10% Glc egg yolk PtdEtn were prepared
similarly. Liposomes in the phosphate buffer, including those
supplemented with Glc PtdEtn, were peroxidized by incubation with
10-mmol/L tert-butyl hydroperoxide as the oxidant. At different times,
aliquots were withdrawn. Lipids were extracted with chloroform-methanol
(2:1 vol/vol) and washed repeatedly with buffer to remove residual
tert-butyl hydroperoxide. Samples were analyzed by liquid
chromatography with online electrospray mass
spectrometry (LC/ESI/MS) immediately after being extracted.
Lipoprotein Isolation
LDL (1.019 to 1.069 g/mL) was obtained by density gradient
ultracentrifugation14 from plasma of
fasted normolipidemic individuals. LDL (2 mg protein per milliliter)
was subsequently dialyzed against 0.1 mol/L phosphate buffer (pH 7.4)
that contained 0.1 mmol/L EDTA for 24 hours (3 buffer changes).
LDL samples were sterilized by being passed through a 0.22-µm filter
(Millipore), stored at 4°C, and used within 1 week. Lipoprotein
concentration was determined by the method of Lowry et
al15 and expressed in milligrams per milliliter. Oxidation
of LDL (5 mg protein per 5 mL) was performed by dialysis against 5
µmol/L CuSO4 .5
H2O in 0.1 mol/L phosphate buffer, pH 7.4, for 24
hours at 37°C in the dark. Lipids were extracted into
chloroform/methanol (2:1 vol/vol) as described above.
Enrichment of LDL With PtdEtn
Glc and non-Glc PtdEtn was incorporated into LDL in a manner
similar to that described by Engelmann et al16 for
enriching human plasma lipoproteins with phospholipids. Glc PtdEtn (1
mg) in chloroform-methanol (2:1 vol/vol) was transferred to a 15-mL
test tube. The solvent was evaporated under nitrogen, and the lipids
were dispersed by centrifuging in 1.5 mL buffer that contained 50
mmol/L Tris/HCl, 1 mmol/L DTT, and 0.03 mmol/L EDTA (pH 7.4).
Solutions were sonicated in a bath sonicator for 5 minutes at 1-minute
intervals while being kept on ice under a stream of nitrogen. The
liposome mixture was centrifuged at 3500g, and the
supernatant was collected and passed through a 0.45µm filter. The
liposomal mixture (1 mL) was added to fresh plasma (4 mL) that
contained 3 mmol/L NaN3, under gentle
mixing. The mixture was incubated under nitrogen at 37°C for 24
hours. Lipoproteins were isolated as described above.
LDL Glucosylation
LDL (2 mg protein per milliliter) in PBS that contained 1
mmol/L EDTA, 0.1 mg/mL chloramphenicol, and 3 mmol/L
NaN3 was incubated with 5 to 400 mmol/L
glucose at 37°C for 1 week under nitrogen. Other incubation mixtures
contained 3 µCi/mL D-glucose-1-14C.
At the end of incubation, LDL samples were extracted with
chloroform/methanol (2:1 vol/vol). LDL samples that contained
radiolabeled glucose were reduced with NaBH4 for
1 hour at 4°C and dialyzed against 0.1 mol/L phosphate buffer that
contained 0.1 mmol/L EDTA for 24 hours to remove free radioactive
glucose. After dialysis was complete, lipids were extracted and
radioactivity measured in both protein and lipid fractions.
Extraction of Atherosclerotic Plaques
Atherosclerotic plaques obtained from patients during carotid
endarterectomy and from postmortem carotid arteries
(24 hours after death) were available from previous
studies.18 Plaques from intermediate to advanced lesions
were immediately placed into PBS, pH 7.4, that contained 0.1% EDTA.
Plaque material (100 mg) was separated from media and adventitia and
minced (0.5 to 1.0 mm2 pieces), and total
lipid extracts were prepared with chloroform/methanol (2:1
vol/vol).19 Tissue lipid extracts were stored in
chloroform at -20°C under nitrogen.
Analysis of Fatty Acid Methyl Esters
LDL lipid classes were isolated by preparative TLC on silica gel
H with heptane/isopropyl ether/acetic acid (60:40:4 by volume) as
solvent.20 In this system, phospholipids were retained at
origin, whereas free fatty acids, triacylglycerols,
and cholesteryl esters were resolved in order of decreasing polarity.
Appropriate areas of the plate were cleared of silica gel, and fatty
acid methyl esters (FAME) were prepared by treating the gel with 6%
H2SO4 in methanol for 2
hours at 80°C. Heptadecanoic acid was included as an internal
standard at 
10% of total fatty acid concentration. After reaction,
the FAME were extracted twice with hexane. The solvent was blown down
under nitrogen and the samples redissolved in hexane. Fatty acids were
analyzed on a polar capillary column (SP 2380, 15 mx0.32
mm ID, Supelco) installed in a Hewlett Packard model 5880 gas
chromatograph equipped with a flame ionization detector.
Hydrogen was the carrier gas at 3 PSI.21
Analysis of Phospholipid Classes
Normal-phase HPLC of phospholipids was performed on a 3-µm
Spherisorb column (100x4.6 mm ID) or a longer 5-µm Spherisorb
column (250x4.6 mm ID; Alltech Associates). Columns were
installed into a Hewlett-Packard model 1090 liquid
chromatograph and eluted with a linear gradient of 100%
solvent A (chloroform/methanol/30% ammonium hydroxide, 80:19.5:0.5 by
volume) to 100% solvent B (chloroform/methanol/water/30% ammonium
hydroxide, 60:34:5.5:0.5 by volume) for 14 minutes, then in 100%
solvent B for 10 minutes.10 22 Flow was set at 1 mL/min.
Peaks were monitored by online ESI/MS.
Analysis of Molecular Species of Phospholipids
Normal-phase HPLC with LC/ESI/MS was performed by splitting HPLC
flow by 1/50, which resulted in admission of 20 µL/mL to a
Hewlett-Packard model 5988B quadrupole mass spectrometer equipped with
a nebulizer-assisted electrospray interface (HP 59987A).22
Tuning and calibration of the system was achieved in the mass range of
400 to 1500 atomic mass units (AMU) by using of the standard
phospholipid mix dissolved in HPLC solvent A and flow-injected at 50
µL/min into the mass spectrometer. Capillary voltage was set at 4 kV,
endplate voltage at 3.5 kV, and cylinder voltage at 5 kV in the
positive mode of ionization. In the negative mode, voltages were 3.5,
3, and 3.5 kV, respectively. Both positive and negative ion spectra
were taken in the mass range 100 to 1100 AMU. Capillary exit (Cap Ex)
voltage was set at 120 Vin the positive and 160 V in the negative ion
mode. For fragmentation studies, Cap Ex voltage was raised to 300 V.
Nitrogen gas was used as both nebulizing gas (40 PSI) and drying gas
(60 PSI, 270°C). Phospholipids, including glucosylated diradylglycero
phosphoethanolamine, were quantified on the basis of standard curves
established for each phospholipid class and for the oxidized
phospholipids (core aldehydes, hydroperoxy, and isoprostanes). Oxidized
phospholipids that were used as standards were prepared as previously
described.20 The detection limit for Glc PtdEtn with
normal scanning range (400 to 1100 AMU) was 20 to 30 pmol, and the
response was linear to 100 nmol. Equimolar ion intensities of different
species of each phospholipid class varied by <5%23 in
each of the ion modes. LC/ESI/MS response to different phospholipid
classes varied greatly and required regular use of standards.
Measurement of Radioactivity
Radioactivity in glucosylated diradyl GroPEtn and proteins was
determined by scintillation counting after lipid extraction and TLC.
The silica gel that contained the Glc Etn phospholipids was scraped
into scintillation vials with CytoScint (ICN Pharmaceuticals). The
radioactivity in the protein precipitate was determined after the
precipitate was dissolved in CytoScint.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Isolation of Glc PtdEtn From Glucosylated LDL
Glc PtdEtn may occur in LDL as the Schiff base adduct or its Amadori rearrangement product (Figure 1
). The glucose adducts of PtdEtn were
originally recognized as part of the front of the PtdCho peak detected
by conventional normal-phase LC/ESI/MS as already described for total
plasma and red blood cell extracts of diabetic subjects.11
By use of longer HPLC columns, a complete resolution was obtained for
the PtdCho, PtdIns, and Glc PtdEtn components overlapping on the
shorter column. Furthermore, with the use of the negative ion mode, the
major PtdCho peak was seen only as minor PtdCho+Cl adduct, which
provided a clear view (Figure 2A) of the
Glc diradylGroPEtn peak after incubation of LDL with 50 mmol/L
glucose. Figure 2B shows single-ion plots of the major species
of both Glc PtdEtn and PlsEtn, which elute with the same retention time
as standard Glc PtdEtn. Figure 2C shows the full-mass spectrum
averaged over the elution time of the Glc diradyl GroPEtn peak. The
spectrum includes many molecular species, which are quantified in
Table 1. Table 1 shows that
the glucosylation of the ethanolamine phospholipids of LDL was largely
indiscriminate, given that both PtdEtn and alkenylacyl GroPEtn
components were glucosylated. However, on average, the diacyl species
represent a somewhat higher proportion of total glucosylated
ethanolamine phospholipids. The species present in untreated LDL
show a decrease in PlsEtn glycation compared with LDL incubated
with glucose in vitro. The 2 major species in both in vivo and in vitro
glycated LDL were 18:0-18:2 and 18:0-20:4 GroPEtn. The difference in
glycosylation in untreated LDL shows that the diacyl species
represent the majority of the glycated ethanolamine
phospholipids.
|
|
|
Although we investigated the possible presence of other aminophospholipid glycation products, such as glycated PtdSer or carboxymethylPtdEtn, in both untreated LDL and LDL incubated with glucose, we did not detect ions in the chromatogram pertaining to these products.
Figure 3 shows the fragment ions
obtained for the palmitoyllinoleoylGroPEtn glucosylate in the negative
and positive ionization mode as obtained for the HPLC peak eluting at
15.8 minutes. In the negative ion mode (Figure 3A),
fragmentation of the molecular ion (m/z 876) gave palmitic (m/z 255)
and linoleic (m/z 279) acids. The ion at m/z 712 is due to loss of
glucose moiety, whereas the ion at m/z 756 results from cleavage of the
attached glucose between carbons 2 and 3. In the positive ion mode
(Figure 3B), the ion at m/z 303 represents Glc PEtn and
the ion at m/z 576 the diacyl Gro fragment. The ion at m/z 180 was due
to cleavage of the adduct with a retention of nitrogen with the glucose
moiety. The same fragmentation pattern was observed for Glc PtdEtn
present in LDL, with more complex fragmentation patterns due to the
presence of multiple species. In both modes of ionization, the
fragments produced were consistent with either Schiff base or
Amadori rearrangement product of Glc PtdEtn, but the relatively
higher stability of the complex suggested the Amadori product, as
concluded previously by Lederer et al.24
|
Rate of Glucosylation of LDL
The concentration dependence of LDL lipid glucosylation was
investigated by determining the time course of Glc PtdEtn and Glc
PlsEtn accumulation in LDL by LC/ESI/MS. Figure 4 shows the concentration dependence of
LDL diradylGroPEtn glucosylation. Even at 5 mmol/L glucose
(24-hour incubation), 4-nmol/mg LDL protein Glc GroPEtn was
present. This represents 8.6% of total LDL diradylGroPEtn
(52 nmol/mg LDL protein) in glucosylated form at
physiological glucose concentrations. These levels
were close to the amounts found in untreated LDL (3.2 nmol/mg LDL
protein), indicative of in vivo levels of PtdEtn glycation. The level
of Glc diradylGroPEtn in LDL incubated at 400 mmol/L glucose for 7
days at 35 nmol/mg LDL protein represents 67% of total LDL
diradylGroPEtn being glucosylated. Even at these high glucose
concentrations, glycation of LDL diradylGroPEtn was incomplete. With
radiolabeled glucose, the differences between the rate of lipid and
protein glycation in LDL were investigated. Figure 5 shows that the initial rate of
incorporation of radioactivity was significantly higher in the lipid
than protein fraction. A TLC examination of the lipid phase indicated
that 92% of the radioactivity in the lipid pool was recovered in
the band that corresponded to the NaBH4 reduction
product of Glc diradylGroPEtn, and further analysis by
LC/ESI/MS gave the correct molecular weights for the reduced Glc
diradylGroPEtn products (data not shown). This demonstrates that
Glc diradylPEtn is the initial and major LDL glucosylation
product.
|
|
) as
measured by radiolabeled glucose incorporation. LDL was incubated with
20 mmol/L glucose containing 3 mCi/mL
D-glucose-1-[14C]. Incubation mixture was reduced with
NaBH4 and dialyzed extensively. Radioactivity was measured
by scintillation counting for both lipid and protein fractions and is
given as moles of glucose per LDL molecule. Molecular weight of ApoB
was assumed to be 500 kDa.
Relative Rates of Oxidation of PtdEtn and Glc PtdEtn
The effect of glycation on PtdEtn oxidation was determined by
comparing the rates of tert-butyl hydroperoxidation of glucosylated and
non-Glc PtdEtn incorporated at 10% mass into unilamellar egg yolk
PtdCho liposomes (Figure 6). Figure 6A shows the time course of hydroperoxidation of 18:0-18:2
GroPEtn and glucosylated 16:0-18:2 GroPEtn when incorporated into
PtdCho liposomes, as determined by LC/ESI/MS. Monohydroperoxides were
the major oxidized molecules monitored in both PtdEtn molecules. Glc
PtdEtn is noted to be peroxidized 2 to 3 times more readily than
non-Glc PtdEtn. Furthermore, Glc PtdEtn displayed pro-oxidant activity
toward non-Glc PtdEtn and PtdCho. Figure 6B shows that 16:0 to
20:4 and 16:0 to 18:2 GroPCho liposomes in presence of Glc PtdEtn were
peroxidized at a 2 to 3 times higher rate than the same PtdCho
liposomes supplemented with non-Glc PtdEtn.
|
Pro-oxidant Activity of Glc PtdEtn in LDL
To investigate further the apparent pro-oxidant effect of
Glc Etn phospholipids, we measured the peroxidation of LDL by copper
ions in the presence and absence of added glucosylated 16:0- 18:2
GroPEtn (10 to 30 nmol/mg LDL protein). This amount corresponds to
levels obtained for LDL incubation with 50 mmol/L for 48 hours.
Control LDL contained 10 to 20 nmol/mg LDL protein of added
nonglucosylated 16:0-18:2 GroPEtn. Extent of lipid peroxidation was
determined by measuring the hydroperoxides and core aldehydes of the
most abundant PtdCho species of the 2 LDL preparations. Figure 7 compares the phospholipid profiles of
untreated LDL (Figure 7A) and copper-oxidized LDL plus PtdEtn
(12 hours, 37°C) (Figure 7B) and copper-oxidized LDL+Glc
PtdEtn (12 hours, 37°C) (Figure 7C) as obtained by
positive-ion LC/ESI/MS. Figure 7A shows the resolution of the
PtdCho, SM, and lysoPtdCho, which are the major phospholipid components
of native LDL. The PtdCho, SM, and lysoPtdCho peaks are split because
of a resolution of the short and long carbon-chain species. After
copper-ion oxidation (Figure 7B), a dramatic change occurs in
the composition of the LDL phospholipids, largely as a result of
conversion of the unsaturated PtdCho into the hydroperoxides, core
aldehydes, and isoprostanes, which are eluted later than the saturated
PtdCho species. In addition, a relative increase has occurred in the
proportion of lysoPtdCho, possibly as a result of hydrolysis of the
peroxidized PtdCho by the phospholipases present in LDL. After
supplementation of the LDL with Glc PtdEtn (Figure 7C), a 4- to
5-fold increase occurs in the lipid peroxidation. The absolute amount
of LDL SM remained unchanged in both PtdEtn-supplemented (27.9% total
phospholipid) and Glc PtdEtn–supplemented LDL preparations (26.7%),
although this is not immediately obvious from either Figure 7B
or 7C.
|
Figure 8A shows the full-mass
spectrum averaged over the range of elution times of the hydroperoxy
PtdCho (17.070 to 18.603 minutes), whereas Figure 8B shows the
single-ion plots corresponding to the monohydroperoxy, dihydroperoxy
and isoprostane derivatives of the major PtdCho species. The
arachidonate-containing species (16:0-20:4 GroPCho) with 2
hydroperoxy groups (2xOOH) are resolved into 3 subfractions presumably
due to the presence of positional and cis, trans isomers of
the hydroperoxides.
|
Figure 9 shows the time course of
PtdCho hydroperoxide formation during dialysis against copper ions of
LDL supplemented with Glc PtdEtn or PtdEtn. The major PtdCho
hydroperoxides, 34:2 1xOOH (m/z 790), 36:2 1xOOH (m/z 818), 34:2 2xOOH
(m/z 822), and 36:4 2xOOH (m/z 846), are all increased
significantly in Glc PtdEtn supplemented LDL, with 34:2 1xOOH the most
abundant. LC/ESI/MS analysis of the NaBH4
reduced peroxidation products indicated that the hydroperoxy PtdCho
had been converted into the corresponding hydroxy PtdCho. Addition of
Glc PtdEtn to LDL resulted in about 5-fold increase in the formation of
phospholipid-bound hydroperoxides as calculated from averaged
differences from all time points.
|
Figure 10A shows the full-mass
spectrum averaged over the elution range (18.76 to 19.19 minutes), of
the C9 core aldehydes of PtdCho, which are eluted
ahead of the C5 core aldehydes. Figure 10B shows the full spectrum averaged over the entire range
(20.78 to 21.39 minutes) of the C5
aldehydes which trail into the SM species (m/z 675, palmitoyl
sphingosinephosphocholine). Figure 10C shows the single-ion
plots for all 4 major molecular species of the PtdCho core aldehydes.
The 9-oxo-nonanonates and 5-oxo-valerates of the palmitoyl and stearoyl
GroPCho represented the major components expected on the
basis of the fatty acid composition of the molecular species of LDL
PtdCho.
|
Figure 11 shows the time course of
PtdCho core aldehyde formation during copper oxidation of LDL
supplemented with Glc PtdEtn or PtdEtn. Major core aldehydes are
the 16:0-9:0 Ald and the 18:0-9:0 Ald species, as expected
from presence of linoleic acid as the major unsaturated fatty acid
(61%) in the sn-2 position of LDL PtdCho. NaBH4
treatment resulted in a reduction of the aldehydes to the corresponding
alcohols, as reflected in the increase of 2 AMU in the molecular ion of
each PtdCho core aldehyde (data not shown). On average, the increase in
core aldehyde formation as a result of Glc PtdEtn supplementation was
4-fold.
|
Addition of Glc PtdEtn to LDL also promoted oxidation of
cholesteryl esters in the interior of the particle. Figure 12 shows a more rapid loss of the 2
major species of polyunsaturated cholesteryl esters during dialysis
against copper ions of LDL supplemented with Glc PtdEtn when compared
with LDL supplemented with similar amounts of PtdEtn. Destruction of
the arachidonoyl ester occurred at a much faster rate than that of the
linoleoyl ester of cholesterol. Furthermore, the
cholesteryl esters appeared to undergo peroxidation much more rapidly
than the glycerophospholipids (compare Figures 10
and 12), which suggests that access to the interior of the particle
was facilitated by addition of Glc PtdEtn.
|
Isolation of Glc PtdEtn and PlsEtn From Atherosclerotic
Tissue
Figure 13 shows the total negative
ion current profile (Figure 13A), the single-ion chromatograms
for major species (Figure 13B), and the mass spectrum averaged
over the elution time (15.01 to 15.89 minutes) of the glycated
ethanolamine phospholipid peak (Figure 13C), as obtained by
normal-phase LC/ESI/MS for a total lipid extract of an atherosclerotic
plaque from a diabetic man. To increase the sensitivity of
detection, scanning was limited to a mass range of 850 to 1000 AMU,
which eliminated any overlap with the PtdEtn, PtdCho+Cl, and
SM+Cl ions also present in the total negative LC/ESI/MS profile. As
a result, only the dimers of the molecular species of lysoPtdCho
remained visible. Table 2 compares the
composition of the molecular species of glycated and nonglycated
ethanolamine phospholipids isolated from the atheroma.
Considerable discrepancy exists between the glycated and nonglycated
sets of species, which excludes direct in situ interconversion and
suggests possible deposition of these lipids from the circulation.
Composition of the molecular species of the ethanolamine phospholipids
of the atheroma also differs significantly from that of
human LDL (Table 1), with the relative proportion of the
polyunsaturated PtdEtn and specially PlsEtn species being reduced. The
most abundant glycated PlsEtn species in the plaque was 38:4 (m/z 928).
The possibility of selective glycation or oxidation is indicated by the
absence of glycated PtdSer species, although PtdSer species were
detected in the atheroma. In samples analyzed from
nondiabetic individuals, we found 2.4-nmol glycated diradylGroPEtn per
milligram total lipid, which represented 1.3% of the total
atheroma phospholipid. The atherosclerotic tissue from
diabetics contained 11.5-nmol glycated diradylGroPEtn per milligram
total lipid, which represented 7.3% of total
atheroma phospholipid, whereas an analysis of
normal aortic tissue did not show any PtdEtn glycation. These results
are comparable to levels of glycated PtdEtn previously measured by our
group in plasma of diabetic individuals (4.8% glycated GroPEtn per
milligram total phospholipid).11
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
LDL oxidation is claimed to play a central role in atherogenesis, which is exacerbated in diabetic hyperglycemia.1 Hyperglycemia is known to promote LDL glycation6 17 and lipid peroxidation,6 25 but the relative importance of protein and aminophospholipid glycation in these processes has not been established. Using previously prepared standards of Glc PtdEtn,10 we have now identified the lipid glycation product of LDL as a mixture of glycated diacyl and alkenylacylGroPEtn. We had previously shown10 11 that the glucose adduct of PtdEtn possesses a molecular weight that corresponds to either the Schiff base or its Amadori rearrangement product. In the present study, we have examined the glucosyl PtdEtn isolated from Glc LDL by pseudo-MS/MS with electrospray23 and have obtained fragment ions, which are consistent with both Schiff base and its Amadori rearrangement product. On the basis of the relatively high stability of the adduct, we must conclude that the structure is more likely to be that of the Amadori product rather than of the simple Schiff base. This conclusion is consistent with the recent report of Lederer et al,24 who have unequivocally proven that Amadori products are formed from D-glucose and PtdEtn in an in vitro model system. Furthermore, Requena et al26 have recently demonstrated the formation of carboxymethylEtn after hydrolysis of the aminophospholipid adduct. Bucala et al6 25 had shown in vitro that glycation of PtdEtn results in the formation of immunochemically detectable AGE. Furthermore, using ELISA, these authors concluded that the bulk of the AGE in LDL isolated from normal and diabetic subjects was located in the lipid phase.27 Pamplona et al28 had obtained evidence for glycated aminophospholipids in rat liver and for their increase in animals with streptozotocin-induced diabetes. With Amadori-product formation from glucose and PtdEtn now clearly established, we undertook further investigations on the effect of Glc PtdEtn on LDL lipid oxidation.
Of special interest to LDL lipid oxidation was the presence of the PlsEtn and its fate during glycation. Previous work had attributed both antioxidant29 and pro-oxidant30 properties to PlsEtn in various biological systems, although the effect of PlsEtn on LDL oxidation had not been directly studied. Myher et al31 showed that of the total diradylGroPEtn of plasma, 71.8% was due to alkenylacyl, 19.9% to diacyl, and 8.3% to alkylacyl species. The bulk of these phospholipids were assumed to originate in plasma LDL. The LDL preparation used in the present studies contained 49.4% PlsEtn, whereas glucosylated LDL contained 35% PlsEtn of total diradylGroPEtn. Thus, only 70% of the LDL PlsEtn was glucosylated under the present working conditions. Furthermore, the glucosylated species of both PtdEtn and PlsEtn tended to be more saturated than the respective nonglucosylated species. Analysis of diabetic plasma and red blood cells for glycated diradylGroPEtn in our earlier study11 had indicated a near absence of glycated PlsEtn species, whereas aminophospholipids isolated from red blood cells and plasma were glucosylated in vitro in proportion to their masses. Other studies5 have shown that hyperglycemic individuals have a significantly lower PtdEtn in LDL compared with normoglycemic controls, but it has not been established whether this is due to decreased PlsEtn. The loss of alkenylacyl species from glycated ethanolamine phospholipids could be due to their greater susceptibility to peroxidation,29 30 given that it has been demonstrated that the vinyl ether bond of alkenylacylGroPEtn is highly sensitive to peroxidation and can act as an antioxidant in protection of other phospholipids.32
Our study shows a more rapid and more extensive glucosylation of PtdEtn compared with apoprotein B in LDL. The distribution of radioactive glucose suggests that PtdEtn is more susceptible to glucosylation than apoB possibly due to more favorable environment for trapping the initial Schiff base adduct in the lipid phase. Our measurements of radioactive glucose distribution are consistent with the results of Bucala et al,25 who showed markedly increased levels of AGE in lipid compared with the aqueous (apoprotein) phase in diabetic and end-stage renal disease patients.
The present study shows that Glc PtdEtn is more easily peroxidized than PtdEtn, and that Glc PtdEtn, included in a liposomal mixture, promotes the oxidation of PtdEtn and PtdCho. Oxidation of phospholipids proceeds by way of hydroperoxides to the core aldehydes, all of which were isolated and identified as products of LDL oxidation. Likewise, addition of Glc PtdEtn to LDL caused more extensive oxidation of PtdEtn and PtdCho than addition of an equal amount of PtdEtn. This oxidation took place in the absence of protein glycation and free glucose in the incubation medium. The increased peroxidation of glucosylated ethanolamine phospholipids can probably be attributed to an increased exposure of these molecules to the oxidizing agent when incorporated into a liposome along with other phospholipids. The increased bulk of the polar-head group in the glucosylated aminophospholipid might permit a more ready access of the oxidizing agent to the unsaturated fatty chains of all glycerophospholipids in the liposomes and in the lipid monolayer of LDL. Furthermore, addition of Glc PtdEtn to LDL in absence of protein glycation resulted in increased oxidation of cholesteryl esters, which are located in the interior of the LDL particle. Clearly, Glc PtdEtn or its oxidation in the LDL lipid monolayer facilitates the access of free radicals to the cholesteryl esters in the interior of the LDL particle, which leads to a rapid destruction of the polyunsaturated cholesteryl esters. Previously it has been reported that LDL molecules from diabetic individuals show increased phospholipid layer fluidity, which could explain the increased susceptibility to oxidation.33
In the past glucose, has been claimed34 to possess pro-oxidant activity in both free and bound form, but more recently, increasing evidence has been obtained favoring the bound glucose pathway of oxidation of biological macromolecules. On the basis of albumin as a model protein, Hunt et al35 have concluded that oxidative alterations in experimental diabetes mellitus are due to protein-glucose adduct oxidation. The promotion of LDL lipid peroxidation by Glc PtdEtn demonstrated in the present study would also be consistent with pro-oxidant action of bound glucose. However, direct measurements of the oxidation potential of aminophospholipid-bound glucose in the absence of protein glucosylation suggest that the role of lipid glycation may have been overlooked. Hunt et al35 did not examine the effect of PtdEtn glycation in their study. In any event, the interplay between glycation, glycoxidation, and lipid peroxidation may be more complex than revealed by model studies, and the separate contributions of each step of the overall process to the final result have yet to be determined. Thus, complications may arise from the interaction of the oxidation products (hydroperoxides and core aldehydes) with free amino groups of both proteins and aminophospholipids as shown elsewhere.23 Greenspan et al36 have demonstrated that LDL associated with negatively charged phospholipids causes a dramatic increase in uptake and deposition of cholesteryl esters in J774 macrophages. We have recently shown that the glycation of PtdEtn not only results in increased negative charge but also promotes cholesteryl ester and triglyceride accumulation in macrophages.37 Extensive data have been reported on the involvement of protein AGE38 39 40 and lipid peroxidation in atheroma development.39 The identification of glycated PtdEtn and PlsEtn in atheroma suggests that glycated phospholipids may also be involved in plaque formation. The presence of PtdCho core aldehydes is of special interest beause these compounds are found in oxidized LDL41 and in atherosclerotic tissue.42 PtdCho core aldehydes have been specifically shown to induce increased monocyte-endothelial interactions in vitro,41 and PtdCho with a short chain aldehyde in the sn-2 position can mimic platelet-activating factor.43 44
In conclusion, the present study demonstrates that Glc PtdEtn is an early product of LDL glucosylation and that it is more readily peroxidized than PtdEtn. Furthermore, Glc PtdEtn promotes the peroxidation of LDL phospholipids and cholesteryl esters when included at a level of 10% of total diradylGroPEtn of LDL. These studies lend experimental support to previous speculation about the role of aminophospholipid glycation in LDL oxidation. This is the first report of isolation and identification of glucosylated diradylGroPEtn from LDL and human atherosclerotic tissue. The present study provides evidence that Glc PtdEtn may promote oxidation of LDL phospholipids and cholesteryl esters in hyperglycemia.
|
Acknowledgments |
|---|
This work was supported by the Heart and Stroke Foundation of Ontario, Toronto, Canada (A.K.) and Spectral Diagnostics Inc, Toronto, Canada (N.A.S.).
Received October 19, 1999; accepted July 30, 1999.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
1. Ross R. The pathogenesis of atherosclerosis. Nature. 1992;362:801–809.
2. Lyons T J, Jenkins AJ. Lipoprotein glycation and its metabolic consequences. Curr Opin Lipidol. 1997;8:174–180.[Medline] [Order article via Infotrieve]
3. Monnier VM. In: Baynes JW, Monnier VM, eds. The Maillard Reaction in Aging, Diabetes, and Nutrition. New York, NY: Alan R. Liss; 1989;1–22.
4. Witztum J L, Mahoney EM, Branks MJ, Fisher M, Elam R, Steinberg D. Nonenzymatic glucosylation of low-density lipoprotein alters its biologic activity. Diabetes. 1982;31:283–291.[Abstract]
5. Sobenin IA, Tertov VV, Koschinsky T, Bunting CE, Slavina ES, Dedov II, Orekhov AN. Modified low density lipoprotein from diabetic patients causes cholesterol accumulation in human intimal aortic cells. Atherosclerosis. 1993;100:41–54.[Medline] [Order article via Infotrieve]
6.
Bucala R, Makita Z, Koschinsky T, Cerami A,
Vlassara H. Lipid advanced glycosylation: pathway for lipid oxidation
in vivo. Proc Natl Acad Sci U S A. 1993;90:6434–6438.
7. Gugliucci A, Menini T, Stahl AJ. Susceptibility to copper-enhanced autoxidation of VLDL+LDL fractions from diabetic patients. Biochem Mol Biol Int.. 1994;32:139–147.[Medline] [Order article via Infotrieve]
8. Kobayashi K, Watanabe J, Umeda F, Nawata H. Glycation accelerates the oxidation of low density lipoprotein by copper ions. Endocrinol J.. 1995;42:461–465.
9. Hunt JV, Bottoms MA, Clare K, Skamarauskas JT, Mitchinson MJ. Glucose oxidation and low-density lipoprotein-induced macrophage ceroid accumulation: possible implications for diabetic atherosclerosis Biochem J.. 1994;300:243–249.
10. Ravandi A, Kuksis A, Marai L, Myher JJ. Preparation and characterization of glucosylated aminoglycerophospholipids. Lipids. 1995;30:885–891.[Medline] [Order article via Infotrieve]
11. Ravandi A, Kuksis A, Marai L, Myher JJ, Steiner G, Lewis G, Kamido H. Isolation and identification of glycated aminophospholipids from red cell and plasma of diabetic blood. FEBS Lett.. 1996;381:77–81.[Medline] [Order article via Infotrieve]
12.
Folch J, Lees M, Sloane-Stanley GH. A simple
method for the isolation and purification of total lipid from animal
tissues. J Biol Chem.. 1957;226:497–509.
13. Kremer JM, Esker MW, Pathmamanoharan C, Wiersema PH. Vesicles of variable diameter prepared by a modified injection method. Biochemistry. 1977;16:3932–3935.[Medline] [Order article via Infotrieve]
14. Havel RJ, Eder HA, Bragdon JH. The distribution and chemical composition of ultracentrifugally separated lipoproteins in human serum. J Clin Invest.. 1955;34:1345–1353.
15.
Lowry OH, Rosebrough NJ, Farr AL, Randall RJ.
Protein measurement with the folin phenol reagent. J Biol
Chem.. 1951;193:265–275.
16. Engelmann B, kogl C, Kulschar R, Schaipp B. Transfer of phosphatidylcholine, phosphatidylethanolamine and sphingomyelin from low- and high-density lipoprotein to human platelets. Biochem J.. 1996;315:781–789.
18. Kamido H, Nonaka K, Kuksis A, Marai L, Ravandi A. Identification of lipid ester core aldehydes and hydroperoxides in atherosclerotic lesions. Circulation. 1997:94(suppl I):I-707. Abstract.
19. Shaikh NA. Assessment of various techniques for the quantitative extraction of lysophospholipids from myocardial tissues. Anal Biochem.. 1994;216:313–321.
20. Ravandi A, Kuksis A, Myher JJ, Marai L. Determination of lipid ester ozonides and core aldehydes by high-performance liquid chromatography with on-line mass spectrometry. J Biochem Biophys Methods.. 1995;30:271–285.[Medline] [Order article via Infotrieve]
21. Yang LY, Kuksis A, Myher JJ. Intestinal absorption of menhaden and rapeseed oils and their fatty acid methyl and ethyl esters in the rat. Biochem Cell Biol.. 1990;68:480–491.[Medline] [Order article via Infotrieve]
22. Becart J, Chevalier C, Biese JP. Quantitative analysis of phospholipids by HPLC with a light scattering evaporating detector: application to raw materials of cosmetic use. J High Resol Chromatogr.. 1990;10:126–129.
23. Ravandi A, Kuksis A, Shaikh NA, Jackowski G. Preparation of Schiff base adducts of phosphatidylcholine core aldehydes and aminophospholipids, amino acids and myoglobin. Lipids. 1997;32:989–1001.[Medline] [Order article via Infotrieve]
24. Lederer MO, Drisbusch CM, Bundschuh RM. Amadori products from model reactions of D-glucose with phosphatidyl ethanolamine: independent synthesis and identification of 1-deoxy-1-(2-hydroxyethylamino)-D-fructose derivatives Carbohydr Res.. 1997;301:111–121.
25.
Bucala R, Makita Z, Vega G, Grundy S, Koschinsky
T, Cerami A, Vlassara H. Modification of low density lipoprotein by
advanced glycation end products contributes to the
dyslipidemia of diabetes and renal insufficiency.
Proc Natl Acad Sci U S A.. 1994;91:9441–9445.
26.
Requena JR, Ahmed MU, Fountain CW, Degenhardt TP,
Reddy S, Perez C, Lyons TJ, Jenkins AJ, Baynes JW, Thorpe SR.
Carboxymethylethanolamine, a biomarker of phospholipid modification
during the Maillard reaction in vivo. J Biol Chem.. 1997;272:17473–17479.
27.
Makita Z, Vlassara H, Rayfield E, Cartwright K,
Friedman E, Rodby R, Cerami A, Bucala R. Hemoglobin-AGE: a circulating
marker of advanced glycosylation. Science. 1992;258:651–653.
28. Pamplona R, Bellmunt MJ, Portero M, Riba D, Prat J. Chromatographic evidence for Amadori product formation in rat liver aminophospholipids. Life Sci. 1995;57:873–879.[Medline] [Order article via Infotrieve]
29. Hofner G, Lichtenberg D, Kostner GM, Hermetter A. Oxidation of fluorescent glycero- and sphingophospholipids in human plasma lipoproteins: alkenylacyl subclass are preferred targets. Clin Biochem.. 1996;29:45–50.
30. Reiss D, Beyer K, Engelmann B. Delayed oxidative degradation of polyunsaturated diacyl phospholipids in the presence of plasmalogen phospholipids in vitro. Biochem J. 1997;323:807–814.
31. Myher JJ, Kuksis A, Pind S. Molecular species of glycerophospholipids and sphingomyelins of human plasma: comparison to red blood cells. Lipids. 1989;24:408–418.[Medline] [Order article via Infotrieve]
32. Felde R, Spiteller G. Plasmalogen oxidation in human serum lipoproteins. Chem Phys Lipids. 1995;76:259–267.[Medline] [Order article via Infotrieve]
33. Rabini RA, Firretti G, Galassi R, Taus M, Curatola G, Tangorra A, Fumelli P, Mazzanti L. Modified fluidity and lipid composition in lipoproteins and platelets membranes from diabetic individuals. Clin Biochem.. 1994;27:381–385.[Medline] [Order article via Infotrieve]
34. Hunt JV, Wolff SP. Oxidative glycation and free radical production: a causal mechanism of diabetic complications. Free Radic Res Commun. 1991:12–13(pt 1):115–123.
35. Hunt JV, Bottoms MA, Mitchinson MJ. Oxidative alterations in the experimental glycation model of diabetes mellitus are due to protein-glucose adduct oxidation. Biochem J.. 1993;291:529–535.
36. Greenspan P, Ryu BH, Mao F, Gutman RL. Association of negatively-charged phospholipids with low-density lipoprotein (LDL) increases its uptake and the deposition of cholesteryl-esters by macrophages. Biochim Biophys Acta.. 1995;1257:257–264.[Medline] [Order article via Infotrieve]
37.
Ravandi A, Kuksis A, Shaikh NA. Glycated lipid
moiety of low-density lipoprotein promotes macrophage uptake
and accumulation of cholesteryl ester and
triacylglycerol. J Biol Chem.. 1999;274:16494–16500.
38. Hunt JV, Skamarauskas JT, Mitchinson MJ. Protein glycation and fluorescent material in human atheroma. Atherosclerosis. 1994;111:255–265.[Medline] [Order article via Infotrieve]
39. Nakamura Y, Horii Y, Nishino T, Shiiki H, Sakaguchi Y, Kagoshima T, Dohi K, Makita Z, Vlassara H, Bucala R. Immunohistochemical localization of advanced glycosylation end products in coronary atheroma and cardiac tissue in diabetes mellitus. Am J Physiol.. 1993;143:1649–1656.
40. Kamido H, Kuksis A, Marai L, Myher J. Lipid ester-bound aldehydes among copper-catalyzed peroxidation products of human plasma lipoproteins. J Lipid Res.. 1995;36:1876–1886.[Abstract]
41. Stitt AW, He C, Friedman S, Scher L, Rossi P, Ong L, Founds H, Li YM, Bucala R, Vlassara H. Elevated AGE-modified ApoB in sera of euglycemic, normolipidemic patients with atherosclerosis: relationship to tissue AGEs. Mol Med.. 1997;3:617–627.[Medline] [Order article via Infotrieve]
42.
Watson AD, Leitinger N, Navab M, Faull KF,
Hörkkö S, Witztum JL, Palinski W, Schwenke D, Salomon RG,
Sha W, Subbanagounder G, Fogelman AM, Berliner JA. Structural
identification by mass spectrometry of oxidized phospholipids in
minimally oxidized low density lipoprotein that induce
monocyte/endothelial interactions and evidence for
their presence in vivo. J Biol Chem. 1997;272:13597–13607.
43.
Stremler KE, Stafforini DM, Prescott SM, McIntyre
TM. Human plasma platelet-activating factor acetylhydrolase:
oxidatively fragmented phospholipids as substrates. J Biol
Chem.. 1991;266:11095–11103.
44. Heery JM, Kozak M, Stafforini DM, Jones DA, Zimmerman GA, McIntyre TM, Prescott SM. Oxidatively modified LDL contains phospholipids with platelet-activating factor-like activity and stimulates the growth of smooth muscle cells. J Clin Invest.. 1995;96:2322–2330.
This article has been cited by other articles:
 |
|
 |
|
  M. Wu, Y. Chen, K. Wilson, A. Chirindel, M. A. Ihnat, Y. Yu, M. E. Boulton, L. I. Szweda, J.-X. Ma, and T. J. Lyons Intraretinal Leakage and Oxidation of LDL in Diabetic Retinopathy Invest. Ophthalmol. Vis. Sci., June 1, 2008; 49(6): 2679 - 2685. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. Higuchi, K. Nakagawa, T. Tsuzuki, T. Suzuki, S. Oikawa, and T. Miyazawa Aminophospholipid glycation and its inhibitor screening system: a new role of pyridoxal 5'-phosphate as the inhibitor J. Lipid Res., May 1, 2006; 47(5): 964 - 974. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Nakagawa, J.-H. Oak, O. Higuchi, T. Tsuzuki, S. Oikawa, H. Otani, M. Mune, H. Cai, and T. Miyazawa Ion-trap tandem mass spectrometric analysis of Amadori-glycated phosphatidylethanolamine in human plasma with or without diabetes J. Lipid Res., November 1, 2005; 46(11): 2514 - 2524. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J.-H. Oak, K. Nakagawa, and T. Miyazawa UV analysis of Amadori-glycated phosphatidylethanolamine in foods and biological samples J. Lipid Res., March 1, 2002; 43(3): 523 - 529. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. J. Jarvisalo, A. Putto-Laurila, L. Jartti, T. Lehtimaki, T. Solakivi, T. Ronnemaa, and O. T. Raitakari Carotid Artery Intima-Media Thickness in Children With Type 1 Diabetes Diabetes, February 1, 2002; 51(2): 493 - 498. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Mazzali, T. Kipari, V. Ophascharoensuk, J.A. Wesson, R. Johnson, and J. Hughes Osteopontin--a molecule for all seasons QJM, January 1, 2002; 95(1): 3 - 13. [Full Text] [PDF]
|
|
|
|
 |
|
 U. Jurt, T. Gori, A. Ravandi, S. Babaei, P. Zeman, and J. D. Parker Differential effects of pentaerythritol tetranitrate and nitroglycerin on the development of tolerance and evidence of lipid peroxidation: a human in vivo study J. Am. Coll. Cardiol., September 1, 2001; 38(3): 854 - 859. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||