Decreased Protection by HDL From Poorly Controlled Type
2 Diabetic Subjects Against LDL Oxidation May Be Due to
the Abnormal Composition of HDL
Maya S. Gowri, Deneys R. Van der Westhuyzen, Susan R. Bridges, James W. Anderson
Downloaded from http://atvb.ahajournals.org/ by guest on June 14, 2017
Abstract—High plasma triglyceride concentrations in diabetic subjects increase their risk for developing coronary heart
disease. Numerous studies have shown that the high density lipoprotein (HDL) composition is abnormal in type 2
diabetic subjects. One study has shown that HDL (lipoprotein A-I) isolated from subjects with non–insulin-dependent
diabetes mellitus exhibits a decreased capacity to induce cholesterol efflux. The current study examined the effect of
HDL2 and HDL3 subfractions from poorly controlled type 2 diabetic and control subjects on THP-1 macrophage–
mediated low density lipoprotein (LDL) oxidation. The composition and protective effects of HDL2, but not of HDL3,
differed significantly between control and diabetic subjects. HDL2 from diabetics were triglyceride enriched and
cholesterol depleted compared with those from controls. Control HDL2 inhibited LDL oxidation, as assessed by lipid
peroxides and electrophoretic mobility, significantly (P,0.05) more than did diabetic HDL2 in both the fasting and
postprandial state. In addition, HDL2 from diabetics did not protect against apolipoprotein B-100 fragmentation in LDL.
Cross-linking in apolipoprotein A-I, oxidized in the presence of LDL, was extensive in HDL2 from diabetics compared
with that from controls. Serum triglyceride concentrations were negatively correlated with protection by HDL2
(r520.673, P,0.05) in diabetic but not in control subjects. HDL2-associated platelet-activating factor acetylhydrolase
activity was positively correlated with protection by HDL2 in control (r50.872, P,0.002) but not in diabetic subjects.
In conclusion, compositional alterations in HDL2 from poorly controlled type 2 diabetic subjects may reduce its
antiatherogenic properties. (Arterioscler Thromb Vasc Biol. 1999;19:2226-2233.)
Key Words: diabetes n hypertriglyceridemia n atherosclerosis n oxidized LDL n HDL
D
(CETP) activity8 that may in turn cause TG enrichment of
HDL at the expense of core cholesteryl esters.9,10 Increased
CETP activity has been reported in type 2 diabetes.10,11
In type 2 diabetic subjects, enhanced LDL oxidation occurs
in vivo, since high titers of autoantibodies to oxidatively
modified LDL are present in the plasma.12 Oxidative modification of LDL appears central to foam cell formation, the
earliest lesion of atherosclerosis.13 HDL may protect from
cardiovascular disease by a number of mechanisms, 1 of
which is by inhibiting LDL oxidation in the subendothelial
space.14 –16 HDL prevents the formation of minimally modified LDL in cocultures of artery wall cells by facilitating
hydrolysis of active, oxidized phospholipids to lysophospholipids, thereby destroying the biologically active lipids in
minimally modified LDL.17 Platelet-activating factor acetylhydrolase (PAF-AH) and paraoxonase (PON) are 2 enzymes
associated with HDL that are known to protect LDL from
oxidation.15,16,18
Cavallero and colleagues19 showed that LpA-I (HDL containing only apo A-I) isolated from subjects with non–insulindependent diabetes mellitus (NIDDM) exhibited a decreased
iabetic individuals have a significantly higher risk for
atherosclerosis than do nondiabetic individuals.1 Persons with type 2 diabetes have significantly increased serum
triglyceride (TG) and decreased HDL cholesterol concentrations. High plasma TG concentrations among diabetic subjects appear to be an independent predictive factor of mortality from coronary artery disease (CAD).2 Over the last
decade, the abnormal HDL composition in type 2 diabetes has
been well studied and is characterized by TG enrichment,
cholesterol depletion, and decreased apo A-I concentrations.3– 6 In subjects with type 2 diabetes, total serum TG
concentrations were found to be inversely correlated with
HDL cholesterol and positively correlated with HDL-TG
concentrations.6 Such abnormalities in HDL might contribute
to the well-established high risk for atherosclerosis in
diabetes.
Type 2 diabetic subjects also have a decreased lipoprotein
lipase activity7 that might cause decreased clearance of
TG-rich lipoproteins, leading to accumulation of these lipoproteins in the plasma. Elevated levels of TG-rich lipoproteins may increase the cholesteryl ester transfer protein
Received April 9, 1998; revision accepted January 19, 1999.
From the Metabolic Research Group, Veterans Affairs Medical Center and Department of Internal Medicine (M.S.G., J.W.A., S.R.B.), and the
Department of Internal Medicine (D.R.V.d.W.), College of Medicine, University of Kentucky, Lexington.
Correspondence to James W. Anderson, MD, Endocrine Section, Room 402B, VAMC, Cooper Drive Division, 2250 Leestown Rd 111C, Lexington,
KY 40511. E-mail [email protected]
© 1999 American Heart Association, Inc.
Arterioscler Thromb Vasc Biol. is available at http://www.atvbaha.org
2226
Gowri et al
Decreased Protective Function of HDL in NIDDM
capacity to induce cholesterol efflux in both the fasting and
postprandial state and hence, had a decreased ability to
accomplish their antiatherogenic role. The ability of compositionally abnormal HDL particles from type 2 diabetic
subjects to perform their antiatherogenic function of protecting LDL against oxidation has not been critically examined.
In our study, we examined the protection exhibited by HDL2
and HDL3 from subjects with poorly controlled type 2
diabetes against macrophage-mediated oxidation of LDL
compared with nondiabetic healthy controls. We found that
the HDL2 subfraction from diabetic subjects was abnormal in
composition, as previously reported,3– 6 and also exhibited
decreased protection against LDL oxidation when compared
with controls.
Methods
Materials
Downloaded from http://atvb.ahajournals.org/ by guest on June 14, 2017
Total cholesterol and free cholesterol kits (COD-PAP) and the total
phospholipid kit (phospholipids B) were purchased from Wako
Chemicals. 1-Palmitoyl-2-[(7-nitrobenzoxadiazoyl)amino]caproyl
phosphatidylcholine (C6NBD PC) was purchased from Avanti Polar
Lipids. Rabbit anti-human apo A-I and goat anti-human apo A-II
were purchased from Calbiochem. Electrophoresis reagents were
purchased from Bio-Rad. FCS, phorbol 12-myristate 13-acetate,
peroxidase-conjugated anti-rabbit IgG and anti-goat IgG, paraoxon,
glucose enzymatic kit (glucose oxidase), and the TG enzymatic kit
(INT reagent) were purchased from Sigma Chemical Co.
Study Protocol
Subjects
Ten poorly controlled type 2 diabetic men who had serum cholesterol
concentrations #240 mg/dL and serum TG concentrations #300
mg/dL were recruited from the Veterans Administration outpatient
clinic, and 10 matched male control subjects were also recruited
from the Veterans Administration Medical Center and University of
Kentucky Medical Center staff. Nine subjects were white and 1 was
African-American in each group. Medications taken by the diabetic
subjects at the time of the study were as follows: sulfonylurea (all 10
subjects), biguanide (3 subjects), nitrates (4 subjects), loop diuretics
(3 subjects), b-blockers (2 subjects), a-blockers (1 subject), and
angiotensin-converting enzyme inhibitors (5 subjects). Two of the
diabetic subjects had no diabetic complications, 1 subject had
cerebral artery occlusion, 4 subjects had CAD, and 3 subjects had
hypertension. Except for 1 diabetic subject who smoked ,1⁄2 pack of
cigarettes per day, all other subjects were nonsmokers. None of the
subjects took any antioxidant supplements or probucol.
Test Meal
We gave our subjects a fat load to see whether there was any
difference between fasting and postprandial HDLs in their antiatherogenic role of protecting LDL against oxidation. Each individual arrived at 8 AM after an overnight fast. Fifty milliliters of blood
was collected for lipid analysis and lipoprotein separation from
serum. Immediately after the blood was drawn, the test meal was
ingested within 10 minutes. The test meal consisted of 2 bowls of
cornflakes, a half pint of whole milk, a half pint of whipping cream,
and 2 teaspoons of granulated sugar. This meal provided '80 g of fat
and 900 calories. During the next 6 hours, the patients were told not
to consume any food or drink other than water or unsweetened black
coffee or tea. Fifty milliliters of blood was collected 6 hours after
ingestion of the test meal for lipoprotein isolation and lipid analysis.
The test meal was well tolerated by all of the subjects except 1 of the
diabetic subjects, who experienced nausea and vomiting.
Serum Measurements
Serum glucose concentrations were determined by using an enzymatic kit (glucose oxidase). Glycosylated hemoglobin (HbA1c) values were determined with an ion-exchange mini-column chromato-
2227
graphic procedure (Bio-Rad Laboratories); normal values are 4.2%
to 6.4%. Total cholesterol was measured with an enzymatic kit from
Wako Chemicals (COD-PAP) and total TGs by an enzymatic kit
from Sigma (INT reagent). HDL cholesterol was quantitated after
precipitation of apo B– containing lipoproteins with magnesium/
phosphotungstic acid. HDL3 cholesterol was quantitated after a
dual-precipitation method with combinations of magnesium/phosphotungstic acid and dextran sulfate. HDL2 cholesterol was calculated as the difference between total HDL cholesterol and HDL3
cholesterol.20 LDL cholesterol was calculated by the Friedewald
formula.21
LDL and HDL Subfraction Preparation
HDLs from the study subjects and LDL from a healthy donor were
isolated by sequential ultracentrifugation.22 LDL was isolated from
the same healthy donor for every oxidation experiment. In brief, the
density of serum was adjusted to 1.09 g/mL and centrifuged for 11
hours, 15 minutes at 50 000 rpm in a VTi50 rotor at 4°C. To isolate
LDL, the density of the VLDL-LDL fraction was adjusted to 1.3
g/mL, and gradient ultracentrifugation was carried out at 50 000 rpm
for 2.5 hours at 4°C. To isolate HDL, the density of the bottom
fraction was adjusted to 1.25 g/mL with solid KBr and centrifuged in
a VTi50 rotor at 50 000 rpm for 20 hours at 4°C. Gradient
ultracentrifugation was carried out at 4°C for 15 hours at 50 000 rpm
for fractionation of HDL.23 LDL, HDL2, and HDL3 were dialyzed
extensively overnight against 3 changes of PBS (pH 7.4), sterilized
through a 0.45-mm-filter unit (Millipore), and stored at 4°C under N2.
Aliquots of HDL subfractions were frozen at 270°C for analysis of
HDL phospholipid, cholesteryl ester, and free cholesterol
concentrations.
HDL Subfraction Composition
Total cholesterol and free cholesterol concentrations were measured
by enzymatic methods with the use of commercially available kits
(COD-PAP), and cholesteryl esters were calculated as the difference
between total and free cholesterol. TG concentrations were determined by using an enzymatic kit (INT reagent). Total phospholipids
were measured by an enzymatic method that uses phospholipase
D– choline oxidase and peroxidase (phospholipids B).
Total protein contents of the HDL subfractions were measured by
the Lowry method.24 Apo A-I and A-II in HDL subfractions were
quantitated by SDS–polyacrylamide gradient gel electrophoresis
(PAGE) followed by Western blotting.23 In brief, electrophoresis
was carried out on a 5% to 20% acrylamide/SDS gel and a
25 mmol/L Tris, 192 mmol/L glycine electrophoresis buffer (pH 8.4)
containing 0.1% SDS.23 The proteins were transferred onto a 0.2-mm
nitrocellulose membrane by using a 25 mmol/L Tris, 192 mmol/L
glycine transfer buffer (pH 8.4) containing 15% (vol/vol) methanol.23 The blots were blocked for 1 hour with 5% (wt/vol) nonfat dry
milk in PBS (pH 7.4). The blots were then exposed to the primary
and secondary antibodies for 1 hour each. The primary antibody for
apo A-I was rabbit anti-human apo A-I, and for apo A-II, goat
anti-human apo A-II. The secondary antibody was a peroxidaseconjugated anti-rabbit IgG and anti-goat IgG, respectively. The
antibodies were then detected with the enhanced chemiluminescence
(ECL) detection reagents (Amersham), and the blots were exposed to
x-ray film. The bands were scanned with a densitometer.
HDL Subfraction Enzyme Activities
PON and PAF-AH activities were measured in fresh HDL samples
within 2 days of isolation. HDL PON activity was measured with a
1.0 mmol/L paraoxon substrate in 50 mmol/L Tris-HCl buffer (pH
7.4) containing 1.0 mmol/L CaCl2 in a total volume of 800 mL.25 One
hundred micrograms of HDL2 or HDL3 was added to start the
reaction, and the increase in absorbance at 412 nm was recorded
continuously for 10 minutes. The amount of 4-nitrophenol formed
was calculated from the molar extinction coefficient of 12 800 M21
cm21. The blank contained substrate without HDL. One unit of PON
activity is defined as 1 nmol of 4-nitrophenol formed per minute
under the above assay conditions.
PAF-AH activity was measured by using a fluorescent substrate,
C6NBD PC, by a modification of the method of Steinbrecher and
2228
Arterioscler Thromb Vasc Biol.
September 1999
TABLE 1. Baseline Characteristics and Serum Lipid Profiles of Control and
Diabetic Subjects*
Controls
n
Diabetics
Fasting (CF)
Postprandial (CP)
10
Fasting (DF)
Postprandial (DF)
10
10
8
Age, y
48.762.6
zzz
61.562.1†
zzz
BMI, kg/m2
26.161.2
zzz
28.069.3
zzz
HbA1c, %
5.360.16
zzz
10.160.6‡
zzz
Fasting glucose, mg/dL
8662
zzz
212615.6‡
zzz
TG
120.6624.3
197.7638.6§
164.3622.5
254.0631.0\
Cholesterol
194.369.4
201614
197.767.7
197.268.1
LDL cholesterol
132.669.7
138.265.6
HDL cholesterol
37.662.2
zzz
36.461.7
Serum lipids, mg/dL
26.662.8¶
zzz
26.562.7#
Downloaded from http://atvb.ahajournals.org/ by guest on June 14, 2017
*Values are mean6SEM.
mg/dL: P vs control: †,0.002; ‡,0.0001; §,0.02, CF vs CP; \,0.002, DF vs DP; ¶,0.01, CF
vs DF; and #,0.005, CP vs DP.
Pritchard.26 Four micromoles of the substrate C6NBD PC was
incubated with 10 mg of HDL2 or HDL3 protein in 1 mL of PBS (pH
7.4) at 37°C for 90 minutes. The reaction was terminated by
vortexing with 1 mL of methanol and 1 mL of chloroform for '1
minute, and the mixture was then centrifuged at 2500g for 10
minutes. The fluorescence of the aqueous phase was measured at 470
nm excitation and 533 nm emission in a fluorospectrometer. The
mass of fluorescent substrate hydrolyzed was calculated by using a
standard curve generated by using C6NBD fatty acid diluted in
methanol.
Cell Culture
Dr Mark Kindy (Department of Biochemistry) provided the THP-1
monocytic cells. THP-1 cells were grown in suspension in RPMI1640 medium containing 10% FCS, 100 U/mL streptomycin, and
100 U/mL penicillin. Cells were maintained at a density of 106 mL21
by pelleting the cells twice weekly. Forty-eight hours before cellmediated oxidation studies, 23106 cells were seeded per well in a
12-well tissue-culture plate and induced to differentiate into macrophages by the addition of 1027 mol/L phorbol 12-myristate
13-acetate.27
Macrophage-Mediated Oxidation of Lipoproteins
The differentiated macrophages were washed 3 times with PBS
before incubation with lipoproteins. Macrophage cells were incubated with 100 mg of LDL protein, with or without 100 mg of HDL2
or HDL3 protein, in 1 mL of serum-free Ham’s F-10 medium
containing gentamicin (50 mg/mL) at 37°C in a 5% CO2 incubator
for 24 hours.28 Additionally, 100 mg of HDL2 or HDL3 protein was
incubated separately with the cells. Oxidation was arrested after 24
hours by the addition of 200 mmol/L EDTA and 20 mmol/L BHT,
and the medium was spun at 1200 rpm for 5 minutes at 4°C. Aliquots
of the medium were used for determining lipid peroxides, electrophoretic mobility, apo B-100 fragmentation in LDL, and apo A-I
cross-linking in HDL subfractions.
Assessment of Oxidative Damage
Lipid peroxides in the oxidized lipoproteins were determined by
mixing 0.1 mL of the medium containing the modified lipoproteins
with 1.0 mL of iodine-color reagent, incubating the mixture for 30
minutes at room temperature in the dark, and reading the absorbance
at 365 nm.29 The concentration of lipid peroxides was calculated
using a molar absorption coefficient of 2.463104 M21 cm21. Lipid
peroxides present in HDL samples oxidized separately were used as
a correction factor for the lipid peroxides formed in the LDL samples
containing HDL during oxidation.
Electrophoretic mobility was determined using 1.0% agarose gels.
Oxidized LDL was subjected to electrophoresis in 0.05 mol/L
barbital buffer (pH 8.6) at 90 V for 45 minutes.30 The gel was dried
and stained in Sudan black B. Electrophoretic mobility was measured as the distance from the origin to the median point of the
lipoprotein peak distribution. The increase in the electrophoretic
mobility of each oxidized LDL sample was calculated as a relative
ratio to that of native LDL and expressed as relative electrophoretic
mobility.
Apo B-100 Fragmentation
Fragmentation and aggregation of apo B-100 in the medium containing 1 mg of LDL, oxidized in the absence or presence of HDL2
or HDL3, was determined using 5% to 20% SDS-polyacrylamide
gradient gels under reducing conditions as detailed above. The gels
were fixed for 1 hour and stained with AgNO3 (Bio-Rad).
Apo A-I Cross-Linking
Apolipoprotein A-I cross-linking in macrophage-modified HDL2 or
HDL3 was determined by subjecting the medium containing 1 mg of
HDL protein to SDS-PAGE under reducing conditions,23 followed
by Western blotting with an antibody against apo A-I.
Data Analysis
Statistical analysis for comparison between groups was performed
using a 2-way ANOVA followed by post hoc testing with Tukey’s
multiple comparisons test by using SYSTAT statistics software.31
Results are expressed as mean6SEM. Correlation coefficients were
determined by simple linear regression analysis. Values of P,0.05
were considered to be statistically significant.
Results
Subject Characteristics
Table 1 provides the clinical characteristics and also the
serum lipid profiles of the diabetic and control subjects in
both the fasting and postprandial states. The diabetic subjects
were poorly controlled and had high HbA1c and fasting serum
glucose concentrations. Although body mass index did not
differ between the diabetic and control subjects, the diabetic
subjects were significantly older (P,0.002). Diabetic and
control subjects had similar serum total and LDL cholesterol
concentrations. Diabetic subjects had slightly but not significantly higher serum TG concentrations, compared with
controls, in both the fasting and postprandial states. However,
HDL cholesterol and HDL2 cholesterol concentrations were
significantly decreased in the diabetics in both the fasting
(P,0.01) and postprandial (P,0.005) states. Serum TG
Gowri et al
TABLE 2.
Decreased Protective Function of HDL in NIDDM
2229
Fasting and Postprandial HDL2 and HDL3 Compositions in Control and Diabetic Subjects*
HDL2
HDL3
Control
Fasting (CF)
n
Diabetic
Postprandial (CP)
Fasting (DF)
Control
Postprandial (DP)
Diabetic
Fasting (CF)
Postprandial (CP)
Fasting (DF)
Postprandial (DP)
10
10
10
8
Protein, %
42.662.2
43.762.5
44.662.2
44.662.9
53.460.76
57.060.64†
53.760.51
57.761.30‡
Phospholipid, %
29.363.2
29.563.1
28.363.3
27.064.8
27.560.65
31.560.77†
27.460.44
30.261.17§
Free cholesterol, %
3.7760.22
4.4160.21
2.9660.15\
3.0160.23¶
2.3760.12
2.9560.09†
2.2360.06
2.6060.16
Cholesteryl ester, %
19.060.8
16.060.6
14.961.1
13.361.6
12.560.28
3.360.09†
11.860.25
Triglyceride, %
5.3060.66
6.4060.69
9.3360.70\
12.061.28¶
4.2560.32
5.1960.33
4.9260.39
6.0660.45
Apo A-I#
0.48960.02
0.49760.02
0.42260.01\
0.44460.01¶
0.47160.03
0.49660.05
0.50860.03
0.44960.04
Apo A-II#
0.21160.01
0.23460.02
0.24360.02
0.24160.03
0.20260.02
0.22460.03
0.20460.01
0.19360.02
PON activity**
1.3560.39
1.3460.32
1.160.30
1.5360.63
7.6661.8
9.7464.0
8.662.1
14.163.5
PAF-AH activity**
437685
350668
569686
4986104
247666
207646
109611
123615.5
3.560.09‡
Downloaded from http://atvb.ahajournals.org/ by guest on June 14, 2017
*Values are mean6SEM.
†P,0.01 CF vs CP; ‡P,0.005 DF vs DP; §P,0.05 DF vs DP; \P,0.05 CF vs DF; and ¶P,0.05 CP vs DP.
#Values are in mg/mg HDL protein.
**Values are in nmol z min21 z mg21 apo AI.
concentrations were significantly increased postprandially
after a fat load in both the control (P50.014) and diabetic
(P50.001) subjects.
Composition of HDL2 and HDL3
Table 2 illustrates composition of HDL2 and HDL3 from
control and diabetic subjects. Two-way ANOVA indicated
that diabetic subjects exhibited significant compositional
abnormalities in their HDL2 fraction in both fasting and
postprandial states compared with controls. Diabetics had
(1) a decrease in HDL2 free cholesterol concentrations and
(2) an increase in HDL2-TG concentrations in both fasting
(P,0.05) and postprandial (P,0.05) states compared with
controls. Apo A-I concentrations were significantly decreased
in HDL2 from diabetic subjects compared with controls in
both fasting (P,0.05) and postprandial (P,0.05) states.
There was a slight but nonsignificant increase in the HDL2
apo A-II concentrations in diabetics compared with control
subjects. PAF-AH and PON activities in HDL2 were not
significantly different between control and diabetic subjects.
Two-way ANOVA indicated that the only significant
differences in HDL3 composition were associated with the
high-fat meal and not due to diabetes. We found no difference
in the HDL3 composition between the control and diabetic
subjects. In both control and diabetic subjects, HDL3 had a
significant increase in proteins (P,0.01) and phospholipids
(P,0.05) but a significant decrease in cholesteryl esters
(P,0.01) postprandially compared with the fasting state.
Free cholesterol concentrations in HDL3 were significantly
increased in the postprandial state (P,0.05) in control but not
in diabetic subjects. There was no significant difference in
either apo A-I and A-II concentrations or PON and PAF-AH
activities in HDL3 from control and diabetic subjects.
Protection Against Macrophage-Mediated
LDL Oxidation
Figure 1 illustrates lipid peroxide accumulation in LDL
oxidized in the absence and presence of HDL2 or HDL3 from
control and diabetic subjects in both fasting and postprandial
states. Lipid peroxides in LDL oxidized alone with the cells
was taken as 100%. Lipid peroxides in LDL oxidized in the
presence of HDL2 from diabetic subjects were significantly
higher, both in the fasting (53.368.3% LDL response versus
29.268.1%; P50.049) and postprandial (63.368.8% versus
35.367.7%; P50.01) states compared with controls. Lipid
peroxide accumulation in LDL oxidized in the presence of
HDL3 was not significantly different between control and
diabetic subjects. HDL2 was significantly more protective
than HDL3 both in fasting (29.268.1% versus 63.068.4%;
P50.002) and in postprandial (35.367.7% versus
57.967.7%; P,0.02) states in controls only but not in
diabetic subjects. We found a significant increase in the
electrophoretic mobility of LDL oxidized in the presence of
HDL2 from diabetic compared with control subjects, in both
Figure 1. Lipid peroxide accumulation in LDL incubated with
macrophages in the presence and absence of HDL2 or HDL3
from control (C) and diabetic (D) subjects in the fasting (CF,
n510; DF, n510) and postprandial (CP, n510; DP, n58) states.
LDL (100 mg/mL) was incubated with 100 mg of either HDL2 or
HDL3 for 24 hours at 37°C in Ham’s F-10 medium. Corrections
were made for the lipid peroxides generated by HDL in LDL
samples that contained HDL during oxidation. Values are
expressed as mean6SEM. *P,0.05 vs LDL1C-HDL2; 1P,0.05
vs LDL1C-HDL3.
2230
Arterioscler Thromb Vasc Biol.
September 1999
Downloaded from http://atvb.ahajournals.org/ by guest on June 14, 2017
Figure 2. Relative electrophoretic mobility of LDL incubated
with macrophages in the presence and absence of HDL2 or
HDL3 from control (C) and diabetic (D) subjects in the fasting
(CF, n510; DF, n510) and postprandial (CP, n510; DP, n58)
states. LDL (100 mg/mL) was incubated with 100 mg of either
HDL2 or HDL3 for 24 hours at 37°C in Ham’s F-10 medium. Values are expressed as mean6SEM. *P,0.05 vs LDL1C-HDL2;
1P,0.05 vs LDL1C-HDL3.
fasting (1.3960.05 versus 1.2260.06; P,0.05) and postprandial (1.5260.06 versus 1.3260.07; P,0.05) states (Figure
2). Again, HDL2 was more efficient in reducing the electrophoretic mobility of LDL compared with HDL3, both in
fasting (1.2260.06 versus 1.4960.07; P50.001) and postprandial (1.3260.07 versus 1.5360.07; P50.05) states in
controls only but not in the diabetic subjects. During oxidation, there was no difference in either the lipid peroxides or
the electrophoretic mobility of HDL2 or HDL3 from control
and diabetic subjects, in both the fasting and postprandial
state (data not shown).
Figure 3 illustrates the fragmentation of apo B-100 in LDL
when oxidized in the absence or presence of HDL2 from
control and diabetic subjects. HDL2 from control subjects, in
both fasting and postprandial states, protected the apo B-100
Figure 3. SDS-PAGE of LDL incubated with THP-1 macrophages in the absence or presence of HDL2 from control (C)
and diabetic (D) subjects, in both the fasting (CF, DF) and postprandial (CP, DP) states. LDL (100 mg/mL) was incubated with
100 mg of HDL2 from either the control or diabetic subjects for
24 hours at 37°C in Ham’s F-10 medium. A broad-range molecular weight marker was used in lane 1.
Figure 4. Western blotting of apo A-I in HDL2 from control (C)
and diabetic (D) subjects, in both the fasting (CF, DF) and postprandial (CP, DP) state, incubated with THP-1 macrophages in
the absence or presence of LDL. HDL2 (100 mg/mL) from either
the control or diabetic subjects was incubated in the absence or
presence of 100 mg of LDL for 24 hours at 37°C in Ham’s F-10
medium.
in LDL from fragmentation during oxidation to a significant
extent. However, the apo B-100 band in LDL oxidized in the
presence of HDL2 from diabetic subjects was almost completely fragmented.
Figure 4 illustrates that HDL2 from controls, when oxidized
in the presence of LDL, had only minimal apo A-I crosslinking in the fasting state, which was amplified in the
postprandial state. However, both fasting and postprandial
HDL2 from diabetic subjects, oxidized in the presence of
LDL, showed extensive apo A-I cross-linking when compared with control HDL2.
Linear Regression Analysis
Simple linear regression analysis was performed to assess
which components in the control and diabetic HDL2 might be
associated with their protective effects against LDL oxidation. Protection by control HDL2 against lipid peroxide
accumulation in LDL was strongly and positively correlated
with its associated PAF-AH activity both in the fasting
(r50.872, P50.002) (Figure 5a) and postprandial (r50.818,
P50.007) state (data not shown), whereas diabetic HDL2associated PAF-AH activity was not correlated with its
protection in either the fasting (Figure 5b) or postprandial
state (data not shown). We found that protection by control
HDL2 against lipid peroxide accumulation in LDL was
positively correlated with PAF-AH activity when expressed
either per milligram of apo A-I or per milligram of HDL
protein. No such correlation existed in the diabetic subjects,
even when the PAF-AH activity was expressed either per
milligram of apo A-I or per milligram of HDL protein. We
did not see any correlation between HDL2-associated PON
activity and protection by HDL2 from control or diabetic
subjects against lipid peroxide accumulation in LDL. In
controls, PAF-AH activity was significantly higher in HDL2
than in HDL3 (437 versus 247 nmol z min21 z mg21 apo A-I,
P,0.05), whereas PON activity was significantly higher in
HDL3 than in HDL2 (7.66 versus 1.35 nmol z min21 z mg21 apo
A-I, P,0.05).
Gowri et al
Decreased Protective Function of HDL in NIDDM
2231
Figure 5. Relationship between control
and diabetic fasting (CF, DF) HDL2associated PAF-AH activity and protection against LDL oxidation. LDL (100
mg/mL) was incubated with THP-1 macrophages along with 100 mg of HDL2 for
24 hours at 37°C in Ham’s F-10 medium.
Relationship between (a) CF HDL2
PAF-AH activity and % protection
against lipid peroxide accumulation in
LDL; (b) DF HDL2 PAF-AH activity and %
protection against lipid peroxide accumulation in LDL.
Downloaded from http://atvb.ahajournals.org/ by guest on June 14, 2017
Correlation analysis indicated that protection by HDL2
from control subjects against lipid peroxide accumulation in
LDL was not correlated with serum TG, HDL2-TG, cholesterol, phospholipid, or apoprotein concentrations. In contrast,
protection by HDL2 from diabetic subjects was inversely
correlated with serum TG concentrations in both the fasting
(r520.673, P,0.05) and postprandial (r520.798, P,0.02)
states (Figure 6a and 6b). Similarly, protection by HDL2 from
lipid peroxide accumulation in the diabetic subjects in the
fasting state was inversely correlated with HDL2-TG concentration (r520.636, P,0.05) (Figure 6c) and positively correlated with HDL2 free cholesterol concentration (r50.820,
P,0.005) (Figure 6d). In the diabetic subjects, we did not see
any correlation between the level of glycemic control and
HDL2 and HDL3 composition, PON and PAF-AH activities,
or protection against LDL oxidation (data not shown).
Discussion
Type 2 diabetic subjects present an increased risk for CAD
that is related to their degree of hypertriglyceridemia.3 Compositional abnormalities in HDL are well characterized in
NIDDM.3– 6 To determine whether abnormal HDL particles in
NIDDM are able to maintain antiatherogenic functions, we
examined the protective effects of HDL2 and HDL3 from
poorly controlled type 2 diabetic and nondiabetic control
subjects in both fasting and postprandial states against macrophage-mediated LDL oxidation.
Control subjects had a slight increase in phospholipid, free
cholesterol, and protein concentrations in both HDL2 and
HDL3 postprandially compared with the fasting state. Alimentary lipemia resulted in an increase in concentrations of
phospholipids and proteins in HDL, particularly in the HDL2
subfraction.32 The magnitude of postprandial lipemia determines the proportion of TGs in postprandial HDL2.33 In our
study, HDL2 from diabetic subjects were enriched in TGs and
depleted of free cholesterol and apo A-I compared with that
in controls. Other studies34,35 reported similar compositional
abnormalities in type 2 diabetes. Similar to our results,
Cavallero et al3 reported a significant postprandial decrease in
free cholesterol and cholesteryl ester concentrations and an
increase in TG concentrations in HDL2 from type 2 diabetics.
Increased cholesteryl ester transfer rates mediated by CETP
Figure 6. Relationship between serum TG,
HDL2-TG, and HDL2 free cholesterol concentrations and protection against LDL oxidation by
HDL2 from diabetic (D) subjects, in both the
fasting (DF) and postprandial (DP) state. a, DF
serum TG concentrations and % protection
against lipid peroxide accumulation in LDL by
DF HDL2; b, DP serum TG concentrations and
% protection against lipid peroxide accumulation in LDL by DP HDL2; c, DF HDL2 TG concentrations and % protection against lipid peroxide accumulation in LDL by DF HDL2; and d,
DF HDL2 free cholesterol concentrations and %
protection against lipid peroxide accumulation
in LDL by DF HDL2.
2232
Arterioscler Thromb Vasc Biol.
September 1999
Downloaded from http://atvb.ahajournals.org/ by guest on June 14, 2017
in type 2 diabetes10,11,36 appear important for the TG enrichment and cholesterol depletion seen in HDL2. In the diabetic
subjects, we found a positive correlation between serum TG
concentrations and HDL2-TG concentrations (data not
shown), as previously reported by Biesbroeck and colleagues.6 Postprandially, we found a significant TG enrichment of HDL2 in diabetic subjects only, although both control
and diabetic subjects had significant increases in their serum
TG concentrations postprandially compared with the fasting
state. Abbott et al37 have shown that serum PON activity is
lower in type 2 diabetic subjects than in controls. However,
we did not see a significant difference in HDL PON or
PAF-AH activity between diabetic and control subjects.
Effects of the individual HDL2 and HDL3 subfractions
against LDL oxidation in relation to their PAF-AH and PON
activities are poorly understood. When we examined the
protective effects of HDL subfractions against LDL oxidation
from control subjects, HDL2 was significantly more efficient
than HDL3 in inhibiting LDL oxidation in all parameters
assayed. Navab and colleagues38 have shown that virtually all
HDL-mediated inhibition of monocyte transmigration induced by minimally modified LDL can be accounted for by
HDL2, whereas HDL3 had no effect. The greater protective
effect of HDL2 seen in our study may be due to a higher
PAF-AH activity associated with HDL2 compared with
HDL3. In our study, HDL2 and HDL3 were not good substrates for macrophage oxidation compared with LDL. This
may be due to a decreased lipid content in HDL. Other studies
have also shown that unlike LDL, HDL is not susceptible to
modification by endothelial cells14 or in cocultures of artery
wall cells.38
In comparing the protective effects of HDL from control
and diabetic subjects against LDL oxidation, no differences
between HDL3 from control and diabetic subjects were
observed. On the other hand, HDL2 fractions from diabetic
subjects exhibited decreased protection against LDL oxidation compared with controls in both fasting and postprandial
states in most of the parameters assayed. There was increased
lipid peroxide accumulation and electrophoretic mobility;
also, their apo B-100 was almost completely fragmented
when LDL was oxidized in the presence of HDL2 from
diabetic compared with control subjects. In controls, in both
fasting and postprandial states, HDL2 exhibited significantly
greater protection compared with HDL3, but this difference
between HDL2 and HDL3 was not seen in diabetic subjects.
This suggests that HDL2 from diabetic subjects may have lost
its efficiency to protect LDL from oxidation. Correlation
analysis indicated that the increase in serum TG concentrations might mediate these effects by altering the composition
of HDL2. Decreased free cholesterol and increased TG
concentrations in HDL2 from diabetic subjects may contribute
importantly to the decreased protection exhibited by diabetic
HDL2.
Of importance, TG-enriched HDL2 from diabetic subjects,
though not susceptible to oxidation, exhibited decreased
protection against LDL oxidation compared with HDL2 from
controls. To determine whether diabetic HDL2 was being
more structurally altered in the presence of LDL compared
with control HDL2, we studied the apo A-I cross-linking in
HDL2. When HDL2 was oxidized in the presence of LDL,
there was more extensive apo A-I cross-linking in HDL2 from
diabetic subjects compared with controls. The greater apo A-I
cross-linking seen in HDL2 from diabetic subjects compared
with controls, when oxidized along with LDL, suggests that
compositionally altered diabetic HDL2 may not be as efficient
as control HDL2 in inhibiting lipid peroxide generation in
LDL. Therefore, the large amounts of aldehydes formed
during LDL oxidation may have modified the apo A-I in
diabetic HDL2. Aldehydes such as 4-hydroxynonenal and
malondialdehyde generated during LDL oxidation can cause
apo A-I cross-linking in HDL.39
Two possible mechanisms for the decreased protection
exhibited by HDL2 from diabetic subjects include the following: (1) The decreased free cholesterol concentration seen in
diabetic HDL2 may alter its surface fluidity, perhaps thereby
facilitating the lipid peroxides generated during LDL oxidation to “seed” the HDL2. Once seeded with lipid peroxides,
the TG-enriched HDL2 from diabetic subjects may be more
susceptible to oxidation and become less protective. (2) The
lipid peroxidation products transferred from oxidized LDL
may directly inactivate the PAF-AH activity in HDL2. Dentan
and colleagues40 have demonstrated that 4-hydroxynonenal
generated during LDL oxidation may inhibit PAF-AH activity, either by a direct modification of amino acid side chains
in PAF-AH or by a modification of the phospholipid environment of the enzyme at the surface of the lipoprotein
particle. In conclusion, our study demonstrates that HDL2
from subjects with poorly controlled NIDDM, in both fasting
and postprandial states, exhibits decreased protection against
macrophage-mediated LDL oxidation, and this may contribute to accelerated atherosclerosis in type 2 diabetes.
Acknowledgments
This research was supported in part by the High Carbohydrate Fiber
Nutrition Foundation (to J.W.A.) and by National Institutes of Health
(Bethesda, Md) grant RO1HL59376 (to D.R.V.d.W.). We gratefully
acknowledge Dr Frederick DeBeer for his support. We thank Dr
Marcielle DeBeer and Darin Cecil for their advice and
technical assistance.
References
1. Bierman EL. Atherogenesis in diabetes. Arterioscler Thromb. 1992;12:
647– 656.
2. Fontbonne A, Eschwege E, Cambien F, Richard J-L, Ducimetiere P,
Thibult N, Warnet J-M, Claude J-R, Rosselin G-E. Hypertriglyceridemia
as a risk factor of coronary heart disease mortality in subjects with
impaired glucose tolerance or diabetes: results from the 11-year follow-up
of the Paris Prospective Study. Diabetologia. 1989;32:300 –304.
3. Cavallero E, Dachet C, Neufcour D, Wirquin E, Mathe D, Jacotot B.
Postprandial amplification of lipoprotein abnormalities in controlled type
II diabetic subjects: relationship to postprandial lipemia and C-peptide/
glucagon levels. Metabolism. 1994;43:270 –278.
4. Bagdade JD, Buchanan WE, Kuusi T, Taskinen MR. Persistent abnormalities in lipoprotein composition in noninsulin-dependent diabetes after
intensive insulin therapy. Arteriosclerosis. 1990;10:232–239.
5. Howard BV, Xiaoren P, Harper I, Foley JE, Cheung MC, Taskinen M-R.
Effect of sulfonylurea therapy on plasma lipids and high-density
lipoprotein composition in non-insulin-dependent diabetes mellitus. Am J
Med. 1986;79(suppl 3B):78 – 85.
6. Biesbroeck RC, Albers JJ, Wahl PW, Weinberg CR, Bassett ML,
Bierman EL. Abnormal composition of high density lipoproteins in noninsulin-dependent diabetics. Diabetes. 1982;31:126 –131.
7. Taskinen M-R. Lipoprotein lipase in diabetes. Diabetes Metab Rev. 1987;
3:551–570.
8. Barter PJ, Jones ME. Kinetic studies of the transfer of esterified cholesterol between human plasma low and high density lipoproteins. J Lipid
Res. 1980;21:238 –249.
9. Tall A, Sammett D, Granot E. Mechanisms of enhanced cholesteryl ester
transfer from high density lipoproteins to apolipoprotein B-containing
Gowri et al
10.
11.
12.
13.
14.
15.
16.
Downloaded from http://atvb.ahajournals.org/ by guest on June 14, 2017
17.
18.
19.
20.
21.
22.
23.
24.
Decreased Protective Function of HDL in NIDDM
lipoproteins during alimentary lipemia. J Clin Invest. 1986;77:
1163–1172.
Bagdade JD, Lane JT, Subaiah PV, Otto ME, Ritter MC. Accelerated
cholesteryl ester transfer in non–insulin-dependent diabetes mellitus. Arteriosclerosis. 1993;104:69 –77.
Sutherland WHF, Walker RJ, Barned NJL, Pratt H, Tillman HC. The
effect of acute hyperinsulinemia on plasma cholesteryl ester transfer
protein activity in patients with non-insulin-dependent diabetes mellitus
and healthy subjects. Metabolism. 1994;43:1362–1366.
Bellomo G, Maggi E, Poli M, Agosta FG, Bollati P, Finardi G. Autoantibodies against oxidatively modified low-density lipoproteins in
NIDDM. Diabetes. 1995;44:60 – 66.
Fuster V, Badimon L, Badimon JJ, Chesebro JH. The pathogenesis of
coronary heart disease and the acute coronary syndromes. N Engl J Med.
1992;326:242–250.
Parthasarathy S, Barnett J, Fong LG. High density lipoprotein inhibits the
oxidative modification of low density lipoprotein. Biochim Biophys Acta.
1990;1044:275–283.
Mackness MI, Abbott C, Arrol S, Durrington PN. The role of highdensity lipoprotein and lipid-soluble antioxidant vitamins in inhibiting
low-density lipoprotein oxidation. Biochem J. 1993;294:829 – 843.
Watson AD, Navab M, Hama SY, Sevanian A, Prescott SM, Stafforini
DM, McIntyre TM, La Du BN, Fogelman AM, Berliner JA. Effect of
platelet activating factor-acetylhydrolase on the formation and action of
minimally oxidized low density lipoprotein. J Clin Invest. 1995;95:
774 –782.
Watson AD, Berliner JA, Hama SY, La Du BN, Faull KF, Fogelman AM,
Navab M. Protective effect of high density lipoprotein associated paraoxonase: inhibition of the biological activity of minimally oxidized low
density lipoprotein. J Clin Invest. 1995;96:2882–2891.
Stafforini DM, Zimmerman GA, McIntyre TM, Prescott SM. The platelet-activating factor acetylhydrolase from human plasma prevents oxidative modification of low density lipoprotein. Trans Assoc Am Phys.
1993;105:44 – 63.
Cavallero E, Brites F, Delfly B, Nicolaiew N, Decossin C, De Geitere C,
Fruchart J-C, Wikinski R, Jacotot B, Castro G. Abnormal reverse cholesterol transport in controlled type II diabetic patients: studies on fasting
and postprandial LpA-I particles. Arterioscler Thromb Vasc Biol. 1995;
15:2130 –2135.
Mackness MI, Durrington PN. Lipoprotein separation and analysis for
clinical studies. In: Converse CA, Skinner ER, eds. Lipoprotein Analysis:
A Practical Approach. New York, NY: Oxford University Press;
1992:1– 42.
Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the
preparative ultracentrifuge. Clin Chem. 1972;18:499 –502.
Strachan AF, de Beer FC, Van der Westhuyzen DR, Coetzee GA. Identification of three isoforms pattern of human serum amyloid A protein.
Biochem J. 1988;250:203–207.
Godenir NL, Jeenah MS, Coetzee GA, Van der Westhuyzen DR, Strachan
AF, de Beer FC. Standardization of the quantitation of serum amyloid A
protein (SAA) in human serum. J Immunol Methods. 1985;83:217–225.
Lowry OH, Rosebrough NJ, Farr AL, Randal RJ. Protein measurement
with the Folin phenol reagent. J Biol Chem. 1951;193:265–275.
2233
25. Gan KN, Smolen A, Eckerson HW, La Du BN. Purification of human
serum paraoxonase/arylesterase: evidence for one esterase catalyzing
both activities. Drug Metab Dispos. 1991;19:100 –106.
26. Steinbrecher UP, Pritchard PH. Hydrolysis of phosphatidylcholine during
LDL oxidation is mediated by platelet-activating factor acetylhydrolase.
J Lipid Res. 1989;30:305–315.
27. Via DP, Pons L, Dennison DK, Fanslow AE, Bernini F. Induction of
acetyl-LDL receptor activity by phorbol ester in human monocyte cell
line THP-1. J Lipid Res. 1989;30:1515–1524.
28. Leake DS, Rankin SM. The oxidative modification of low-density
lipoproteins by macrophages. Biochem J. 1990;270:741–748.
29. El-Saadani M, Esterbauer H, El-Sayed M, Goher M, Nassar AY, Jurgens
G. A spectrophotometric assay for lipid peroxides in serum lipoproteins
using a commercially available reagent. J Lipid Res. 1989;30:627– 630.
30. Noble RP. Electrophoretic separation of plasma lipoproteins in agarose
gel. J Lipid Res. 1968;9:693–700.
31. Wilkinson L. SYSTAT, the System for Statistics. Evanston, Ill: SYSTAT Inc:
1987.
32. Tall AR, Blum CB, Forester GP, Nelson CA. Changes in the distribution
and composition of plasma high density lipoproteins after ingestion of fat.
J Biol Chem. 1982;257:198 –207.
33. Patsch JR, Karlin JB, Scott L, Smith L, Gotto AM Jr. Inverse relationship
between blood levels of high density lipoprotein subfraction 2 and magnitude of postprandial lipemia. Proc Natl Acad Sci U S A. 1983;80:
1449 –1453.
34. Deckelbaum RJ, Granot E, Oschry Y, Rose L, Eisenberg S. Plasma
triglyceride determines structure-composition in low and high density
lipoproteins. Arteriosclerosis. 1984;4:225–231.
35. Durlach V, Attia N, Zahouani A, Leutenegger M, Girard-Globa A. Postprandial cholesteryl ester transfer and high density lipoprotein composition in normotriglyceridemic non-insulin-dependent diabetes patients.
Atherosclerosis. 1996;120:155–165.
36. Bhatnagar D, Durrington PN, Kumar S, Mackness MI, Dean J, Boulton
AJM. Effect of treatment with a hydroxymethylglutaryl coenzyme a
reductase inhibitor on fasting and postprandial plasma lipoproteins and
cholesteryl ester transfer activity in patients with NIDDM. Diabetes.
1995;44:460 – 465.
37. Abbott CA, Mackness MI, Kumar S, Boulton AJ, Durrington PN. Serum
paraoxonase activity, concentration, and phenotype distribution in
diabetes mellitus and its relationship to serum lipids and lipoproteins.
Arterioscler Thromb Vasc Biol. 1995;15:1812–1818.
38. Navab M, Imes SS, Hama SY, Hough GP, Ross LA, Bork RW, Valente
AJ, Berliner JA, Drinkwater DC, Laks H, Fogelman AM. Monocyte
transmigration induced by modification of low density lipoprotein in
cocultures of human aortic wall cells is due to induction of monocyte
chemotactic protein 1 synthesis and is abolished by high density
lipoprotein. J Clin Invest. 1991;88:2039 –2046.
39. McCall MR, Tang JY, Bielicki JK, Forte TM. Inhibition of lecithin-cholesterol acyltransferase and modification of HDL apolipoproteins by
aldehydes. Arterioscler Thromb Vasc Biol. 1995;15:1599 –1606.
40. Dentan C, Lesnik P, Chapman MJ, Ninio E. PAF-acether-degrading
acetylhydrolase in plasma LDL is inactivated by copper- and cellmediated oxidation. Arterioscler Thromb. 1994;14:353–360.
Downloaded from http://atvb.ahajournals.org/ by guest on June 14, 2017
Decreased Protection by HDL From Poorly Controlled Type 2 Diabetic Subjects Against
LDL Oxidation May Be Due to the Abnormal Composition of HDL
Maya S. Gowri, Deneys R. Van der Westhuyzen, Susan R. Bridges and James W. Anderson
Arterioscler Thromb Vasc Biol. 1999;19:2226-2233
doi: 10.1161/01.ATV.19.9.2226
Arteriosclerosis, Thrombosis, and Vascular Biology is published by the American Heart Association, 7272
Greenville Avenue, Dallas, TX 75231
Copyright © 1999 American Heart Association, Inc. All rights reserved.
Print ISSN: 1079-5642. Online ISSN: 1524-4636
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://atvb.ahajournals.org/content/19/9/2226
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Arteriosclerosis, Thrombosis, and Vascular Biology can be obtained via RightsLink, a service of the
Copyright Clearance Center, not the Editorial Office. Once the online version of the published article for
which permission is being requested is located, click Request Permissions in the middle column of the Web
page under Services. Further information about this process is available in the Permissions and Rights
Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Arteriosclerosis, Thrombosis, and Vascular Biology is online
at:
http://atvb.ahajournals.org//subscriptions/