64
Biochimica
et Biophysics Acta, 1046 (1990) 64-72 Biochimica et Biophysics Acta,
BBALIP 53467
Fluidity changes and chemical composition of lipoproteins in type IIa hyperlipoproteinemia C. Dachet ‘, C. Motta 2, D. Neufcour
’
and B. Jacotot
’
’ UnitP de Recherches SW les DyslipidPmies et I’AthtkosclProse (INSERM and ’ Laboratoire
de Biochimie,
HBtel-Dieu,
U 32), Hapita Henri-Mondor, Clermont-Ferrand (France)
Cr&eil
(Received 12 March 1990)
Key words: Lipoprotein structure; Hypercholesterolemia;
Fluorescence polarization
The chemical composition and the physical properties of lipoproteins (VLDL, LDL and HDL) were studied in two groups of patients: 14 healthy normolipidemic subjects and 15 type Da familial hypercholesterolemic patients. The steady-state fluorescence anisotropy r, was estimated in lipoproteins by the fluorescence depolarization of two fluorescent probes: the DPH (1,6-diphenyl-1,3,lhexatriene) and the TMA-DPH (1,4-trimethylammonium phenyld1,3,5_hexatriene). A structured order parameter S was calculated from the DPH fluorescence anisotropy. The flow activation energies were calculated for LDL and HDL from both groups from the Arrhenius plots (log r DPH versus l/ T). By using TNBS (trinitrobenzene sulfonic acid) as a distance control quencher, the two probes were located in the outer shell of LDL. In HDL, TMA-DPH remained at the surface of the particles, while DPH was more deeply embedded in the lipid core. There was no difference in the physico-chemical properties of VLDL between the two groups studied. DPH fluorescence anisotropies were significantly increased in LDL and HDL from the hypercholesterolemic group compared to the control particles (P < 0.05 and P < 0.01, respectively). In LDL this modification of the fluorescence anisotropy can be related to a change in the lipid composition of particles. LDL from hypercholesterolemic patients contained significantly less triacylglycerol (P < 0.01) and more cholesteryl ester (N.S.). Their cholesteryl ester to triacylglycerol ratio was significantly higher. In HDL, there was no difference in chemical composition between the two groups. The increase in DPH fluorescence anisotropy can be related to the presence of smaller particles in I-IDL from HC group. No difference was noted in the TMA-DPH fluorescence anisotropy at 37 o C in the LDL from the two groups. In contrast, TMA-DPH fluorescence anisotropy in HDL from hypercholesterolemic group was significantly higher than in control HDL. The flow activation energy of DPH was also significantly higher in both LDL and HDL from the hypercholesterolemic group than in control group particles. In both LDL and HDL from the control group, DPH fluorescence anisotropy was negatively correlated with TG/protein and TG/PL ratios and positively correlated with the CE/TG ratio. No correlation was observed between lipid composition and DPH fluorescence anisotropy values in hypercholesterolemic particles. The modification in fluidity parameters, especially the increase in the flow activation energies in LDL and HDL from hypercholesterolemic patients, could lead to a restriction of cholesterol movements in these particles. From a physiological point of view, this could represent a loss of functional capacity.
Introduction Low-density (LDL) and high-density (HDL) lipoproteins are the two major lipoprotein classes involved in cholesterol transport. Many epidemiological studies have shown that coronary heart disease is positively corre-
Correspondence: B. Jacotot, Unitt de Recherches sur les Dyslipidtmies et I’Athtrosclkrose (INSERM U 32), Hapital Henri Mondor, 94070, Crtteil Cedex, France.
lated with LDL levels and negatively correlated with HDL levels [l-3]. New information has been acquired concerning the heterogeneity in the composition and characterization of plasma LDL particles in normolipidemic subjects [4], as well as in hypercholesterolemic patients [5]. It has been suggested that some forms of LDL are more atherogenic than others and that the physico-chemical structure of particles may play an important role in their atherogenicity. Thus, LDL from patients with familial hypercholesterolemia are larger and contain
0005-2760/90/$03.50 0 1990 Elsevier Science Publishers B.V. (Biomedical Division)
65 more cholesteryl esters and less triacylglycerols than LDL from normal subjects [6]. Moreover, LDL from monkeys or swine fed and atherogenic diet were also large cholesteryl ester-enriched particles [7,8]. In monkeys, the presence of these enlarged LDL has been associated with the development of atherosclerotic lesions [9]. On the other hand, the study of the thermodynamic behavior of LDL in cholesterol-fed animals showed that the core cholesteryl esters were below the phase transition temperature and consequently less fluid at 37” C [lO,ll]. Human lipid nutrition studies have shown that LDL from subjects fed polyunsaturated fat as a hypocholesterolemic diet, are more fluid than those from subjects fed a saturated fat diet [12,13]. Finally, it was suggested that increased lipoprotein lipid phase fluidity could modify their interactions with cellular components [14,15]. The protective role of HDL in the pathogenesis of coronary heart disease can be attributed to their role in tissue cholesterol elimination [16]. We consider that the HDL plasma level is not the only important factor and that their physico-chemical structure could influence their efficacy in cholesterol reverse transport. Many works have investigated the thermodynamic behavior of HDL in normal subjects to study the structure of the particle itself [17], but few studies have considered the physico-chemical characteristics of HDL in different pathologies or after different diets. However, an increase was noted in HDL fluidity after a polyunsaturated fatty acid enriched diet in human beings [18]. These considerations prompted us to study the physico-chemical properties of lipoproteins, very-lowdensity lipoprotein (VLDL), low-density lipoprotein (LDL) and high-density lipoprotein (HDL) in normalipidemic subjects and in patients with type IIa familial hypercholesterolemia. The thermodynamic behavior of lipoproteins was studied by fluorescence polarization. Materials and Methods Subjects
Two groups of subjects were studied (Table I). The control group included 14 healthy normolipidemic subjects (9 women and 5 men) ranging in age from 20 to 40
years. None of these subjects was taking drugs affecting lipid metabolism. The hypercholesterolemic group included 15 type IIa hypercholesterolemic patients (11 women and 4 men) ranging in age from 17 to 50 years. None of these patients had been treated by lipid-lowering drugs in the previous 6 months. Ten had associated xanthomas. All secondary causes of hypercholesterolemia were excluded by appropriate tests. Lipoprotein
analyses
Blood samples for lipid analyses and lipoprotein isolation were drawn after a 16 h overnight fast. Lipoproteins were isolated by utlracentrifugation as described by Have1 [19]. A Beckman L8 centrifuge equipped with a 40.3 rotor was operated at 10 “C and 40000 rpm. The density range for the lipoproteins was as follows: VLDL, 1.006; LDL, 1.006-1.063; and HDL, 1.063-1.210. Each lipoprotein was refloated once at its density. Total free and esterified cholesterol, triacylglycerols and phospholipids were determined in serum and lipoprotein fractions, using commercial kits (Boehringer Mannheim for cholesterol and triacylglycerols, B test WAKO Biolyon for phospholipids). Protein concentrations were determined as described by Lowry [20] using serum albumin as the standard. Determination
of particle size
HDL particle size was determined by nondenaturing polyacrylamide gradient gel electrophoresis (Pharmacia PAA 4-30) according to the method of Nichols et al. [21]. After staining and destaining, the gels were scanned in a laser densitometer (LKB, Broma, Sweden). Thyroglobulin (Sigma) (1.75 mg/ml) with molecular diameter of 17.0 nm, was used as an internal standard. The size of HDL subfractions was calculated from a calibration curve using standards of known diameters. The standards included thyroglobulin (diameter 17.0 nm), apoferritin (diameter 12.2 nm), lactate dehydrogenase (diameter 8.1 mn) and bovine serum albumin (diameter 7.1 run). As described by Nichols et al. [21] and Blanche et al. [26], the separation of control HDL particles according to the size identified five subpopulations. The subpopulations, HDL,, and HDL,,, were located
TABLE I Plasma lipid and apolipoprotein levels and the distribution of cholesterol in lipoprotein fractions in the study groups The values are the means f S.D. Number of subjects
Plasma concentrations (mg/dl) cholesterol
TG
Ape A-I
Control ’
14
194*
36
70*21
164k32
99*17
Hypercholesterolemic b
15
363 f 115
76*28
134*45
168*61
i Controls, healthy normofipidemic subjects. Hypercholesterolemic, type IIa familial hypercholesterolemic patients.
AQO B
VLDL-C
LDL-C
9.6*5.4
135*
7.8 f 6.1
302k118
HDL-C 30
49f15
56521
66 within the ultracentrifuge HDL, subclass. Three subpopulations: HDL,,, HDL,,, HDL,, were located within the ultracentrifuge HDL, subclass.
sensitivity in the anisotropy measurements by making differential measurements either at 24 or at 37 o C. A fresh dispersion of fluorescent probe (DPH or TMA-DPH) was incubated in 3 ml of PBS 0.1 M (pH 7.4), with 100 to 500 ~1 of lipoprotein (VLDL, LDL, HDL). The probe concentration was chosen in respect to a final probe/phospholipid ratio always less than l/1000. The probe was incubated with lipoprotein for 1 h at 24°C with gentle agitation. After labelling, the fluorescence measurements were carried out on an Aminco SPF 500 spectrofluorimeter equipped with a fluorescence polarization accessory and a controlled temperature cell holder. The temperature was maintained at 0.5 o C and the excitation and emission wavelength settled to 360 and 460 nm, respectively.
Fluorescence polarization studies Lipoprotein fluidity was assessed by determining the steady-state fluorescence anisotropy rs from two probes: DPH (1,6-diphenyl-1,3,5-hexatriene) and TMA-DPH (1,4-trimethylammonium phenyl-6-1,3,5-hexatriene). (For a review of the theory and the technique, see Refs. 22, 23 and 24). A structural order parameter, S, for the lipid phase was calculated with the equation defined by Heyn [25]. Arrhenius plots (log r versus l/T) were used to calculate A E, the flow activation energy for each lipoprotein particle in the temperature range 1545°C. This thermodynamic parameter can be considered to be an index of physiological properties of some biological systems, such as membranes and lipoproteins. In order to relate the physical data to the biochemical composition of lipoproteins, it was necessary to locate as precisely as possible the fluorescent probes after controlled passive incorporation in the various particles. Trinitrobenzene sulfonic acid (T’NBS) was used as a distance control quencher (4.0 nm). TNBS binds to the surface lipoproteins and can therefore be used to measure the average depth of the probes. Considering the different localizations of the two probes and the difference in the chemical composition of the different parts of the particles, it was possible to enhance the TABLE
Statistical analyses All values are reported as mean rt_S.D. Data were analysed for statistical significance by Student’s t-test. Graphical data were subjected to linear regression analyses to determine the line of best fit. Results
Lipoprotein chemical composition No significant difference between the two groups was observed for the biochemical composition of VLDL and HDL (Table II). However, the CE/TC ratio was signifi-
II
Chemical composition of lipoproteins from variow groups All values are expressed
as mean f S.D. Lipid composition of lipoproteins (mg/lOO mg of proteins)
VLDL control hypercholesterolemic
LDL control hyperholesterolemic
HDL control hypercholesterolemic
* P < 0.05 vs. control. ** P i 0.01 vs. control. *** P i 0.001 VS. control.
CE/TC
CE/PL
CE/TG
TG/PL
total cholesterol
cholesteryl esters
triacylglycerol
phospholipids
(TC)
(CE)
(TG)
(PL)
99 It59 76 f34
52 *31 40 f20
356 f 248 305 *134
135 f73 102 *45
0.507 f 0.062 0.518 0.078
0.376 f 0.098 0.441 f 0.286
0.149 f 0.043 0.15 *0.13
2.53 f 0.41 3.29 f 1.68
135 rt25 153 f32
96 *15 117 k23
22.7 f4.6 17.4 * * f 4.2
90.7 * 9.0 97 k16
0.734 f 0.014 0.726 + 0.014
1.066 f 0.085 1.139 * f 0.078
4.44 f 0.99 6.57 * * * *1.44
0.253 f 0.047 0.181* * * f 0.039
5.6 i-2.8 4.7 f 1.8
48 *21 45 *17
0.828 f 0.032 0.808 * f 0.017
0.502 f 0.084 0.560 f 0.081
4.49 izl.33 5.41 f 1.53
0.118 k 0.025 0.114 f 0.044
29.2 lt15.3 30.4 f 10.3
20.6 k2.3 24.6 f 8.3
67
(2b)
(2b)
(Za) (3aX3bXJc) HDL Subpopulat ions
(Za)(3aX3bX3c) HDL Subpopulations
Fig. 1. Densitometric scans of the 4-30% polyacrylamide gradient gel electrophoresis of total HDL from two hypercholesterolemic patients (HCl and HC2) and two control subjects (Cl and C2). Scale at bottom corresponds to size intervals of five major HDL subpopulations identified in polyacrylamide gradient gel electrophoresis.
cantly reduced in the hypercholesterolemic HDL patients. The more significant changes occurred in LDL. The TG content of hypercholesterolemic LDL patients was reduced when compared to normal subjects (P < O.Ol), while CE and PL were slightly increased (N.S.). The CE/TG ratio was therefore significantly increased (P < O.OOl),while the TG/PL ratio was decreased (P < 0.001). HDL particle size
Densitometric
scans of the 4/30% polyacrylamide
gradient gel electrophoresis of the HDL from two hypercholesterolemic patients and two control subjects are shown in Fig. 1. These scans are representative of those obtained for the other studied HDL in each group. We observed in HDL from hypercholesterolemic patients the lack or the reduction of the larger-size subclasses (HDL,, or HDL,,) while the smaller size subspecies (HDL,, or HDL,,) were increased. Moreover, in HDL from HC patients, the mean particle size of two subfractions, HDL,,, and HDL,,, appeared to be significantly smaller than those of the normal subjects (Table III).
TABLE III High-density
lipoprotein particle size in I1 hypercholesterolemic
patients and 11 control subjects
High-density lipoprotein (HDL) particle sire (nm) was determined by 4-30s polyacrylamide gradient gel electrophoresis. HDL subfactions were identified as described by Nichols et al. [21] for densitometry scans. See Fig. 1. The values are the meansf S.D. HDL subfractions 2b
2a
3a
Hypercholesterolemic patients
10.7*0.1
9.6 f 0.2
8.6f0.2
Controls
10.9kO.2
9.7kO.2
8.8 *0.2
* P < 0.05 vs. control. ** P -z 0.01 vs. control.
*
3b
3c
8.OkO.2 * *
7.6 f 0.3
8.3 f 0.1
7.6 f 0.1
68 TABLE
IV
Probe location in LDL and HDL All results are expressed as the mean value obtained in lipoproteins from five controls or five hypercholesterolemic patients. (Percentage of the fluorescence probe quenched by trinitrobenzene sulfonate : TNBS). % of fluorescence quenched by TNBS DPH LDL HDL
12.3f3.5 r53.1*1.4
Th4A DPH LDL HDL
3.3 f 2.1 12.3 f 2.1
Lipoprotein jluorescence polarization studies
In LDL, almost all the fluorescence of DPH and TMA-DPH was quenched by TNBS. Table IV reports that only 12% and 3% of the initial fluorescence intensity remained for DPH and TMA-DPH, respectively. In HDL, the results were different. The residual fluorescence intensity after TNBS was 12% for TMA-DPH and 64% for DPH. These observations suggest that both DPH and TMA-DPH are located in the outer shell of LDL, while in HDL, TMA-DPH reflected phenomena of the outlayer and DPH, more deeply embedded, reflected the lipid dynamics of the core. As briefly explained in Material and Methods, it is possible to enhance the sensitivity of the fluorescence polarization measurements. Because of the localization of the probes, DPH reflects a very fluid part of the lipoproteins, especially in HDL, so it can be very practical to measure fuidity at 24” C, in order to enhance the perception of subtle changes which can be hindered at 37 OC because of the higher lipid dynamics. In contrast, TMADPH is located in high constraint zones of the particles (the surface of the outlayer) and it may be more judicious to work at 37’C, in order to enhance the lipid TABLE
Composition-structure correlations in lipoproteins
Table VI reports the different correlations observed between DPH fluorescence anisotropy and the chemical composition in lipoproteins from the both groups. In VLDL from both groups a negative correlation was observed between fluorescence anisotropy r and the TG/protein ratio (-0.5179 for controls and - 0.5105 for hypercholesterolemic patients, P < 0.05). A more significant correlation existed with the TG/PL ratio (- 0.6020 for controls, P < 0.05 and - 0.6484 for HC, P < 0.02). No significant correlation was observed with the CE/TG ratio. In LDL from normal subjects, DPH fluorescence anisotropy was negatively correlated with TG/protein (-0.7834, P < 0.01) and TG/PL (- 7357, P < 0.01)
V
Steady-state fluorescence anisotropy rs of DPH and TMA-DPH
DPH 24°C Fluorescence anisotropy
r,
Structure order parameter S TMA-DPH 37 o C Fhiorescence anisotronv r.
and structure order parameter
of lipid phase S in lipoproteins from various groups HDL
LDL
VLDL
The * ** l **
dynamic of these parts and to allow more perceptible TMA-DPH movements and their influence on the polarization measurements. In the LDL and HDL from hypercholesterolemic patients, the mean DPH fluorescence anisotropy r was higher than in the control group (P < 0.05 and P < 0.01, respectively) (Table V). The order parameter S was also significantly enhanced in both LDL and HDL (P < 0.01). No particular changes in VLDL were reported by DPH or TMA-DPH in the two study groups. No difference was noted in the TMA-DPH fluorescence anisotropy in LDL from either groups. However, in HDL, TMA-DPH reported higher values of fluorescence anisotropy in hypercholesterolemic patients than in controls (P < 0.001). The study of the thermodynamic behavior of LDL and HDL from the two groups with DPH between 15’ and 45 ’ C showed that all systems were invariant phases (Fig. 2). The flow activation energy A E was calculated from Arrhenius plots. The mean A E values were higher in LDL and HDL from hypercholesterolemic patients than in control particles (P < 0.01 and P < 0.001, respectively).
control
hyperCT
control
hyperCT
0.033
0.259 f 0.010
0.272 f 0.015 *
0.229 f 0.006
0.244 f 0.014 * *
0.449 * 0.079
0.501 kO.106
0.825 f 0.024
0.849 f 0.035 * *
0.752 k 0.016
0.786 f 0.033 * *
0.212 f 0.068
0.235 f 0.067
0.311 f 0.014
0.317 * 0.010
0.257 f 0.013
0.283 & 0.018 * * *
control
hyperCT
0.131~0.018
0.142*
values are expressed as the mean f S.D. P < 0.05 vs. control. P < 0.01 vs. control. P -z 0.001 vs. control.
69
A
9876-
AE (kcal LDL control LDL HC
/mole) 6.94 9.03
f 1.0/i + 1.36a
20.69
HDL control
6.52
HOL HC
9.30 + 0.75b
2-
L
I
I
I
I
3.2
3.3
3.L
3.5
l/T
)
I
I
I
I
3.2
3.3
3.4
x lo3
l/T
b
x103
Fig. 2. Arrhenius plots of steady-state microviscosity of DPH and flow activation energy AE in LDL (left panel) and HDL (right panel)from ) and hypercholesterolemic (- - - - - -) groups. Flow activation energy of DPH was determined between 15 and 45 o C. Statistical control (test: (a) P < 0.01 vs. control group; (b) P < 0.001 vs. control group.
ratios and positively correlated with CE/TG ratio (0.6976, P -z0.01). In hypercholesterolemic LDL, there was no correlation with the TG/protein and CE/TG ratios (Fig. 3). Only a slightly negative correlation was noted with the TG/PL ratio (-0.5376, P < 0.05). Very similar results were observed in HDL. In the HDL from the control group, DPH fluorescence anisotropy was negatively correlated with the TG/protein ratio (-0.6199, P < 0.02) and to a less extent with the TG/PL ratio (-0.5605, P < 0.05). A good positive correlation was noted with the CE/TG ratio (0.6649,
P -c0.01). None of these correlations was observed in HDL from hypercholesterolemic
patients (Fig. 4).
Discussion In the present work, we examined the physical state of lipoproteins (VLDL, LDL and HDL) from hypercholesterolemic or control subjects related to their chemical composition or their size. The physical lipoprotein properties were studied by fluorescence depolarisation of two fluorescent probes DPH and Th4A-
TABLE VI Correlations between DPH fluorescence anisotropy and lipid composition in lipoproteins isolated from various groups Control group r VLDL Triacylglycerol/proteins Triacylglycerol/phospholipids Cholesteryl esters/triacylglycerol
-0.5179 - 0.6020 0.2483
LDL Triacylglycerol/proteins Triacylglycerol/phospholipids Cholesterol ester/triacylglycerol
-0.7834 -0.7357 0.6976
HDL Triacylglycerol/proteins Triacylglycerol/phospholipids Cholesteryl esters/triacylglycerol
- 0.6199 - 0.5605 0.6649
P: Correlation coefficient. * PcO.05; ** P < 0.02; *** P < 0.01.
Hypercholesterolemic group P * * n.s. *** *** ***
** * ***
r
P
- 0.5105 - 0.6484 0.3447
* ** n.s.
- 0.2962 - 0.5376 0.4616
n.s.
- 0.0064 - 0.2389 0.2332
n.s. n.s. n.s.
n.s. *
70 LDL
Control
i
A
B 0
0.3
a
1
0
0.2 c
0.3 -
og
0
\
HC LDL
u
a -.
0 00
0.2 -
w
I-
l
s
8.
c
0
n 0.11 220
I
I 240
I
260 I
9
a
0
\
280
8
l l
1
240
I
280
260
x 1000
a
l l
l
I
O.I220
300
300
r x 1000
c
D
9-
7
d
,”
5-
z
3-
\
3i
1
I
I
I
1
220
240
260
280
300
c
l
l
l
I
.i
.
m
l
8 8
m
l
m
l
m
11 220
l
:
8
8
I
I
I
1
240
260
280
300
r x 1000
r x 1000
Fig. 3. Plots of steady-state DPH fluorescence anisotropy rS vs. triacylglycerol to protein (A, B) and cholesterol ester to triacylglycerol (C, D) ratios in LDL from control (left panel) and hypercholesterolemic (right panel) groups. Each point represents values for one subject. The lines of best fit are shown for control LDL. Correlation coefficients are given in Table VI. Control
HC
HDL
HDL
9
#
B
00
.
7
5
Biochimica
et Biophysics Acta, 1046 (1990) 64-72 Biochimica et Biophysics Acta,
BBALIP 53467
Fluidity changes and chemical composition of lipoproteins in type IIa hyperlipoproteinemia C. Dachet ‘, C. Motta 2, D. Neufcour
’
and B. Jacotot
’
’ UnitP de Recherches SW les DyslipidPmies et I’AthtkosclProse (INSERM and ’ Laboratoire
de Biochimie,
HBtel-Dieu,
U 32), Hapita Henri-Mondor, Clermont-Ferrand (France)
Cr&eil
(Received 12 March 1990)
Key words: Lipoprotein structure; Hypercholesterolemia;
Fluorescence polarization
The chemical composition and the physical properties of lipoproteins (VLDL, LDL and HDL) were studied in two groups of patients: 14 healthy normolipidemic subjects and 15 type Da familial hypercholesterolemic patients. The steady-state fluorescence anisotropy r, was estimated in lipoproteins by the fluorescence depolarization of two fluorescent probes: the DPH (1,6-diphenyl-1,3,lhexatriene) and the TMA-DPH (1,4-trimethylammonium phenyld1,3,5_hexatriene). A structured order parameter S was calculated from the DPH fluorescence anisotropy. The flow activation energies were calculated for LDL and HDL from both groups from the Arrhenius plots (log r DPH versus l/ T). By using TNBS (trinitrobenzene sulfonic acid) as a distance control quencher, the two probes were located in the outer shell of LDL. In HDL, TMA-DPH remained at the surface of the particles, while DPH was more deeply embedded in the lipid core. There was no difference in the physico-chemical properties of VLDL between the two groups studied. DPH fluorescence anisotropies were significantly increased in LDL and HDL from the hypercholesterolemic group compared to the control particles (P < 0.05 and P < 0.01, respectively). In LDL this modification of the fluorescence anisotropy can be related to a change in the lipid composition of particles. LDL from hypercholesterolemic patients contained significantly less triacylglycerol (P < 0.01) and more cholesteryl ester (N.S.). Their cholesteryl ester to triacylglycerol ratio was significantly higher. In HDL, there was no difference in chemical composition between the two groups. The increase in DPH fluorescence anisotropy can be related to the presence of smaller particles in I-IDL from HC group. No difference was noted in the TMA-DPH fluorescence anisotropy at 37 o C in the LDL from the two groups. In contrast, TMA-DPH fluorescence anisotropy in HDL from hypercholesterolemic group was significantly higher than in control HDL. The flow activation energy of DPH was also significantly higher in both LDL and HDL from the hypercholesterolemic group than in control group particles. In both LDL and HDL from the control group, DPH fluorescence anisotropy was negatively correlated with TG/protein and TG/PL ratios and positively correlated with the CE/TG ratio. No correlation was observed between lipid composition and DPH fluorescence anisotropy values in hypercholesterolemic particles. The modification in fluidity parameters, especially the increase in the flow activation energies in LDL and HDL from hypercholesterolemic patients, could lead to a restriction of cholesterol movements in these particles. From a physiological point of view, this could represent a loss of functional capacity.
Introduction Low-density (LDL) and high-density (HDL) lipoproteins are the two major lipoprotein classes involved in cholesterol transport. Many epidemiological studies have shown that coronary heart disease is positively corre-
Correspondence: B. Jacotot, Unitt de Recherches sur les Dyslipidtmies et I’Athtrosclkrose (INSERM U 32), Hapital Henri Mondor, 94070, Crtteil Cedex, France.
lated with LDL levels and negatively correlated with HDL levels [l-3]. New information has been acquired concerning the heterogeneity in the composition and characterization of plasma LDL particles in normolipidemic subjects [4], as well as in hypercholesterolemic patients [5]. It has been suggested that some forms of LDL are more atherogenic than others and that the physico-chemical structure of particles may play an important role in their atherogenicity. Thus, LDL from patients with familial hypercholesterolemia are larger and contain
0005-2760/90/$03.50 0 1990 Elsevier Science Publishers B.V. (Biomedical Division)
65 more cholesteryl esters and less triacylglycerols than LDL from normal subjects [6]. Moreover, LDL from monkeys or swine fed and atherogenic diet were also large cholesteryl ester-enriched particles [7,8]. In monkeys, the presence of these enlarged LDL has been associated with the development of atherosclerotic lesions [9]. On the other hand, the study of the thermodynamic behavior of LDL in cholesterol-fed animals showed that the core cholesteryl esters were below the phase transition temperature and consequently less fluid at 37” C [lO,ll]. Human lipid nutrition studies have shown that LDL from subjects fed polyunsaturated fat as a hypocholesterolemic diet, are more fluid than those from subjects fed a saturated fat diet [12,13]. Finally, it was suggested that increased lipoprotein lipid phase fluidity could modify their interactions with cellular components [14,15]. The protective role of HDL in the pathogenesis of coronary heart disease can be attributed to their role in tissue cholesterol elimination [16]. We consider that the HDL plasma level is not the only important factor and that their physico-chemical structure could influence their efficacy in cholesterol reverse transport. Many works have investigated the thermodynamic behavior of HDL in normal subjects to study the structure of the particle itself [17], but few studies have considered the physico-chemical characteristics of HDL in different pathologies or after different diets. However, an increase was noted in HDL fluidity after a polyunsaturated fatty acid enriched diet in human beings [18]. These considerations prompted us to study the physico-chemical properties of lipoproteins, very-lowdensity lipoprotein (VLDL), low-density lipoprotein (LDL) and high-density lipoprotein (HDL) in normalipidemic subjects and in patients with type IIa familial hypercholesterolemia. The thermodynamic behavior of lipoproteins was studied by fluorescence polarization. Materials and Methods Subjects
Two groups of subjects were studied (Table I). The control group included 14 healthy normolipidemic subjects (9 women and 5 men) ranging in age from 20 to 40
years. None of these subjects was taking drugs affecting lipid metabolism. The hypercholesterolemic group included 15 type IIa hypercholesterolemic patients (11 women and 4 men) ranging in age from 17 to 50 years. None of these patients had been treated by lipid-lowering drugs in the previous 6 months. Ten had associated xanthomas. All secondary causes of hypercholesterolemia were excluded by appropriate tests. Lipoprotein
analyses
Blood samples for lipid analyses and lipoprotein isolation were drawn after a 16 h overnight fast. Lipoproteins were isolated by utlracentrifugation as described by Have1 [19]. A Beckman L8 centrifuge equipped with a 40.3 rotor was operated at 10 “C and 40000 rpm. The density range for the lipoproteins was as follows: VLDL, 1.006; LDL, 1.006-1.063; and HDL, 1.063-1.210. Each lipoprotein was refloated once at its density. Total free and esterified cholesterol, triacylglycerols and phospholipids were determined in serum and lipoprotein fractions, using commercial kits (Boehringer Mannheim for cholesterol and triacylglycerols, B test WAKO Biolyon for phospholipids). Protein concentrations were determined as described by Lowry [20] using serum albumin as the standard. Determination
of particle size
HDL particle size was determined by nondenaturing polyacrylamide gradient gel electrophoresis (Pharmacia PAA 4-30) according to the method of Nichols et al. [21]. After staining and destaining, the gels were scanned in a laser densitometer (LKB, Broma, Sweden). Thyroglobulin (Sigma) (1.75 mg/ml) with molecular diameter of 17.0 nm, was used as an internal standard. The size of HDL subfractions was calculated from a calibration curve using standards of known diameters. The standards included thyroglobulin (diameter 17.0 nm), apoferritin (diameter 12.2 nm), lactate dehydrogenase (diameter 8.1 mn) and bovine serum albumin (diameter 7.1 run). As described by Nichols et al. [21] and Blanche et al. [26], the separation of control HDL particles according to the size identified five subpopulations. The subpopulations, HDL,, and HDL,,, were located
TABLE I Plasma lipid and apolipoprotein levels and the distribution of cholesterol in lipoprotein fractions in the study groups The values are the means f S.D. Number of subjects
Plasma concentrations (mg/dl) cholesterol
TG
Ape A-I
Control ’
14
194*
36
70*21
164k32
99*17
Hypercholesterolemic b
15
363 f 115
76*28
134*45
168*61
i Controls, healthy normofipidemic subjects. Hypercholesterolemic, type IIa familial hypercholesterolemic patients.
AQO B
VLDL-C
LDL-C
9.6*5.4
135*
7.8 f 6.1
302k118
HDL-C 30
49f15
56521
66 within the ultracentrifuge HDL, subclass. Three subpopulations: HDL,,, HDL,,, HDL,, were located within the ultracentrifuge HDL, subclass.
sensitivity in the anisotropy measurements by making differential measurements either at 24 or at 37 o C. A fresh dispersion of fluorescent probe (DPH or TMA-DPH) was incubated in 3 ml of PBS 0.1 M (pH 7.4), with 100 to 500 ~1 of lipoprotein (VLDL, LDL, HDL). The probe concentration was chosen in respect to a final probe/phospholipid ratio always less than l/1000. The probe was incubated with lipoprotein for 1 h at 24°C with gentle agitation. After labelling, the fluorescence measurements were carried out on an Aminco SPF 500 spectrofluorimeter equipped with a fluorescence polarization accessory and a controlled temperature cell holder. The temperature was maintained at 0.5 o C and the excitation and emission wavelength settled to 360 and 460 nm, respectively.
Fluorescence polarization studies Lipoprotein fluidity was assessed by determining the steady-state fluorescence anisotropy rs from two probes: DPH (1,6-diphenyl-1,3,5-hexatriene) and TMA-DPH (1,4-trimethylammonium phenyl-6-1,3,5-hexatriene). (For a review of the theory and the technique, see Refs. 22, 23 and 24). A structural order parameter, S, for the lipid phase was calculated with the equation defined by Heyn [25]. Arrhenius plots (log r versus l/T) were used to calculate A E, the flow activation energy for each lipoprotein particle in the temperature range 1545°C. This thermodynamic parameter can be considered to be an index of physiological properties of some biological systems, such as membranes and lipoproteins. In order to relate the physical data to the biochemical composition of lipoproteins, it was necessary to locate as precisely as possible the fluorescent probes after controlled passive incorporation in the various particles. Trinitrobenzene sulfonic acid (T’NBS) was used as a distance control quencher (4.0 nm). TNBS binds to the surface lipoproteins and can therefore be used to measure the average depth of the probes. Considering the different localizations of the two probes and the difference in the chemical composition of the different parts of the particles, it was possible to enhance the TABLE
Statistical analyses All values are reported as mean rt_S.D. Data were analysed for statistical significance by Student’s t-test. Graphical data were subjected to linear regression analyses to determine the line of best fit. Results
Lipoprotein chemical composition No significant difference between the two groups was observed for the biochemical composition of VLDL and HDL (Table II). However, the CE/TC ratio was signifi-
II
Chemical composition of lipoproteins from variow groups All values are expressed
as mean f S.D. Lipid composition of lipoproteins (mg/lOO mg of proteins)
VLDL control hypercholesterolemic
LDL control hyperholesterolemic
HDL control hypercholesterolemic
* P < 0.05 vs. control. ** P i 0.01 vs. control. *** P i 0.001 VS. control.
CE/TC
CE/PL
CE/TG
TG/PL
total cholesterol
cholesteryl esters
triacylglycerol
phospholipids
(TC)
(CE)
(TG)
(PL)
99 It59 76 f34
52 *31 40 f20
356 f 248 305 *134
135 f73 102 *45
0.507 f 0.062 0.518 0.078
0.376 f 0.098 0.441 f 0.286
0.149 f 0.043 0.15 *0.13
2.53 f 0.41 3.29 f 1.68
135 rt25 153 f32
96 *15 117 k23
22.7 f4.6 17.4 * * f 4.2
90.7 * 9.0 97 k16
0.734 f 0.014 0.726 + 0.014
1.066 f 0.085 1.139 * f 0.078
4.44 f 0.99 6.57 * * * *1.44
0.253 f 0.047 0.181* * * f 0.039
5.6 i-2.8 4.7 f 1.8
48 *21 45 *17
0.828 f 0.032 0.808 * f 0.017
0.502 f 0.084 0.560 f 0.081
4.49 izl.33 5.41 f 1.53
0.118 k 0.025 0.114 f 0.044
29.2 lt15.3 30.4 f 10.3
20.6 k2.3 24.6 f 8.3
67
(2b)
(2b)
(Za) (3aX3bXJc) HDL Subpopulat ions
(Za)(3aX3bX3c) HDL Subpopulations
Fig. 1. Densitometric scans of the 4-30% polyacrylamide gradient gel electrophoresis of total HDL from two hypercholesterolemic patients (HCl and HC2) and two control subjects (Cl and C2). Scale at bottom corresponds to size intervals of five major HDL subpopulations identified in polyacrylamide gradient gel electrophoresis.
cantly reduced in the hypercholesterolemic HDL patients. The more significant changes occurred in LDL. The TG content of hypercholesterolemic LDL patients was reduced when compared to normal subjects (P < O.Ol), while CE and PL were slightly increased (N.S.). The CE/TG ratio was therefore significantly increased (P < O.OOl),while the TG/PL ratio was decreased (P < 0.001). HDL particle size
Densitometric
scans of the 4/30% polyacrylamide
gradient gel electrophoresis of the HDL from two hypercholesterolemic patients and two control subjects are shown in Fig. 1. These scans are representative of those obtained for the other studied HDL in each group. We observed in HDL from hypercholesterolemic patients the lack or the reduction of the larger-size subclasses (HDL,, or HDL,,) while the smaller size subspecies (HDL,, or HDL,,) were increased. Moreover, in HDL from HC patients, the mean particle size of two subfractions, HDL,,, and HDL,,, appeared to be significantly smaller than those of the normal subjects (Table III).
TABLE III High-density
lipoprotein particle size in I1 hypercholesterolemic
patients and 11 control subjects
High-density lipoprotein (HDL) particle sire (nm) was determined by 4-30s polyacrylamide gradient gel electrophoresis. HDL subfactions were identified as described by Nichols et al. [21] for densitometry scans. See Fig. 1. The values are the meansf S.D. HDL subfractions 2b
2a
3a
Hypercholesterolemic patients
10.7*0.1
9.6 f 0.2
8.6f0.2
Controls
10.9kO.2
9.7kO.2
8.8 *0.2
* P < 0.05 vs. control. ** P -z 0.01 vs. control.
*
3b
3c
8.OkO.2 * *
7.6 f 0.3
8.3 f 0.1
7.6 f 0.1
68 TABLE
IV
Probe location in LDL and HDL All results are expressed as the mean value obtained in lipoproteins from five controls or five hypercholesterolemic patients. (Percentage of the fluorescence probe quenched by trinitrobenzene sulfonate : TNBS). % of fluorescence quenched by TNBS DPH LDL HDL
12.3f3.5 r53.1*1.4
Th4A DPH LDL HDL
3.3 f 2.1 12.3 f 2.1
Lipoprotein jluorescence polarization studies
In LDL, almost all the fluorescence of DPH and TMA-DPH was quenched by TNBS. Table IV reports that only 12% and 3% of the initial fluorescence intensity remained for DPH and TMA-DPH, respectively. In HDL, the results were different. The residual fluorescence intensity after TNBS was 12% for TMA-DPH and 64% for DPH. These observations suggest that both DPH and TMA-DPH are located in the outer shell of LDL, while in HDL, TMA-DPH reflected phenomena of the outlayer and DPH, more deeply embedded, reflected the lipid dynamics of the core. As briefly explained in Material and Methods, it is possible to enhance the sensitivity of the fluorescence polarization measurements. Because of the localization of the probes, DPH reflects a very fluid part of the lipoproteins, especially in HDL, so it can be very practical to measure fuidity at 24” C, in order to enhance the perception of subtle changes which can be hindered at 37 OC because of the higher lipid dynamics. In contrast, TMADPH is located in high constraint zones of the particles (the surface of the outlayer) and it may be more judicious to work at 37’C, in order to enhance the lipid TABLE
Composition-structure correlations in lipoproteins
Table VI reports the different correlations observed between DPH fluorescence anisotropy and the chemical composition in lipoproteins from the both groups. In VLDL from both groups a negative correlation was observed between fluorescence anisotropy r and the TG/protein ratio (-0.5179 for controls and - 0.5105 for hypercholesterolemic patients, P < 0.05). A more significant correlation existed with the TG/PL ratio (- 0.6020 for controls, P < 0.05 and - 0.6484 for HC, P < 0.02). No significant correlation was observed with the CE/TG ratio. In LDL from normal subjects, DPH fluorescence anisotropy was negatively correlated with TG/protein (-0.7834, P < 0.01) and TG/PL (- 7357, P < 0.01)
V
Steady-state fluorescence anisotropy rs of DPH and TMA-DPH
DPH 24°C Fluorescence anisotropy
r,
Structure order parameter S TMA-DPH 37 o C Fhiorescence anisotronv r.
and structure order parameter
of lipid phase S in lipoproteins from various groups HDL
LDL
VLDL
The * ** l **
dynamic of these parts and to allow more perceptible TMA-DPH movements and their influence on the polarization measurements. In the LDL and HDL from hypercholesterolemic patients, the mean DPH fluorescence anisotropy r was higher than in the control group (P < 0.05 and P < 0.01, respectively) (Table V). The order parameter S was also significantly enhanced in both LDL and HDL (P < 0.01). No particular changes in VLDL were reported by DPH or TMA-DPH in the two study groups. No difference was noted in the TMA-DPH fluorescence anisotropy in LDL from either groups. However, in HDL, TMA-DPH reported higher values of fluorescence anisotropy in hypercholesterolemic patients than in controls (P < 0.001). The study of the thermodynamic behavior of LDL and HDL from the two groups with DPH between 15’ and 45 ’ C showed that all systems were invariant phases (Fig. 2). The flow activation energy A E was calculated from Arrhenius plots. The mean A E values were higher in LDL and HDL from hypercholesterolemic patients than in control particles (P < 0.01 and P < 0.001, respectively).
control
hyperCT
control
hyperCT
0.033
0.259 f 0.010
0.272 f 0.015 *
0.229 f 0.006
0.244 f 0.014 * *
0.449 * 0.079
0.501 kO.106
0.825 f 0.024
0.849 f 0.035 * *
0.752 k 0.016
0.786 f 0.033 * *
0.212 f 0.068
0.235 f 0.067
0.311 f 0.014
0.317 * 0.010
0.257 f 0.013
0.283 & 0.018 * * *
control
hyperCT
0.131~0.018
0.142*
values are expressed as the mean f S.D. P < 0.05 vs. control. P < 0.01 vs. control. P -z 0.001 vs. control.
69
A
9876-
AE (kcal LDL control LDL HC
/mole) 6.94 9.03
f 1.0/i + 1.36a
20.69
HDL control
6.52
HOL HC
9.30 + 0.75b
2-
L
I
I
I
I
3.2
3.3
3.L
3.5
l/T
)
I
I
I
I
3.2
3.3
3.4
x lo3
l/T
b
x103
Fig. 2. Arrhenius plots of steady-state microviscosity of DPH and flow activation energy AE in LDL (left panel) and HDL (right panel)from ) and hypercholesterolemic (- - - - - -) groups. Flow activation energy of DPH was determined between 15 and 45 o C. Statistical control (test: (a) P < 0.01 vs. control group; (b) P < 0.001 vs. control group.
ratios and positively correlated with CE/TG ratio (0.6976, P -z0.01). In hypercholesterolemic LDL, there was no correlation with the TG/protein and CE/TG ratios (Fig. 3). Only a slightly negative correlation was noted with the TG/PL ratio (-0.5376, P < 0.05). Very similar results were observed in HDL. In the HDL from the control group, DPH fluorescence anisotropy was negatively correlated with the TG/protein ratio (-0.6199, P < 0.02) and to a less extent with the TG/PL ratio (-0.5605, P < 0.05). A good positive correlation was noted with the CE/TG ratio (0.6649,
P -c0.01). None of these correlations was observed in HDL from hypercholesterolemic
patients (Fig. 4).
Discussion In the present work, we examined the physical state of lipoproteins (VLDL, LDL and HDL) from hypercholesterolemic or control subjects related to their chemical composition or their size. The physical lipoprotein properties were studied by fluorescence depolarisation of two fluorescent probes DPH and Th4A-
TABLE VI Correlations between DPH fluorescence anisotropy and lipid composition in lipoproteins isolated from various groups Control group r VLDL Triacylglycerol/proteins Triacylglycerol/phospholipids Cholesteryl esters/triacylglycerol
-0.5179 - 0.6020 0.2483
LDL Triacylglycerol/proteins Triacylglycerol/phospholipids Cholesterol ester/triacylglycerol
-0.7834 -0.7357 0.6976
HDL Triacylglycerol/proteins Triacylglycerol/phospholipids Cholesteryl esters/triacylglycerol
- 0.6199 - 0.5605 0.6649
P: Correlation coefficient. * PcO.05; ** P < 0.02; *** P < 0.01.
Hypercholesterolemic group P * * n.s. *** *** ***
** * ***
r
P
- 0.5105 - 0.6484 0.3447
* ** n.s.
- 0.2962 - 0.5376 0.4616
n.s.
- 0.0064 - 0.2389 0.2332
n.s. n.s. n.s.
n.s. *
70 LDL
Control
i
A
B 0
0.3
a
1
0
0.2 c
0.3 -
og
0
\
HC LDL
u
a -.
0 00
0.2 -
w
I-
l
s
8.
c
0
n 0.11 220
I
I 240
I
260 I
9
a
0
\
280
8
l l
1
240
I
280
260
x 1000
a
l l
l
I
O.I220
300
300
r x 1000
c
D
9-
7
d
,”
5-
z
3-
\
3i
1
I
I
I
1
220
240
260
280
300
c
l
l
l
I
.i
.
m
l
8 8
m
l
m
l
m
11 220
l
:
8
8
I
I
I
1
240
260
280
300
r x 1000
r x 1000
Fig. 3. Plots of steady-state DPH fluorescence anisotropy rS vs. triacylglycerol to protein (A, B) and cholesterol ester to triacylglycerol (C, D) ratios in LDL from control (left panel) and hypercholesterolemic (right panel) groups. Each point represents values for one subject. The lines of best fit are shown for control LDL. Correlation coefficients are given in Table VI. Control
HC
HDL
HDL
9
#
B
00
.
7
5
