119

RBA 54048

Differential effect of subspecies of lipoprotein containing apolipoprotein A-I on cholesterol efflux from cholesterol-loaded ma~rophag~s: functional correlation with lecithin : cholesterol a~yltra~sferase Takao Ohta a, Rie Nakamura ‘, Y&him Ikeda ‘, M~r~~s~~u ~~i~~~ar~ ht Akira Miyazaki b, Seikoh Horiuchi b and Ichiro Matsuda ’ Departments

of ’ Pediatrics

and ’ Biochemistry, Kumamoto Uniwsity School

of Medicine,

Kumamotn (Japan)

(Received 10 April 1992) (Revised manuscript received 3 July 1992)

Two species of lipoprotein col?tai~~n~ apoA-I, one co~tai~~~~ only apoA-I KpA-I), and the other containing apoA-I and apoA-II (LpA-I/A-II), were tested for their effects on macrophage foam cells. Rat macrophages were converted to foam ceIIs by incubation with radiolabeled acetylated LDL. Incubation with LpA-I or LpA-I/A-II decreased the cellular cholesteryl esters (CE) mass. However, the free cholesterol (FC) mass was only reduced by LpA-I. All the radioactivity excreted into the medium was associated with LpA-I or LpAmI/A-II; 39% of the excreted radioactivity was esterified in LpA-I and 10% in LpA-I/A-II. Upon complete inactivation of lecithin : cholesterol acyltransferase &CAT) activity with dithiobisnitrobenzoic acid, the cholesterol reducing capacity of LpA-I was weakened significantly. However, the CE mass reducing capacity of LpA-I/A-II was not affected. When LpA-I and LpA-I/A-II were combined, the cholesterol reducing capacity of the mixture was similar to that of l&A-I alone. However, LpA-I re-isolated from the medium showed a lower esterification rate than did the se-isolated LpA-I/A-II, thereby indicating that the cholesterol esterified in LpA-I was transferred to LpA-I/A-II. These results suggest that 6) the function of LpA-I is closely linked to the LCAT activity white that of LEA-IDA-II is not, and (iif LpA-I in concert with LpA-f/A-II induces a series of e~traceIIular events; L~AT-mediated ester~ficat~o~ of excreted FC by LpA-I and a subsequent CE transfer to LpA-f/A-II. These mechanisms might be important for net cholesterol efflux from macrophage foam cells in ~hys~o~~~jcal states,

Introduction Lipoproteins containing apoA-I isolated from plasma by immunological methods are divided into two subspecies: one contains apoA-I but no apoA-II (LpA-I) and the other is lipoprotein containing apoA-I and apoA-II ~L~A-I~A-I~~. These Ii~opro~eins may reflect a more native state of HDL particles than does the MIX isolated by ultracentrifugation f l-31. Recent clinical studies have shown that LpA-I levels in patients with atherosclerotic coronary heart disease (ACHD) are preferentially lower than corresponding normal subjects [4,.5], and that changes in LpA-I levels occur

more prominently than those in LpA-I/A-II levels in patients with chronic renal failure who are at an increased risk for ACHD 161, LpA-I levels are lower in males than in females who are generally resistant to ACHD, whereas no sex difference was found in plasma LpA-I/A-II levels f’7,81.These lines of clinical evidence suggest that LpA-I might represent the anti-atherogenie fraction of HDL. One of major physiolog~~l functions of HDL is attributed to ‘reverse cholesterol transport’, i.e.: the transport of cholesterol from peripheral cells to the liver for excretion into the bile [9-111. Coupled with this nation, the anti-atherogenic effect of HDL at clinical levels is explained by acceleration of the cholestero1 efflux from cholesteryl esters laden foam cells 1121.Studies provided evidence that macrophages or macrophage-derived cells are precursors of foam cells, particularly in an early stage of atherosclerotic lesions [13,14]. Therefore, we examined the effect of

120 LpA-I and LpA-I/A-II on cholesterol efflux from macrophage foam cells in an attempt to elucidate their capacity for net cholesterol efflux from foam cells. Changes in cellular levels of free cholesterol (FC) and cholesteryl esters (CE) were determined in a tracer study using [3H]cholesterol and by mass determinations. In addition, we examined the fate of cholesterol after excretion from macrophage foam cells. The current use of LpA-I and LpA-I/A-II revealed a series of extracellular events during cholesterol efflux from macrophages; lecithin : cholesterol acyltransferase (LCAT)-mediated esterification of excreted FC by LpA-I and a subsequent CE transfer to LpA-I/A-II. Materials

and Methods

Materials All chemicals used were the best grade available from commercial sources. [7-“HIcholesterol (21.9 Ci/ mmol) was purchased from Du Pont New England Nuclear. Plasma for isolation of lipoproteins containing apoA-I, HDL and LDL were obtained from five independent groups of healthy adult normolipidemic volunteers from Kumamoto University Medical School staff and students (each group included 5 men and 5 women, 18-29 years of age, all who had fasted overnight). Five independent pooled plasma samples were used for the following studies. Isolation of LpA-I, LpA-I/A-II, HDL,, HDL, and LDL from plasma LpA-I and LpA-I/A-II were isolated from each fresh pooled plasma sample by a combination of antiapoA-I and anti-apoA-II immunosorbent columns, as described [3,7]. Briefly, fresh plasma was applied on an anti-apoA-I column. After washing extensively with 0.01 mol/L Tris, 0.5 mol/L NaCI, 1 mmol/L EDTA, pH 7.5 (buffer A), the column was eluted with 0.1 mol/L acetic acid, 1 mmol/L EDTA, (pH 3.0). Each effluent was immediately adjusted to pH 7.4 with 1.0 mol/L Tris solution and dialyzed against 0.15 mol/L NaCl and 1 mmol/L EDTA, pH 7.4 (buffer B). Finally, the sample was concentrated and then applied on an anti-apoA-II column. The column was washed with buffer A to obtain LpA-I. The bound fraction was eluted from the column to obtain LpA-I/A-II. Both LpA-I and LpA-I/A-II were dialyzed and concentrated in buffer B. In this procedure, > 90% of lipids and apolipoproteins applied were recovered in the unbound and bound fractions. LDL (d = 1.019-1.063 g/ml), HDL, (d = 1.063-1.125 g/ml), HDL, (d = 1.125-1.21 g/ml) and lipoprotein deficient fraction (LDF) (d > 1.21 g/ml> were isolated from a portion of the fresh pooled plasma sample by sequential ultracentrifugation at 150000 x g for 22 h at 4°C 1151 and dialyzed against buffer B.

[“HIFC-labeled ace@-LDL Acetyl-LDL was prepared as described [ 161. [ ‘H]FC was incorporated into acetyl-LDL by a modification of the method of Jonas et al. [171. Two ml of chloroform was added to 100 mg of washed and dried Celite particles in a 10 ml conical glass tube, and then 25 PCi of [“HIFC per mg protein of acetyl-LDL was added. The tube was dried by nitrogen flush and resuspended in 3.0 ml of Tris-HCI buffer (pH 8.0) containing 4-5 mg of acetyl-LDL and 0.02% NaN,, After 15-h incubation at 37°C with gentle rotatory shaking, the tube was centrifuged and the supernatant was filtered, followed by dialysis against buffer B. More than 98% of the radioactivity was recovered from the band corresponding to FC seen in the TLC analysis. Cells All cell culture experiments were done at 37°C in humidified air containing 5% CO, [18]. Peritoneal macrophages were harvested in PBS from nonstimulated male Wistar rats (170-200 g), centrifuged at 800 x g for 3 min, and suspended at 2-3.10’ cells/ml in Dulbecco’s modified minimal essential medium containing 3% BSA, 100 U/ml penicillin and 100 pg/ml streptomycin (medium A). To each 22-mm plastic dish was added 1.0 ml of the cell suspension and the preparation was incubated for 4 h. Cell monolayers formed were washed 3 times with 1 ml of PBS and used for cholesterol efflux experiments. Cholesterol efflux from macrophage foam cells Adhered macrophages were converted to foam cells by a 16-h incubation with 100 pg/ml of [3H]FC-labeled acetyl-LDL (5000-8000 cpm/pg protein) in 1.0 ml of medium A. Each well was then washed twice with 1.0 ml of PBS containing 0.2% BSA and twice with PBS, followed by a 6-h incubation in 1.0 ml of medium A for equilibration. Preliminary experiments showed that a 6-h equilibration was sufficient to allow surface bound and undegraded acetyl-LDL to undergo complete processing. These foam cells were washed twice with PBS containing 0.2% BSA and twice with PBS and subjected to efflux assays by incubating with medium A containing LpA-I, LpA-I/A-II and the combination of LpA-I and LpA-I/A-II or other effecters to be tested. Parallel incubations were performed without lipoproteins, as the controls. In studies using dithiobisnitrobenzoic acid (DTNB), LpA-I and LpA-I/A-II were preincubated with 4 mmol DTNB for 4 h, and dialyzed against PBS for 16 h to remove unreacted DTNB. At 24 h after the onset of efflux experiments, the culture medium was removed to isolate the lipoproteins; the wells were then washed twice with 1.0 ml of PBS containing 0.2% BSA, 4 times with PBS, and the cellular lipids were then extracted. Unless otherwise specified, the data derived from these efflux assays

121 were the mean of quintuplicate experiments. Re-isolation of LpA-I, LpA-I/A-II, LDL from culture medium

runs in five separate

HDL,,

HIILj

and

LpA-I and LpA-I/A-II were re-isolated from culture medium by anti-apoA-I and anti apoA-II immunosorbent column in a manner similar to those from plasma as described above. HDL, and HDL, were re-isolated from culture medium on an anti-apoA-I immunosorbe~t column as apoA-I containing lipoprotein. LDL was re-isolated from culture medium on an anti-apoB immunosorbent column. This column was prepared by the same methods used for the anti-apoA-I and anti-apoA-II immunosorbent columns. Briefly, antisera to apoB were obtained from rabbits, as described [191. Antibodies specific for apoB were isolated from antiserum on an LDL-CNBr-activated Sepharose 4B column and coupled to Formyl-Cellufine, a formylated derivative of cellulose gel, as described [3,7]. Lipid extraction from macrophage monolayer and lipoproteins isolated from capture medium

FC and CE were extracted directIy from ma~rophage monolayers by the method of Hara and Radin [20], an approach used by Brown et al. 1211. Briefly, to each well was added 0.6 ml of hexane/isopropanol (3: 2) and the cells were incubated for 30 min at room temperature. The organic solvent was saved and the cells were extracted again with 0.4 ml of the same solvent. The combined extracts were dried under nitrogen and dissolved in 75 yl of isopropanol. Aliquots were used both for radioactivity and the mass estimations described below. After lipid extraction, cells in each well were washed with 1.0 ml PBS and dissolved in 0.5 ml of 0.1 N NaOH for 10 min at 3’7°C. The supernatant was saved and the wells were again treated with 0.4 ml of 0.1 N NaOH. The combined supernatant was used to determine cell protein concentrations. Lipoproteins isolated from culture medium were dried under nitrogen before lipid extraction. Lipid was extracted by the same procedure mentioned above. Radioactivity determination

Aliquots (20 $1 of lipid extracts were spotted in duplicate on a thin layer chromatography (TLC) plate (Merck) and developed in n-hexane/ diethyl ether/ acetic a~id/methanol (8.5 : 20 : 1: 1, v/v). Spots corresponding to FC and CE were cut out from the plate and the radioactivities were determined. Mass determination

The mass of FC and CE was quantified by a modification of the enzymatic/fluorometric method of Heider and Boyett [22]. Briefly, a remaining portion (10 ~1) of the lipid extract was added to 0.4 ml of enzyme

mixture and the preparation was incubated at 37°C for 1 h (for FC) or for 2 h (for total cholesterol), followed by the addition of 0.81 ml of 0.5 N NaOH to halt the reaction. Enzyme mixtures were similar to those used by Heider and Boyett except that Carbowax-6000 was replaced with 0.01% Triton X-100 and the enzyme concentrations used were twice as high (cholesterol oxidase; 0.16 U/ml and cholesteryl hydrolase; 60 U/ml>. Fluorescence intensity was measured with excitation at 320 nm and emission at 407 nm. Values were quantitated by comparison with standard curves obtained using cholesterol and choleste~l oleate for FC and total cholestero1, respectively. CE was calculated by subtracting FC from the total cholesterol. Standard curves were constructed for each set of experiments. LCAT assay

The LCAT activities in LpA-I and LpA-I/A-II were measured using endogenous substrates. [ ‘H]FC was incorporated onto polystyrene tissue culture wells (Corning) as follows: absolute ethanol (100 ~1) containing 0.2 ,uCi of i3H]FC was placed in wells and the ethanol was dried off by nitrogen flush. Then, each Iipoprotein (400 gg of protein) in 500 ,uI of phosphate buffer saline (PBS) was added to the well and the 13H]FC was equilibrated with FC in each lipoprotein by incubation at 4°C for 16 h. Total radioactivity of LpA-I and LpA-I/A-II was similar (62000 + 6800 cpm; mean f S.D., y1= 3) and specific radioactivity (cpm/ nmol FC) was 680 t_ 100 in LpA-I and 1050 + 230 in LpA-I/A-II (mean + S.D., n = 3). [“HIFC-labeled lipoproteins showed a pattern identical to that of the unlabeied preparations, using agarose gel and sodium dodecylsulfate polya~~iamide gel electrophoresis. ~~H~FC-1abeIed lipoproteins were then incubated at 37°C. Aliquots of each lipoprotein sampte were taken out at 0, 0.5, 1, 3, 5, 7, 9, 12 and 24 h to observe the time course of cholesterol esterification. The enzyme reaction was stopped by immersing the sample tubes in an ice bath. 13H]FC and [3H]CE were separated by TLC as described above. LCAT activity was expressed as the difference between the percentage of radioactive choIestero1 esterified before and after incubation at 37°C. Assay uf c~~~este~~ esters tra~fer

The transfer of CE from LpA-I to LpA-I/A-II was studied by isotopic CE transfer assay. [“HIFC was incorporated onto tissue culture wells as described above. Then, LpA-I (400 pg of protein) in 500 ~1 of PBS was added to the well and incubated at 37°C for 16 h. With this procedure, 33-40% of total radioactivity in LpA-I was present on the spot corresponding to CE in the TLC analysis. The radiolabeled LpA-I was incubated with the same amount of unlabeled LpAI/A-II at 37°C for 24 h in the presence of 4 mmol

122

DTNB to inhibit the LCAT activity. It has been previously shown that DTNB does not affect the CE transfer between lipoproteins [23]. LpA-I and LpA-I/A-II were then re-isolated from incubation medium by immunoaffinity chromatography. [ “H]FC and [ “H]CE in LpA-I and LpA-I/A-II were separated by TLC as described above. The percent of labeled CE transferred to LpA-I/A-II was calculated. Protein and lipid analysis The apoA-I, apoA-II and apoE concentrations of plasma, LpA-I and LpA-I/A-II were measured by radial immunodiffusion assay [ 19,241. Cholesterol and triglyceride concentrations of these samples were analyzed on an ABA 100 Autoanalyzer (Abott Laboratory), using enzymatic methods [25,26]. CE was measured by enzymatic methods [27]. Phospholipid was analyzed by the method of Bartlett [28]. The protein content of each fraction from the immunosorbent columns was determined by the method of Lowry et al. [29]. Values of lipids and apolipoproteins in LpA-I and LpA-I/A-II were corrected based on the % recoveries during the isolation. Electrophoretic analysis Agarose gel electrophoresis was performed using a Pol-E Film system for lipoprotein (Corning). Slab gel electrophoresis was done according to Weber and Osborn [30]. The Stokes diameters of the lipoprotein particles were estimated by gradient polyactylamide gel electrophoresis on Pharmacia precast PAA 4/30 gels, according to the procedure specified by the manufacturer. Electrophoresis calibration kits for high molecular weight proteins (Pharmacia) were used. Statistical evaluation The Wilcoxon signed rank test were used to evaluate the data.

and

paired

TABLE

I

Chemical compositions of LpA-I and LpA-I/ A-II LpA-I Protein Total cholesterol Phospholipid Triglyceride CE/TC x 100 FC/PL ratio

53.1 17.1 25.3 4.5 71.5 0.20

s + 1.1 +0.4 kO.9 +0.4 kO.6 * 0.3

LpA-I/A-II 58.7 + 1.0 14.4 kO.5 23.9 kO.8 3.4 kO.4 75.9 kO.5 0.13kO.2

:

h b

Values are expressed as meankS.E. of % weight or %(CE/TC). S, significance between LpA-I and LpA-I/A-II; TC: total cholesterol; CE: cholesteryl ester; a, P < 0.01; b. P < 0.005.

I/A-II particles and the remainder was on LpA-I. With gradient gel electrophoresis, LpA-I migrated as two distinct particles and LpA-I/A-II as three particles (Fig. 1). Effect of LpA-I and LpA-I/A-II on cholesterol efflux from macrophage foam cells When foam cells were subjected to a 24-h incubation with an increasing concentration of LpA-I or LpA-I/A-II, 50-60% of the [3H]CE and 60-70% of [ “H]FC were similarly reduced significantly (P < 0.005) (Fig. 2). A net cholesterol flux is defined as a combination of an influx from effector lipoproteins into foam cells and an efflux from foam cells. To determine the net cholesterol flux, changes in the mass level of cellular CE and FC were measured and findings were compared with data obtained by the tracer study. As Fig. 3 shows, the CE mass was reduced by incubation

t-test

Results

Characterization of LpA-I and LpA-I/A-II Table, I summarizes the chemical composition of LpA-I and LpA-I/A-II isolated from five independent pooled plasma. The percent protein was lower in LpA-I than in LpA-I/A-II, whereas the % total cholesterol of LpA-I was higher than that of LpA-I/A-II. The ratio of CE to total cholesterol of LpA-I was slightly lower than LpA-I/A-II, thereby indicating that the LpA-I was rich in FC. The molar ratio of apoA-I to apoA-II in LpA-I/A-II was 1.57 + 0.12 (mean + S.D.). FC/ Phospholipid(PL) ratio (pgg/pg) was significantly higher in LpA-I than in LpA-I/A-II. ApoE content in whole unfractionated lipoproteins containing apoA-I was 40 _t 5% (mean + S.D.) of the total plasma apoE; 68 + 4% (mean f S.D.) of which was localized on LpA-

11.1

nm-

10.1

nm

d-

9.0

nm

c-

8.1

nm

8.8

nm -

AB Fig. 1. Nondenaturing gradient polyacrylamide gel electrophoresis of LpA-I (A) and LpA-I/A-II (B). Stokes diameter of each particle is inserted.

123 CE

FC

100

100

80

80

60

60

40

40

20

20 0 100 200 300

Effector Concentration ( pgfml 1 Fig. 2. Effect of LpA-I and LpA-I/A-II on [‘HJCE and FC levels in macrophage foam cells. Macrophage foam cells were incubated for 24 h with varying concentrations of LpA-I and LpA-I/A-II. Control incubations were performed without LpA-I and LpA-I/A-II (0 pg/ml). Cellular lipids were extracted, subjected to TLC and the radioactivity of CE and FC was determined. Data are the mean f S.E. of five experiments. Mean values of CE and FC for 100% were 35’ lo4 cpm and 53 X lo4 cpm/mg cell protein, respectively.

either with LpA-I or LpA-I/A-II. These patterns of CE-reduction were similar to those observed with [ ‘H]CE (Fig. 2). The FC mass was significantly reduced by LpA-I at 200 and 400 pg/ml (P < 0.01). However, the extent of FC mass reduction by LpA-I (30-400/o) was much smaller than that seen in the tracer study (60-70%). In contrast, incubation with LpA-I/A-II had no apparent effect on the FC mass. Thus, an influx of FC from these lipoproteins into foam cells might occur concurrent with the efffux. Since our LpA-I and LpA-I/A-II contain apoE as a protein component [6,7J, it is possible that the FC influx from these lipoproteins could be due to endocytic uptake of these CE

FC 100

so 60 40

20 ~

-0i 100

200

No

a IOU 200 MO -.- @IA-I L&-"&-n

400

Effector Concentration

q HIM.?

J&I

LpA-I

4&I

( @g/ml )

Fig. 3. Effect of LpA-I and LpA-I/A-II on CE and FC mass in macrophage foam cells. Macrophage foam cells were incubated for 24 h with varying concentrations of LpA-I and LpA-I/A-II. Control experiments were performed without LpA-I and LpA-I/A-II (0 yg/ml). Cellular lipids were extracted and determined for FC and CE. Data are the mean&SE. of five experiments. Mean values of CE and FC for 100% were 113 and 122 nmol/ mg cell protein, respectively.

q H1X.l

Fig. 4. Effect of HDL, and HDL, on CE and FC mass in macrophage foam celk. Macrophage foam cells were incubated for 24 h with 4oU pg/mi of HDL, and HDL,. Control experiments were performed without HDL, and HDL,. Cellular lipids were extracted and determined for FC and CE. Data are the mean + SE. of five experiments. Mean values of CE and FC for 100% were 114 and 165 nmol/mg cell protein, respectively.

lipoproteins by the apoB/E receptor of macrophages. After a 24-h preincubation, macrophages were incubated for another 24-h with 400 pg/ml of LpA-I or LpA-I/A-II or [3H]FC-labeled LpA-I or 13H]FClabeled LpA-I/A-II. Both cellular FC and CE mass levels remained unchanged. A parallel radioactive analysis showed that although about 5% of the radioactivity of these lipoproteins was transferred to cells, none of which was converted to CE. Thus, it is likely that FC influx mediated by apoB/E receptor is negligible. Effect of HDL,, from foam cells

HDL3 and LDL on cholesterol efflux

Incubation with 400 pug/ml of HDL, or HDL, led to a reduction in (“H]CE (5-15% and 40-50%, respectively), and in [3H]FC (60-70%, similar in both). The difference of the CE-reduction by HDL, and HDL, was significant (P < 0.005). As shown in Fig. 4, the CE mass was reduced by HDL, (30-40%). HDL, did not reduce significantly the CE mass. These lipoproteins had virtually no effect on the FC mass level. We also tested the effect of LDL. Incubation with 400 ,ug/ml of LDL resulted in a significant reduction (60-700/o) of [3H]FC, whereas the [ 3H]CE Ievel did not change. LDL either had no effect or slightly increased both CE and FC mass levels. Fate of ~e~~~larradioactive ~~o~estero~ after excretion into the culture medium

Amounts of the radioactivity released into the culture medium were at much the same level among the lipoproteins tested (data not shown). To determine to what extent the released radioactivity associated with added lipoproteins, lipoproteins were re-isolated from the culture medium. More than 98% of the radioactivity was recovered in the re-isoIated lipoproteins. To

124 50,

0

3

6

9 1215182124

Hours Fig. 5. Time-course of LCAT activity in LpA-I and LpA-I/A-II. Values are mean k SE., n = 3. Error bars of several points in LpAI/A-II are buried in the symbols.

assess the subsequent fate, we determined whether or not the released radioactive cholesterol could be esterified. The radioactivity in the medium was separated into FC and CE to determine their ratio. When incubated with LpA-I, 39 f 3% of the total excreted radioactivity was found to be esterified, whereas 10 * 3% of the total radioactive cholesterol was esterified when chased with LpA-I/A-II (mean f SD.). In parallel experiments, the percent esterification of the excreted radioactivity was less than 1% for HDL,, HDL, and LDL. LCAT activity of LpA-I and LpA-I /A-II As shown in Fig. 5, LCAT activity of LpA-I was 4 times higher than that of LpA-I/A-II. However, when we take into account the fact that total radioactivity of these lipoproteins was similar and specific radioactivity (cpm/nmole FC) in LpA-I/A-II was 1.5 times higher than that in LpA-I before incubation, the CE mass formed by the action of LCAT (real activity) was approximately 6 times greater in LpA-I than in LpAI/A-II. In similar experiments using HDL,, HDL, and LDL, no esterification was found (data not shown). To confirm the effect of ultracentrifugation on LCAT activity, we did following experiments. i3HlFC was equilibrated with FC in HDL, as described in Material and Methods. Then, nonbinding plasma fraction of anti apoA-I immunosorbent column (plasma containing no apoA-I) (100 ~1) and lipoprotein deficient plasma fraction obtained by ultracentrifugation (LDF) (100 ~1) were incubated with [“HIFC labeled HDL, for 12 h. Plasma containing no apoA-I and LDF were obtained from same plasma and the volumes were adjusted to original plasma volume before experiments. In case of plasma containing no apoA-I, O-3% of labeled cholesterol was found to be esterified. To the contrary, LDF esterified 20 + 5% of labeled cholesterol (n = 5). Thus, it is likely that a large portion of LCAT is dissociated from apoA-I containing particles during ultracentrifugation as suggested by Cheung et al. previously [2l.

LCAT-inactivation of LpA-I and LpA-I/A-II and its subsequent effect on cholesterol efflux from macrophage foam cells To confirm that [“HIFC excreted from foam cells was esterified extracellularly to CE by LCAT associated with LpA-I and LpA-I/A-II as well as to determine a role of the LCAT-mediated esterification in cholesterol efflux from foam cells, we inactivated the LCAT activity of LpA-I and LpA-I/A-II by DTNB and tested these DTNB-treated lipoproteins for their effect on cholesterol efflux. When labeled foam cells were incubated with 400 pg/ml of DTNB-treated LpA-I (D-LpA-I) or LpA-I/A-II (D-LpA-I/A-II), the total radioactivity in the medium did not differ from that in case of incubation with LpA-I or LpA-I/A-II, but the subsequent conversion to CE did not occur, thereby confirming the notion that LCAT associated with these lipoproteins is responsible for esterification of the excreted [“H]FC. Upon incubation with D-LpA-I or D-LpA-I/A-II, cellular [“HICE was decreased by D-LpA-I (lo-15%) (P < 0.05) and by D-LpA-I/A-II (35-45%) (P < 0.005). Under the identical conditions, the amounts of cellular [“HIFC were significantly decreased, by both lipoproteins (60-70%) (P < 0.005). To examine the effect on the net cholesterol efflux, cellular CE and FC mass levels were determined (Fig. 6). D-LpA-I decreased the CE mass only slightly but D-LpA-I/A-II decreased it significantly when compared to control (P < 0.05 or 00051. D-LpA-I and D-LpA-I/A-II did not decrease the cellular FC mass (P > 0.1). Upon treatment with DTNB, the capacity of LpA-I to reduce CE and FC mass was significantly weakened (P < O.OOS), but that of LpA-I/A-II was not significantly affected. These results suggest that LCAT-mediated esterification is

LpA-I

CE

, 100,

LpA-I/A-II -

CE

FC

1

FC

W LCAT(+) q

LCAT(-)

Fig. 6. Effect of LCAT-inactivation of LpA-I and LpA-I/A-II on CE and FC mass in macrophage foam cells. Macrophage foam cells were incubated for 24 h with 400 pg/ml of LpA-I and LpA-I/A-II pretreated with DTNB [LCATt-I], and native LpA-I and LpA-I/A-II [LCAT( + I]. Control experiments were performed without LpA-I and LpA-I/A-II. Cellular lipids were extracted and determined for CE and FC mass. Data are the mean of five experiments with showing the standard error. Mean values of CE and FC for 100% were 114 and 165 nmol/mg cell protein, respectively.

125 important for LpA-I particles to maintain their cellular CE and FC reducing capacity, but not for the LpAI/A-II particles.

TABLE

II

Excreted [‘H]cholesterol re-isolated LpA-I/ A-II

Relationship between subfractions of HDL isolated by ultracentrifu-gation and those isolated by immunological methods

As shown in Figs 4 and 6, cholesterol reducing capacity of HDL, resembles LCAT inactivated LpA-I and that of HDL, resembles LCAT inactivated LpAI/A-II. Thus, to elucidate the relationship between subfractions of HDL and those of lipoprotein containing apoA-I, we isolated LpA-I and LpA-I/A-II from HDL, and HDL,. Based on the protein concentration, 67-75% of the HDL, particles were LpA-I and 2533% were LpA-I/A-II. On the contrary, 70-80% of the HDL, particles were LpA-I/A-II and 20-30% were LpA-I. However, different from LpA-I and LpAI/A-II isolated directly from plasma, LpA-I and LpAI/A-II from HDL, and HDL, did not possess LCAT activity, suggesting LCAT was lost during ultracentrifugation. CE transfer from LpA-I to LpA-I/A-II tion by LCAT

after esterifica-

As both LpA-I and LpA-I/A-II are present in plasma, we examined the combined effects of LpA-I and LpA-I/A-II on cholesterol efflux and the subsequent esterification by LCAT. LpA-I and LpA-I/A-II were combined at an equi-protein ratio and incubated with foam cells. The effects on FC and CE at radioactive and mass levels were indistinguishable from those of LpA-I alone (Figs 2 and 3). Thus, the cholesterol-reducing capacity of LpA-I was not inhibited by LpAI/A-II in our system. This was suggested by in vitro cholesterol efflux experiments using adipocytes [31]. Although these results indicate that the combined effect of LpA-I and LpA-I/A-II on cholesterol efflux is similar to that of LpA-I alone, extracellular events

LpA-I Re-isolated LpA-I LpA-I/A-II Re-isolated LpA-I/A-II

in LpA-I, re-isolated LpA-I, LpA-I/A-II

10m4 cpm/mg

protein

Total [ 3H]cholesterol

Free

Esterified

6.4kO.l 5.4kO.l 6.6kO.l

3.8 + 0.2 3.6 f 0.2 5.9kO.l

2.5 k 0.2 1.8kO.2 * 0.7+0.1

7.OkO.6

*

4.2kO.5

**

2.8kO.3

and

**

Values are expressed as mean + S.E. * Significantly different from LpA-I; P < 0.005. * * Significantly different from LpA-I/A-II; P < 0.005.

occurring after FC excretion into medium showed a marked contrast. The re-isolated LpA-I and LpA-I/AII from culture medium were determined for the esterification of excreted [3Hlcholesterol. As Fig. 7 shows, the % esterification was lower in the re-isolated LpA-I (33 + 2%: mean 5 S.D.1 than with LpA-I alone (39%). In contrast, the esterification of the cholesterol in re-isolated LpA-I/A-II was higher (39 + 5%: mean k S.D.) than LpA-I/A-II alone (10%). The protein ratio of re-isolated LpA-I to LpA-I/A-II remained unchanged. Table II shows the content of [‘HICE of re-isolated LpA-I was significantly lower than that of LpA-I alone, whereas that of re-isolated LpA-I/A-II inversely became higher than that of LpA-I/A-II alone. Since the content of 13H]FC per LpA-I particle was virtually constant during the experiment, it is strongly suggested that CE in LpA-I formed by LCAT is transferred to LpA-I/A-II. In case of LpA-I/A-II, [3H]FC per particle re-isolated was significantly lower than that in LpA-I/A-II alone (Table II). This suggests that LpA-I/A-II, when co-existing with LpA-I, may serve as an acceptor for CE from LpA-I, rather than that for cellular FC. To further confirm the CE transfer from LpA-I to LpA-I/A-II, CE transfer in the cell free system was determined as described under Materials and Methods. These results show that 66 + 5% (mean * S.D., n = 3) of L3HlCE of LpA-I was recovered in LpA-I/A-II, thereby indicating that CE transfer from LpA-I to LpA-I/A-II did occur. Discussion

0

LO

40 (90)

60

80

100

Fig. 7. Percent esterification of excreted [3H]FC on LpA-I and LpA-I/A-II re-isolated from the mixture of LpA-I and LpA-I/A-II. Macrophage foam cells were incubated for 24 h with 400 pg/ml of the mixture of LpA-I and LpA-I/A-II. LpA-I and LpA-I/A-II were re-isolated from the mixture (A-ILp) of LpA-I and LpA-I/A-II in culture medium as described under Materials and Methods. The lipoprotein lipids were extracted, subjected to TLC and radioactivities of FC and CE were determined. Data are the mean of five experiments.

The present studies demonstrate that (i) interaction of LpA-I or LpA-I/A-II with macrophage foam cells induces mass reduction in cholesterol from these cells, (ii) mechanisms involved in this reduction differ. The cholesterol mass reducing capacity of LpA-I is closely linked to the LCAT activity associated with LpA-I while that of LpA-I/A-II is not, (iii) cellular cholesterol esterified in LpA-I is transferred to LpA-I/A-II,

126 and (iv) LpA-I/A-II, when coexisting with LpA-I, may serve as an acceptor for CE rather than that for cellular FC. According to the current concept of the cholesteryl ester cycle in foam cells [21,32], CE-accumulation induced by acetyl-LDL undergoes a continuous cycle of hydrolysis and re-esterification. Incubation with HDL, did not enhance the hydrolysis of CE, but did disrupt the cycle by removing acyl CoA: cholesterol acyltransferase (ACAT) accessible FC (referred to as intraceilular FC in the foIlo~ing discussion) from the cells, thereby suppressing the re-esterification step, leading to a net CE-reduction. In the present study, LpA-I and LpA-I/A-II showed a similar CE-reducing capacity, both in radioactive and mass levels. The experiment using LCAT-inactivated LpA-I and LpA-I/A-II showed that only FC was excreted from foam cells. This indicates that the cellular CE-reduction by these lipoproteins must proceed by removing intracellular FC, not by direct removal of CE. The mechanism for removal of intracellular FC by lipoproteins containing apoA-I remains controversial. Schmitz et al. [33] and Aviram et al. 1341have proposed that specific interaction of apoA-I with plasma membrane is required for removal of intracellular FC. However, Mahlberg et al. presented the data against their notion (specific interaction of apoA-I is not necessary) (351. According to our recent study, treatment of HDL with tetranitromethan (TNM) and dithiobissuccinimidyl-propionate (DSP) which diminished ligand activity (specific binding to its receptor) of apoA-I [36] abolished CE-reducing capacity of HDL [37]. Thus, we assume that the specific interaction of LpA-I and LpAI/A-II with plasma membrane is necessary for CE-reduction. As intracellular FC removal leads to CE-reduction 121,321, our data indicate that the FC generated from CE hydrolysis are similarly removed by LpA-I and LpA-I/A-II. Nonetheless, there is a discrepancy between findings in a tracer study using [3H]FC and the FC mass data; the FC efflux must be accompanied by a fairly large amount of FC influx from these lipoproteins (Fig 2 and 3). Treatment of HDL with TNM and DSP had no effect on the efflux of the cellular radiolabeled FC from foam cells [371. In addition, the incubation of LDL with these cells under similar conditions elicited a response similar to that seen with TN~-HDL. These results suggest that most of the efflux of [“HIFC might reflect a non-specific reaction which occurs between plasma membrane and LpA-I or LpA-I/A-II particles. Thus, the different FC influx from each lipoprotein into foam cells might be responsible for the different capacity of net FC mass reduction between these lipoproteins. To identify the mechanism for the FC influx from LpA-I and LpA-I/A-II into foam cells, we studied the

effect of LCAT on the cholesterol reducing capacity of these Lipoproteins. Fielding and Fielding 1381suggested that a net FC removal from the plasma membrane of fibroblast is closely linked to the LCAT activity. Inhibition of the LCAT activity resulted in an increase in the FC influx from lipoproteins to cells, but had no effect on the corresponding FC efflux from the plasma membrane. This indicates that LCAT might promote the net FC removal from the plasma membrane by reducing a FC influx into fibroblasts. The present study confirmed this in case of LpA-I (FC mass reducing capacity was weakened by inactivation of LCAT with no effect on [“H]FC reductions. LCAT activity is required to make the diffusion gradient between foam cell plasma membrane and LpA-I. In addition, f3H]CE and CE mass reduction by LpA-I was also weakened (Fig. 61. These results suggest that at least in LpA-I, the efficient efflux of FC in the plasma membrane might be a key step in CE reduction as well as specific interaction of LpA-I with plasma membrane. The reason for the lack of effect of LCAT on LpA-I/A-II is not clear, but particle size (size of major LpA-I/A-II particles are smaller than that of LpA-If (Fig. 1) and lipid composition (FC/PL ratio in LpA-I/A-II is smaller than that of LpA-I; 0.13 vs 0.20) may be responsible, as suggested by Phillips et al. 1331.In other words, the diffusion gradient from plasma membrane to LpA-I/A-II would be readily facilitated, without LCAT activity. However, FC influx from LpA-I/A-II was greater than that of LpA-I with LCAT, as judged by the FC mass reducing capacity. Recently, Johnson et al. [40] reported that LpA-I and LpA-I/A-II function equally well in removing cholesterol from rat hepatoma cells, fibroblast and rabbit aortic smooth muscle cells. Differed from our LpA-I and LpA-I/A-II, their samples did not possess LCAT activity, possibly because of different isolation procedures. If they used LpA-I with LCAT activity, cholesterol reducing capacity of their LpA-I might be enhanced. The percent esterification of the excreted cholesterol in LpA-I/A-II was increased, and that in LpA-I was decreased when LpA-I and LpA-I/A-II were combined (Fig 7). Based on the radioactive CE content per mg effector protein (Table II), cholesterol esterified in LpA-I seemed to be transferred to LpA-I/A-II, Experiments with a cell free system confirmed that a significant portion of CE esterified in LpA-I was subsequently transferred to LpA-I/A-II. Cheung et al. already reported that LpA-I particles contained most of cholesteryl esters transfer activity (CETA) [2]. Thus, our present data may suggest that CETP could mediate the CE transfer from LpA-I to LpA-I/A-II. In physiological states, LpA-I/A-II may serve as a specific CE reservoir for LpA-I, rather than that for cellular FC and may be involved in a subsequent CE transfer to the liver or other organs. LpA-I/A-II, when the func-

127 tion of LpA-I being inhibited, may serve as an acceptor for cellular FC. All of these data suggest that both LpA-I and LpA-I/A-II might play an important antiatherogenic role in vivo. However, Puchois et al. and Stampfer et al. recently reported that only the concentration of LpA-I particles were inversely related to the risk of myocardial infarction, but that of LpA-I/A-II were not [4,5]. This may be due to the fact that the concentration of LpA-I/A-II are fairfy constant and that of LpA-I are variable in normolipidemic and dysIipidemic subjects [5-81. Finaily, the relationship between lipoproteins containing apoA-I isolated by immunological methods and the HDL isolated by ultracentrifugation needs to be addressed. When we consider ‘reverse cholesterol transport’ as a function of HDL, we need to explain the discrepancy between the results of epidemiologicai studies [5,41] and experimental in vitro studies (10,421. Epidemiological studies indicated that the pIasma Ievels of HDL, 1411 or both HDL, and HDL, [5] are inversely correlated with the incidence of ischemic heart disease, but e~eriments showed that onty HDL, is involved in the cholesterol efffux. Our data on HDL, and HDL, are consistent with these findings; only HDL, was effective for cholesterol efflux from the foam cells (Fig 4). In the present study, more than 70% of HDLz particles were LpA-I and > 70% of HDL, particles were LpA-I/A-II. These data, taken together, suggest close relationship between HDL, and LpA-I, and between HDL, and LpA-I/A-II. However, LCAT activity was not observed in HDL, and HDL,. This suggests that LpA-I and LpA-I/A-II in HDL, and HDL, fractions iost LCAT activity during ultracentr~fugation. When LCAT activity in LpA-I was inhibited, LpA-I lost most of cholesterol reducing capacity, yet LpA-I/A-II was not affected (Fig. 6). Thus, we assume that the loss of LCAT activity associated with LpA-I in HDL, during ultracentrifugation might explain why HDL, is not effective in reducing cholesterol from foam cells. On the other hand, the loss of LCAT activity associated with LpA-I/A-II make no effect on cholesterol reducing capacity of LpA-I/A-II. This might be a reason why HDL, is effective in cellular cholesterol reduction. All this evidence suggests that in physioIogica1 states, particles in the HDL, fraction must be more active than those in HDL, in reverse cholesterol transport, leading to no discrepancy between epidemiological and experimental studies. Studies using HDL isolated by ultracentrifugation seems to underestimate the actual functions of HDL. Acknowledgments

We are grateful to Drs. Paul Nestel and Noel Fidge for comments on the manuscript and for pertinent advice on technique. This work was supported in part

by Grant-in-aid No. 02807094 for Scientific Research from the Ministry of Education, Science, and Culture of Japan. References 1 Cheung, MC. and Albers, J.J. (1984) J. Biol. Chem. 259, 1220112209. 2 Cheung, MC, Wolf, AC., Lum, K.D., Tollefson, J.H. and Albers, J.J (1986) J. Lipid. Res. 27, 1135-1144. 3 Ghta, T., Hattori, S., Murakami, M., Horiuchi, S., Nishiyama, S. and Matsuda, I. (1989) Clin. Chim. Acta 179, 183-190. 4 Puchois, P., Kandoussi, A., Fievet, P., Fourrier, J.L., Bertrand, M., Koren, E. and Fruchart, J.C. (1987) Atherosclerosis 68, 35-40. 5 Stampfer, M.J., Sacks, F.M., Salvini, S., Will&t, W.C. and Hennekens, C.H. (1991) N. Engl. J. Med. 325, 373-381. 6 Ohta, ‘I’., Hattori, S., Nishiyama, S., Higashi, A. and Matsuda, I. (1989) Metabolism 38, 843-849. 7 Ohta, T., Hattori, S., Nishiyama, S. and Matsuda, I. (1988) J. Lipid. Res. 29, 721-728. 8 Ohta, T., Hattori, S., Murakami, M., Nishiyama, S. and Matsuda, I. (1989) Arteriosclerosis 9, 90-95. 9 Glomset, J.A. (1968) J. Lipid. Res. 9, 155-167. 10 Miller,N.E., La Ville, A. and Crook, D. (1985) Nature 314, 109-Ill. 11 Phillips, M.C., Johnson, W.J. and Rothblat, GH, (1987) Biochim. Biophys. Acta 906, 223-276. 12 Brown, M.S. and Goldstein, G.L. (1983) Ann. Rev. Biochem. 52, 223-261. 13 Gerrity, R.G. (1980) Am. J. Pathol. 103, 181-190. 14 Faggiotto, A., Ross, R. and Harker, L. (1984) Arteriosclerosis 4, 323-340. 15 Have& R.J., Eder, H.A. and Brangdon, J.H. (1955) J. Clin. Invest. 34, 1345-1353. 16 Basu, SK., Goldstein, J.L., Anderson, R.G.W. and Brown, MS. (1976) Proc. Natl. Acad. Sci. USA. 73, 3178-3182. 17 Jonas, A., Hesterberg, L.K. and Drengler, SM. (1978) Biochim. Biophys. Acta 528, 47-57. 18 Takata, K., Horiuchi, S., Araki, M., Shiga, M., Saitoh, M. and Morino, Y. (1988) J. Biol. Chem. 29, 14819-14825. 19 Ohta, T., Matsuda, I. (1981) Clin. Chim. Acta 117133-143. 20 Hara, A. and Radin, N.S. (1978) Anal. Biochem. 90, 420-426. 21 Brown, M.S., Ho, Y.K. and Goldstein, J.L. (1980) J. Biol. Chem. 2.55, 9344-9352. 22 Heider, J.G. and Boyett, R.L. (1978) J. Lipid. Res. 19, 514-518. 23 Fielding, C.J., Fielding, P.E. (1981) J. Biol. Chem. 256, 2102-2104. 24 Goto, Y., Akanuma, Y., Harano, Y. (1986) J. Clin. Biochem. Nutri. 1, 73-88. 25 Allain, CC., Poon, L.S., Chan, C.S.G., Richmond, W., Fu, P.C. (1974) Clin. Chem. 20, 470-475. 26 Bucolo, G., Yabut, J., Chang, T.Y. (1975) Ciin. Chem. 21, 420424. 27 Huang, H.J., Kuwan, J.W., Guilbaut, G.G. (197.5) Clin. Chem. 21, 1605-1608. 28 Bartlett, G.R. (1959) J. Biol. Chem. 234, 466-468. 29 Lowry, O.H., Rosebrough, N.J, Farr, A.L., Randall, R.J. (1951) J. Biol. Chem. 193, 265-275. 30 Weber, K. and Gsborn, M. (1969) J. Biol. Chem. 244, 4406-4412. 31 Barkia, A., Puchois, P., Ghalim, N., Torpier, G., Barbaras, R., Ailhaud, G. and Fruchart, J.C. (1991) Atherosclerosis 87, 135-146. 32 HO, Y.K., Brown, MS. and Goldstein, J.L. (1980) J. Lipid. Res. 21,391-398. 33 Schmitz, G., Robenek, H., Lohma~n, U. and Assmann, G. (1985) EMB0.J. 4: 613-622.

128 34 Aviram, M., Bierman, E.L. and &am, J.F. (19891 J. Lipid. Res.

35 36 37 38

30. 65-76. Mahlberg, FM., Glick, J.M., Lund-Katz, S. and Rothblat, G.H. (1991) J. Biol. Chem. 26619930-19937. Miyazaki, A., Rahim, A.T.M., Araki, S., Morino, Y. and Horiuchi, S. (1991) Biochim. Biophys. Acta 1082, 143-151. Miyazaki, A., Rahim, A.T.M., Ohta. T., Morino, Y., and Horiuchi, S. (1992) Biochim. Biophys. Acta, 1126, 73-80. Fielding, C.J. and Fielding, P.E. (1981) Proc. Natl. Acad. Sci. USA. 78, 3911-3Y14.

39 Phillips, M.C., Johnson, W.J. and Rothblat, G.H. (1987) Biochim. Biophys. Acta 904, 223-276. 40 Johnson, W.J., Kilsdonk, E.P.C., Tol, A.V., Phillips, M.C., and Rothblat, G.H. (1991) J. Lipid. Res. 32, 1993-2000. 41 Miller, N.E., Hammett, F. and Saltissi, S. (1981) Br. Med. .i. 2X2. 1741-1744. 42 Oram, J.F. (1983) Arteriosclerosis 3, 420-432.