INTRODUCTION

Lipoprotein-dependent reverse cholesterol transport is believed to be the most important mechanism for removal of cholesterol from extrahepatic cells. In some cells, however, there are alternative mechanisms. Recently, we showed that cultured human alveolar macrophages have a very high capacity to eliminate cholesterol as 27-hydroxycholesterol and 3-hydroxy-5-cholestenoic acid (1). This oxidative mechanism has the potential to eliminate up to about half of the content of cholesterol in the cultured cells in 24 h. Both the alcohol and the acid are primary products of the sterol 27-hydroxylase, and 3-hydroxy-5-cholestenoic acid is thus formed by three consecutive hydroxylations at the same methyl group (2). We have also demonstrated that there is a continuous net flux of 27-oxygenated products to the liver, where these metabolites are efficiently converted into bile acids (3). It was shown that up to 4% of the total formation of bile acids in humans may occur by a mechanism involving extrahepatic 27-hydroxylation of cholesterol (3).

The importance of this oxidative mechanism for cholesterol homeostasis may be illustrated by the fact that subjects with a genetic disease lacking sterol 27-hydroxylase (cerebrotendinous xanthomatosis, CTX)¹ may get premature atherosclerosis despite normal circulating levels of cholesterol (4).

The relative importance of macrophages for the net flux of 27-oxygenated steroids to the liver is not known. The sterol 27-hydroxylase seems to be present in most organs and tissues (5-7). Human umbilical endothelial cells (1) as well as bovine aortic endothelial cells (8) have also been shown to have some capacity to convert cholesterol into both 27-hydroxycholesterol and 3-hydroxy-5-cholestenoic acid, although the relative amount of 3-hydroxy-5-cholestenoic acid appeared to be very low. On the basis of this finding, Javitt (9) suggested that there might be a flux of 27-hydroxycholesterol from the vascular endothelium to the liver.

Secretion of 27-oxygenated products from both macrophages and endothelial cells in culture appeared to be stimulated by the addition of cholesterol to the culture medium (1, 8). It was shown that macrophages could take up deuterium-labeled cholesterol added to the culture medium and excrete part of this labeled cholesterol back to the medium as 27-oxygenated metabolites (1). It was also shown that loading of the macrophages with a 10-fold excess of cholesterol results in a doubled excretion of 27-oxygenated products (3). The cholesterol-induced increase in secretion of 27-oxygenated products did not appear to be a consequence of increased amount of sterol 27-hydroxylase enzyme protein (3). However, the effect of a cholesterol load on the pattern of products formed has not been studied systematically, and the importance of different acceptors in the medium for the flux of 27-oxygenated metabolites from the cells is not known.

The method used for the addition of cholesterol to cultured cells is critical for the degree of uptake and metabolism of this cholesterol. In our experiments with macrophages, cholesterol was added as lipoprotein or dissolved in ethanol. In a study by Reiss et al. (8) concerning production of 27-hydroxycholesterol from cultured bovine aortic endothelial cells, cholesterol was added dissolved in cyclodextrin. To compare the capacity of different cell types to produce 27-oxygenated metabolites, standardized conditions must be used.

In the present work, we have cultured human lung macrophages, human monocyte-derived macrophages, human umbilical vein endothelial cells, and bovine aortal endothelial cells as well as skin fibroblasts under identical conditions and measured the capacity of all these cells to secrete 27-oxygenated products into the culture medium. We have also used monocyte-derived macrophages from patients with CTX, known to lack sterol 27-hydroxylase. In addition, we have studied the effect of the addition of serum lipoproteins, albumin, or a lipophilic acceptor to cultures of macrophages and endothelial cells and compared the efficiency of the sterol 27-hydroxylase mechanism with that of the HDL-mediated reverse cholesterol transport in macrophages under different conditions.

EXPERIMENTAL PROCEDURES

Deuterium-labeled 27-hydroxycholesterol and unlabeled 3-hydroxy-5-norcholestenoic acid were the same as those used previously (1, 3). Tritium-labeled 27-hydroxycholesterol (200 × 10⁶ cpm/mg) was prepared as described previously (10). Medium 199, fetal calf serum, and delipidized serum were obtained from Life Technologies, Inc. The rabbit antibody toward human sterol 27-hydroxylase was a generous gift from Prof. David Russell (University of Texas Southwestern Medical Center, Dallas, TX). This antibody is directed against amino acids 15-28 of the sterol 27-hydroxylase protein. The secondary antibody (goat antirabbit IgG-horseradish peroxidase conjugate) was obtained from Bio-Rad. Apolipoprotein A1 was obtained from Sigma.

The patients with CTX were those previously defined with respect to the site of mutation in the sterol 27-hydroxylase gene. Two of the patients had a deletion of thymine in exon 4, and one had a splice junction mutation in intron 4 as described previously (11). The patients (20, 21, and 15 years, respectively) had xanthomas and elevated serum cholestanol levels typical for the disease (12). Blood was drawn from these patients and from healthy volunteers for isolation of monocytes (see below). Bronchoalveolar lavage was performed on patients with pulmonary malignancies undergoing bronchoscopy for diagnostic purposes. From a clinical point of view, the patients represented a homogenous population. The ethical aspects of the study were approved by the Ethical Committee at Huddinge Hospital.

Human alveolar macrophages were isolated from bronchoalveolar lavage fluid as described (1, 13). The cells were allowed to adhere to plastic culture flasks and cultured in minimum essential medium supplemented with 10% fetal calf serum, benzylpenicillin (400 units/ml), and streptomycin (0.2 mg/ml) as described previously in Ref. 1. The cells were cultured (3-ml volumes) in dishes at 37 °C in an atmosphere of 5% CO₂ in air for 24 h. The number of macrophages in each well varied between 1 and 3 × 10⁶ in different experiments, corresponding to 0.4-1.5 mg of cell protein. In some experiments, the concentration of fetal calf serum was varied between 0.8 and 13%. In other experiments, the fetal calf serum was replaced by 60 µl of delipidized serum (Ultroser). In some experiments, cholesterol (20 µM) was added to the culture medium dissolved in Ultroser (60 µl) or a solution containing 4.5 µmol of cyclodextrin.

Peripheral blood was drawn from healthy volunteers or CTX patients (55 ml from each), and polymorphonuclear cells were isolated by centrifugation in Ficoll-Hypaque and incubated (RPMI 1640 supplemented with pyruvate, glutamine, antibiotics, and 2% heat-inactivated human AB serum) for 1.5 h at 37 °C at a concentration of 4-5 × 10⁶ in six-well plates. The nonadherent lymphocytes were removed, and the adherent monocytes (approximately one million/well) were incubated for 6 days to differentiate to macrophages. At that time, the medium was changed to contain 10% fetal calf serum, and culture was performed under the same conditions as above. In some experiments, medium was removed and analyzed after different periods of time (0-60 h) to ensure a linear conversion of cholesterol into 27-oxygenated products with time.

Human endothelial cells were isolated from umbilical veins as described previously (14, 15). More than 99% of the cells recovered were endothelial cells. Medium 199 with 20% fetal calf serum and 1% L-glutamine was used as the feeding medium. Each milliliter contained 100 µg of endothelial mitogen, 100 µg of benzylpenicillin, and 100 µg of streptomycin. The medium was adjusted to pH 7.4 with 2.5% NaHC0₃ and filter-sterilized. The cell suspension was transferred to tissue culture dishes coated with gelatin. Feeding medium was added to a final volume of 1.5 ml/well. In general, each well contained 1.5 × 10⁶ cells, corresponding to about 0.8 mg of protein. Finally, the dishes were incubated at 37 °C under the same conditions as above.

Bovine aortae (25-30 cm) were excised from newly killed animals and immediately immersed in phospate-buffered saline containing 0.02 M glucose and gentamycin (50 µg/ml). Small arteries were sutured, and the vessels were perfused with 100 ml of phosphate-buffered saline. The aorta was clamped and filled with 100 ml of phosphate-buffered saline containing 0.5 mg/ml collagenase and incubated for 15 min at 37 °C. The solution was removed, and the collagenase-treated aorta was washed gently with 50-60 ml of Dulbecco's modified Eagle's medium. The aorta was filled with 20-30 ml of medium containing 10% fetal calf serum. The vessel was vigorously shaken back and forth 15-20 times. The medium, containing released endothelial cells, was then carefully withdrawn and divided into 6-10 plastic dishes (3-ml volume) and incubated at 37 °C in an atmosphere of 5% CO₂ in air for 1-2 h. Nonadherent cells were washed away, and fresh medium supplemented with 10% fetal calf serum, streptomycin (l00 µg/ml), and benzylpenicillin (100 units/ml) was added. The cells were cultured under these conditions for 5-7 days until they became confluent, changing the medium every 48 h. The final experiments with the cells were performed under the same conditions as above, using about 2 × 10⁶ cells/well, corresponding to about 1 mg of protein.

Human skin fibroblasts were isolated from a control subject and a CTX patient under conditions described previously (7). The final experiments with the cells were performed under the same conditions as above, using about 3 × 10⁶ cells/well, corresponding to about 0.9 mg of protein.

Lung macrophages were cultured in minimum essential medium containing 2% Ultroser together with 1.5 × l0⁶ cpm [4-¹⁴C]cholesterol (55 mCi/mmol) for 48 h and then recultured with the same amount of [4-¹⁴C]cholesterol for another 48 h under the same conditions. On the fifth day, the medium was removed, and the cells were washed twice with phosphate-buffered saline. The macrophages were then cultured in minimal essential medium alone or with the addition of LDL (50 or 150 µg/ml on a protein basis), HDL (50, 150, or 450 µg/ml on a protein basis), albumin (50 or 150 µg/ml), or apolipoprotein A1 (50 or 150 µg/ml). The LDL fraction contained 3.1 µg of cholesterol/µg of protein, and the HDL fraction contained 0.6 µg/µg of protein. The medium was then removed, and the cells were harvested. An aliquot of the medium (1 ml) was extracted with Folch solution and applied to thin layer chromatography using toluene/ethyl acetate/acetic acid (30:70:1) as the mobile phase. The radioactivity in the labeled cholesterol and in labeled 27-oxygenated products was measured using a plate scanner (LB 284/285 from Berthold Co.). In this chromatography, the labeled 27-hydroxycholesterol could not be separated from the labeled 3-hydroxy-5-cholestenoic acid, and the two products appeared as one homogenous peak. In some experiments (see Fig. 1), part of the above lipid extract was subjected to radio HPLC, using an A-280 radioactivity detector (Radiomatic Inc., Tampa, FL) and a YMC pack ODS-A column. The column was eluted with methanol/water/acetic acid 90:10:0.01 (v/v/v) for 20 min, followed by a linear gradient to pure methanol for 5 min and finally isocratic elution with pure methanol for another 25 min. The flow rate was 1.0 ml/min. This chromatographic system was able to separate the two 27-oxygenated products. In addition, the content of cholesterol and 27-oxygenated sterols in the medium and in the cells was determined by combined gas chromatography-mass spectrometry as described below.

Lung macrophages were cultured in minimal essential medium as above together with 0.26 × 10⁶ cpm 27-hydroxycholesterol for 24 h in the presence and absence of 20 µM cyclosporin A, an inhibitor of sterol 27-hydroxylase (1). The medium was extracted and analyzed by radio HPLC as above.

To a defined volume (1-3 ml) of the cell medium, 0.5 µg of 3-hydroxy-5-norcholestenoic acid and 2 µg of 27-[²H₅]hydroxycholesterol were added. The medium was acidified with hydrochloric acid, diluted to 10 ml with water, and extracted with 15 ml of diethyl ether. The ether phase was washed with water until neutral, and the solvent was removed under reduced pressure. The residue was dissolved in 0.5 ml of chloroform and fractionated on a Bond-Elut NH₂ cartridge (1). The neutral lipid fraction, containing 27-hydroxycholesterol and the fatty acid fraction containing 3-hydroxy-5-cholestenoic acid, were collected. The fractions were blown to dryness under a stream of argon, the fatty acid fraction was methylated with diazomethane, and both fractions were converted into trimethylsilyl ethers.

A Hewlett-Packard 5890 GC coupled to a Hewlett-Packard 5970 MSD mass spectrometer was used for the analysis. The chromatographic conditions were those described previously (1). ²H₅-Labeled 27-hydroxycholesterol was used as an internal standard for 27-hydroxycholesterol, and 3-hydroxy-5-norcholestenoic acid was the internal standard for 3-hydroxy-5-cholestenoic acid as described previously (1, 3).

Crude homogenates of the cells were subjected to SDS-polyacrylamide gel electrophoresis as described by Laemmli (16) with 10% gels. The proteins were transferred to a nitrocellulose membrane by the method of Towbin et al. (17). Blocking was performed with the use of 3% gelatin in phosphate-buffered saline for 1.5 h. The nitrocellulose membrane was incubated first with the primary antibody (diluted 1:500) for 2.5 h and then with goat anti-rabbit IgG carrying horseradish peroxidase (after dilution, 1:3000). The bands were visualized with a reagent for peroxidase conjugates.

RESULTS

A number of different cells of human and bovine origin were cultured and tested with respect to their capacity to secrete 27-oxygenated products of cholesterol into the medium under identical conditions in the presence of 10% fetal calf serum. Table I summarizes the results. When expressed as the secretion of 27-oxygenated products into the medium/million cells/24 h, the capacity of the human lung alveolar macrophages was about 5-fold higher than that of human monocyte-derived macrophages and 35-fold higher than that of human endothelial cells from the umbilical cord. The activity of human fibroblasts was about half that of human endothelial cells. There was a relatively high variation in the capacity of different macrophage preparations to secrete 27-oxygenated products. No correlation could be found between this capacity and presence of a lung malignancy in the donor.

Table I. Capacity of different cultured cells to secrete 27-oxygenated products into the culture medium Human lung alveolar macrophages were isolated from three different lung cancer patients; human monocyte-derived macrophages were from three different healthy volunteers and three CTX patients as described under "Materials and Methods." Endothelial cells were isolated from three different umbilical cords and two different bovine aortas as described. Human fibroblasts were isolated from one healthy volunteer and one CTX patient. All experiments were performed under the same standard conditions in the presence of 10% fetal calf serum in 3 ml of minimum essential medium for 24 h at 37 °C (see "Experimental Procedures"). The amount of 27-oxygenated products was determined by combined gas chromatography-mass spectrometry. Cultured cells Secretion of 27-oxygenated products into the medium ng/106 cells/24 h Human lung alveolar macrophages (n = 3) 1400 (630-3050) Human monocyte-derived macrophages (n = 3)  300 (60-700) Monocyte-derived macrophages from CTX patients (n = 3) <2 Human endothelial cells from umbilical cord (n = 3) 38 ± 4a Human fibroblasts (n = 1) 20 Fibroblasts from a patient with CTX <2 Bovine endothelial cells (n = 2)  9 (4 and 14) a Mean ± S.D.

Monocyte-derived macrophages and fibroblasts from CTX patients had no significant capacity to secrete 27-oxygenated products into the culture medium. Thus, it is evident that the sterol 27-hydroxylase is critical for the formation of these products.

The formation of 27-oxygenated products appeared to be about linear during 24-60 h of culture in all the above active cells (results not shown). In accordance with our previous studies, the ratio of 27-hydroxycholesterol to 3-hydroxy-5-cholestenoic acid varied between 0.05 and 0.5 in the experiments with human alveolar macrophages. Similar ratios (0.2-0.5) between the two products were obtained also in the experiments with monocyte-derived macrophages. In the experiments with human endothelial cells, however, this ratio varied between 5 and 10, and in the experiments with bovine endothelial cells the ratio varied between 1.5 and 3.

No significant amounts of 3-hydroxy-5-cholestenoic acid were formed from the fibroblasts under the conditions employed.

7-Hydroxylation of 27-hydroxycholesterol has been demonstrated in human diploid fibroblasts (18). There was however no significant formation of metabolites of 27-hydroxycholesterol or 3-hydroxy-5-cholestenoic acid in any of the above experiments with macrophages and endothelial cells. Due to the small overall conversion, a slight further metabolism of 27-hydroxycholesterol could not be ruled out in the experiments with fibroblasts.

Oxysterols are known to be esterified in the circulation, most likely due to the action of lecithin:cholesterol acyltransferase. Under the in vitro conditions used here, however, there was no significant esterification of 27-hydroxycholesterol as judged from measurements before and after alkaline hydrolysis and experiments with radioactive cholesterol (see below).

Western blotting was performed under the conditions previously described (1) on human alveolar macrophages, human umbilical vein endothelial cells, and human fibroblasts. Under the conditions employed, a significant band was obtained only with the alveolar macrophages (results not shown). Thus, alveolar macrophages contain considerably more sterol 27-hydroxylase protein than the other cells.

The formation of 3-hydroxy-5-cholestenoic acid in the above experiments with macrophages may be due to a sterol 27-hydroxylase-mediated hydroxylation of 27-hydroxycholesterol or to a dehydrogenase-dependent oxidation of the same compound. As shown in Fig. 1, there was a very efficient conversion of labeled 27-hydroxycholesterol into 3-hydroxy-5-cholestenoic acid in cultured lung macrophages. In the presence of 20 µM cyclosporin A, an efficient inhibitor of sterol 27-hydroxylase (1), very little conversion of 27-hydroxycholesterol was obtained. This finding supports the contention that the sterol 27-hydroxylase is responsible for formation of 3-hydroxy-5-cholestenoic acid from 27-hydroxycholesterol. In accordance with this, the addition of an alcohol dehydrogenase inhibitor, 4-methylpyrazole, had no effect on the conversion of labeled 27-hydroxycholesterol into 3-hydroxy-5-cholestenoic acid in the macrophages (results not shown).

In all the studies above, cholesterol was added as lipoproteins present in fetal calf serum. As shown in Table II, the addition of fetal calf serum to cultured macrophages doubled the secretion of 27-oxygenated products into the medium. The addition of more than 100 µl of fetal calf serum did not further increase this secretion. The ratio between the secreted 27-hydroxycholesterol and 3-hydroxy-5-cholestenoic acid increased, however, with the amount of fetal calf serum added from 0.09 to 0.46 

Table II. Effect of the addition of fetal calf serum on secretion of 27-oxygenated products into the medium and accumulation of the products in the macrophages Human lung alveolar macrophages were isolated from one specific patient as described under "Experimental Procedures" and cultured under standard conditions for 24 h at 37 °C without or with 25-400 µl of fetal calf serum (corresponding to 0.8-13.3%). The amount of 27-oxygenated products was determined by combined gas chromatography-mass spectrometry. Addition Total secretion into medium of 27-OH + 27-COOH Ratio of 27-OH to 27-COOH Total content in cells of 27-OH + 27-COOH Ratio of 27-OH to 27-COOH ng/mg protein × 24 h ng/mg protein None 447 0.09 438 6.5 Fetal calf serum   25 µl 780 0.07 334 7.8   100 µl 972 0.16 246 7.2   400 µl 951 0.46 268 12

The residual content of 27-oxygenated products in the macrophages, harvested after the above secretion experiment, decreased with an added amount of fetal calf serum from 438 ng/mg cell protein to 268 ng/mg cell protein. The ratio of 27-hydroxycholesterol to 3-hydroxy-5-cholestenoic acid was 25-75 times higher in the cells than in the medium, demonstrating that 3-hydroxy-5-cholestenoic acid was secreted efficiently from the cells into the medium.

The total production of 27-oxygenated products (sum of products in the medium and in the cells) thus increased by 37% in the above experiment, whereas the content of 27-oxygenated products in the cells decreased by 40%. In addition to serving as a source of cholesterol for the sterol 27-hydroxylase system, it is evident that fetal calf serum must facilitate transport of the products from the cells into the medium.

Human alveolar macrophages and bovine aortic endothelial cells were used for a more detailed study on the effect of the addition of cholesterol by different means.

As shown in Table III, experiments 2 and 3, the addition of cyclodextrin to macrophages cultured in the presence of fetal calf serum or delipidized serum (Ultroser) had a slight stimulatory effect on the secretion of 27-oxygenated products to the medium but a marked effect on the ratio between 27-hydroxycholesterol and 3-hydroxy-5-cholestenoic acid. This effect is probably due to a stimulation of the transport of 27-hydroxycholesterol from the cells, thus reducing the substrate for formation of 3-hydroxy-5-cholestenoic acid. Similar findings were obtained also in experiments with endothelial cells. When additional cholesterol was added to the cells together with the cyclodextrin, the total secretion of 27-oxygenated products increased 3-5-fold in the experiments with the macrophages and 8-11-fold in the experiments with the endothelial cells. Most likely, this stimulation is due to increased substrate availability for the enzyme. When the same amount of cholesterol as above was added dissolved in ethanol, there was, however, only a very slight degree of stimulation (results not shown).

Table III. Effect of the addition of cyclodextrin and cyclodextrin plus cholesterol on secretion of 27-oxygenated products from macrophages and endothelial cells Human lung alveolar macrophages isolated from three different patients and endothelial cells isolated from two bovine aortas were cultured under standard conditions for 24 h at 37 °C with and without certain additions. When added, the amounts were as follows: cyclodextrin, 12 µl; Ultroser, 60 µl; fetal calf serum, 0.3 ml; cholesterol, 20 µM. The two 27-oxygenated products were determined by combined gas chromatography-mass spectrometry. Addition Macrophages Endothelial cells Total secretion Ratio of 27-OH to 27-COOH Total secretion Ratio of 27-OH to 27-COOH ng/mg protein × 24 h ng/mg protein × 24 h Exp. 1    None 162 <0.05 4  a   Cyclodextrin 557 1.3 10   a   Cyclodextrin +   cholesterol 9711 2.1 382 17 Exp. 2    Fetal calf serum 2227 0.09 27 1.5   Fetal calf serum   + cyclodextrin 2659 1.00 42 3.7   Fetal calf serum   + cyclodextrin   + cholesterol 8034 0.91 492 7.0 Exp. 3    Ultroser 542 0.25 8   a   Ultroser +   cyclodextrin 818 1.3 18   a   Ultroser +   cyclodextrin +   cholesterol 3915 2.1 151 8.4 a , in these experiments, the conversion into 3-hydroxy-5-cholestenoic acid was too low to allow an accurate determination of the ratio.

In separate experiments, it was shown that the optimal amount of cyclodextrin was 1.5 µmol/ml of culture medium, and this amount of cyclodextrin was used in all of the experiments.

In all the above experiments, the total formation of 27-oxygenated products/mg of cell protein was 16-80-fold higher in the macrophages than in the endothelial cells. Several additional experiments with different preparations with or without cyclodextrin were performed with similar results (results not shown).

Macrophages were also cultured in the complete absence of any source of exogenous cholesterol or delipidized serum (Table III, experiment 1). The very low secretion of 27-oxygenated products increased 3-fold after the addition of cyclodextrin. The addition of cholesterol dissolved in cyclodextrin gave the same very high secretion as in the other experiments.

It is evident that fetal calf serum and probably also cyclodextrin must have two effects in the above experiments: stimulation of transport of 27-hydroxycholesterol from the cells into the medium and transport of substrate cholesterol into the cells and the sterol 27-hydroxylase system.

Part of the stimulatory effect of the cyclodextrin and the fetal calf serum on the conversion of cholesterol into 27-oxygenated products could be due to a product inhibition of the sterol 27-hydroxylase that might be released after the increased transport of the product from the cells.

25-Hydroxycholesterol is a structural analogue to the product 27-hydroxycholesterol and could be expected to be an inhibitor of the sterol 27-hydroxylase.

As shown in Table IV, the addition of 25-hydroxycholesterol (5 µg/ml) to the culture medium decreased the secretion of 27-oxygenated products into the medium by about 40%. This inhibition was associated with a marked increase in the ratio between 27-hydroxycholesterol and 3--hydroxy-5-cholestenoic acid from 0.04 to 0.21. 

Table IV. Effect of the addition of 25-hydroxycholesterol on secretion of 27-oxygenated products from cultured macrophages Human lung alveolar macrophages were isolated from one specific patient as described under "Experimental Procedures." The cells were cultured in triplicate under standard conditions in the presence of 10% fetal calf serum with or without 25-hydroxycholesterol (5 µg/ml) for 24 h at 37 °C. The oxysterol was added in ethanol (15 µl). The two 27-oxygenated products were determined by combined gas chromatography-mass spectrometry. Addition Total conversiona Ratio of 27-OH to 27-COOH ng/3 × 106 cells Ethanol 3035  ± 117 0.04 Ethanol + 25-OH-chol 1836  ± 508 0.21 a Mean ± S.D.

Albumin and lipoproteins are likely to be the most important factors in fetal calf serum affecting secretion of 27-oxygenated products from the macrophages. The effect of these factors were therefore studied in separate experiments with use of cultured lung macrophages that had been preloaded with labeled cholesterol. This experimental design also allowed measurement of secretion of cholesterol. In addition, the amounts of the 27-oxygenated products were measured by combined gas chromatography-mass spectrometry in the medium and in the cells.

When cultured macrophages preloaded with labeled cholesterol were exposed to medium only, there was a significant secretion of both labeled 27-oxygenated products and labeled cholesterol (Table V). The addition of albumin in relatively low concentrations (50 and 150 µg/ml) increased secretion of the radioactive 27-oxygenated products about 5-fold with little or no increase in the secretion of radioactive cholesterol. 3-Hydroxy-5-cholestenoic acid was the major labeled product (Fig. 2), and the ratio between the alcohol and the acid in the medium varied between 0.1 and 0.15 in different experiments.

Table V. Flux of radioactivity from cultured macrophages preloaded with [4-14C]cholesterol: effect of albumin, LDL, and HDL in the culture medium * Human alveolar macrophages were preloaded with [4-14C]cholesterol as described under "Experimental Procedures" and cultured in the presence of albumin, LDL, HDL, and apolipoprotein A1 (0-150 µg/ml) for 24 h. Different preparations of alveolar cells were used for each set of additions. The content of radioactive cholesterol in the cells (expressed as cpm/106 cells) at the start of the experiment is indicated. The amount of radioactivity (cpm/106 cells) in cholesterol and in 27-oxygenated metabolites present in the culture medium at the end of the experiment is also indicated. The content of cholesterol in the medium was less than 0.01 mg in the experiments without additions to the medium or an addition of albumin or lipid-free apoA1. In the LDL experiments, the content of cholesterol in the medium was 0.45 and 1.35 mg, respectively, and in the HDL experiments it was 0.09 and 0.28 mg, respectively.

In these experiments with albumin, 91 and 79%, respectively, of the total amounts of 27-hydroxycholesterol was retained in the macrophages, indicating a relative block in the efflux of this steroid from the cells. In contrast, the corresponding figures for retention of 3-hydroxy-5-cholestenoic acid were 8 and 5%, indicating an efficient efflux of this polar metabolite of cholesterol.

When LDL was added to the cultured macrophages, preloaded with labeled cholesterol, in physiological concentrations (extracellular fluid), the secretion of labeled 27-oxygenated products increased about 3-fold (Table V). The secretion of labeled cholesterol into the medium was also increased about 3-fold. The ratio between the radioactive alcohol and acid secreted varied between 1 and 2 in different experiments (Fig. 2).

In these experiments with LDL, 15-32% of the total amounts of the two 27-oxygenated products were retained in the macrophages.

When HDL was added to the cultured macrophages preloaded with cholesterol in the same concentrations as for LDL (50 and 150 µg/ml), the secretion of labeled 27-oxygenated products increased about 3-fold (Table V). The ratio between the labeled alcohol and acid varied between 2 and 5 in different experiments with HDL (Fig. 2). The secretion of radioactive cholesterol increased about 6-fold, however, as a consequence of the addition of HDL.

In these experiments with HDL, between 20 and 33% of the two 27-oxygenated products were retained in the macrophages.

Increasing the concentration of HDL up to 450 µg/ml did not further change the yield or pattern of products (results not shown). This concentration is believed to be a physiological concentration of HDL in extracellular fluid (see below).

The HDL-particles used in the above experiments contain both apolipoprotein A and lipids. To study the relative importance of these two components, experiments were also performed with lipid-free apolipoprotein A1 (50 and 150 µg/ml) (Table V). When lung macrophages preloaded with labeled cholesterol were exposed to such particles, the flux of labeled cholesterol into the medium was considerably lower than in the experiment with HDL. The effect of these particles on total efflux of labeled 27-oxygenated products was similar to that obtained in the HDL and LDL experiments. However, the dominating product secreted was 3-hydroxy-5-cholestenoic acid (Fig. 2).

It is evident from the results shown in Table V that the ratio of secreted 27-oxygenated products to secreted cholesterol from the macrophages was about 2:1 in the albumin experiment, about 1:2 in the LDL experiment, and about 1:6 in the HDL experiment. This change was paralleled by an increase in the ratio of the alcohol to the acid, from about 1:8 in the albumin experiment to about 1:1 in the LDL experiment and about 2:1 in the HDL experiment.

By adding together the amounts of 27-oxygenated steroids in the medium and in the cells under the different conditions, the stimulatory effect of the three different additions on the total production of these compounds could be calculated. The maximal stimulatory effect of albumin was about 60%, whereas the corresponding figures for HDL and LDL were 50 and 140%, respectively.

DISCUSSION

It is evident that macrophages have considerably higher capacity to convert cholesterol into 27-oxygenated products and to secrete these products into the medium than have endothelial cells and fibroblasts. In accordance with this, macrophages but not endothelial cells or fibroblasts gave a significant signal for sterol 27-hydroxylase in our Western blotting experiments. In a recent study using a more sensitive histochemical technique, we found that the highest sterol 27-hydroxylase immunoreactivity in human atherosclerotic plaques was in macrophages (19). There was, however, also some staining in endothelial cells.

The present results show that the production and pattern of 27-oxygenated products are dependent upon three major factors: the amount of active sterol 27-hydroxylase enzyme, the availability of substrate cholesterol, and the presence of an acceptor in the medium.

Sterol 27-hydroxylase enzyme activity is critical for the formation of 27-oxygenated products as evident from the experiments with macrophages from the sterol 27-hydroxylase-deficient patients. Since there was no conversion of labeled 27-hydroxycholesterol into 3-hydroxy-5-cholestenoic acid when the sterol 27-hydroxylase was inhibited by cyclosporin, sterol 27-hydroxylase must be responsible also for the conversion of 27-hydroxycholesterol into 3-hydroxy-5-cholestenoic acid.

Since formation of 3-hydroxy-5-cholestenoic acid requires three hydroxylation steps (Fig. 3) whereas formation of 27-hydroxycholesterol requires only one step, relatively more acid can be expected as product if the ratio of hydroxylase enzyme to substrate is high. In accordance with this, the relative formation of the acid was always highest in cells with a high content of sterol 27-hydroxylase. A reduced enzyme activity can be expected to result in a decreased ratio of 3-hydroxy-5-cholestenoic acid to 27-hydroxycholesterol. The result of the present experiment with the inhibitor 25-hydroxycholesterol is in accordance with this contention (Table IV).

Very little is known about the regulation of sterol 27-hydroxylase activity in macrophages. In a previous study we showed, however, that the cholesterol content in the cells was not correlated to the enzyme protein mass (3). Attempts to up-regulate the enzyme activity in macrophages by the addition of various hormones have failed thus far.²

As previously shown (3), exposure of cultured macrophages and endothelial cells to cholesterol in the medium resulted in increased secretion of 27-oxygenated products. The secretion was higher when cholesterol was added in cyclodextrin than when added in ethanol or in the form of lipoproteins. A clear positive correlation between circulating levels of cholesterol and 27-hydroxycholesterol has been reported (20), consistent with substrate availability as a limiting factor also under in vivo conditions.

If substrate availability and degree of substrate saturation of the sterol 27-hydroxylase is a critical factor, transfer of cholesterol to this enzyme system may be limiting. It is noteworthy that the sterol 27-hydroxylase is located at the inner membranes of the mitochondria (21). The nature of the mechanism behind transfer of cholesterol from the outer to the inner mitochondrial membranes in the present cells is not known. In adrenals, ovaries, and testis in rodents and humans, a steroidogenic acute regulatory protein (StAR) may be responsible for such a transfer (for a review, see Ref. 22). It was recently shown that StAR could enhance also sterol 27-hydroxylase activity in a COS cell system (23). Whether a StAR-mediated mechanism is of importance for the cholesterol transport in macrophages remains to be established.

An increased load of cholesterol results in a decreased ratio of enzyme to substrate. Since cholesterol competes with the primary product 27-hydroxycholesterol for the active site, such a change can be expected to lead to reduced formation of acid (Fig. 3). In accordance with this, loading both lung macrophages and endothelial cells with cholesterol resulted in increased conversion and increased ratios of 27-hydroxycholesterol to 3-hydroxy-5-cholestenoic acid (Tables II, III, and V).

Our results are consistent with a need for some type of lipophilic acceptor for efflux of 27-hydroxycholesterol whereas albumin is required for efflux of the more polar 3-hydroxy-5-cholestenoic acid. In the presence of albumin only in the medium, 3-hydroxy-5-cholestenoic acid was found to be excreted very rapidly from the cells. In the absence of a suitable lipophilic acceptor, an intracellular accumulation of 27-hydroxycholesterol was observed. As expected, such an accumulation led to increased conversion into the acid with a subsequent lipoprotein-independent efflux of this acid. In the presence of an effective lipophilic acceptor for 27-hydroxycholesterol like cyclodextrin or HDL, the relative excretion of 3-hydroxy-5-cholestenoic acid decreased, most probably due to reduced amount of substrate for further oxidation to the acid (Fig. 3, Tables II and III).

Lipoprotein-dependent removal of steroids from plasma membranes occurs by two different mechanisms: one aqueous diffusion mechanism and one specific apoprotein-mediated mechanism (for a review, see Ref. 24). Use of cyclodextrin represents the former type of mechanism. The addition of cyclodextrin alone to the macrophages resulted in a significant increase in total secretion of 27-oxygenated products and a marked decrease in the ratio of 3-hydroxy-5-cholestenoic acid to 27-hydroxycholesterol. This effect was observed also in the absence of cholesterol in the medium. It was recently shown that cyclodextrin causes a desorption of sterols from model membranes that is affected by the relative polarity of the sterols (25). Oxidized sterols were desorbed much faster than cholesterol. 27-Hydroxycholesterol can thus be expected to be efficiently transported from the cells by cyclodextrin. When cyclodextrin was added together with cholesterol, the marked increase in secretion of 27-oxygenated products may be due both to the acceptor function of cyclodextrin and to a facilitated transport of cholesterol to the sterol 27-hydroxylase system.

When the preloaded macrophages were exposed to lipid-deficient apolipoprotein A1, the efflux of cholesterol from the cells was considerably lower than in experiments with equimolar concentrations of HDL (Table V). However, the efflux of total labeled 27-oxygenated products was similar, and 3-hydroxy-5-cholestenoic acid was the predominant product excreted. This suggests that the lipid component of the lipoprotein particle is important for the efflux of 27-hydroxycholesterol. In the absence of the lipid, the initially formed 27-hydroxycholesterol is accumulated in the cells and is converted into 3-hydroxy-5-cholestenoic acid. (see Fig. 3).

It has been reported that unesterified 25-hydroxycholesterol in serum is associated with albumin to about the same extent as lipoproteins (26). In view of this, it may appear surprising that the addition of albumin to macrophages preloaded with labeled cholesterol did not increase the efflux of labeled 27-hydroxycholesterol from the cells and that only the flux of 3-hydroxy-5-cholestenoic acid was affected. Albumin seems to be less interactive with the oxysterols dissolved in the cell membranes than lipoproteins, at least under the conditions used here.

In a recent publication (27), it was suggested that excess plasma membrane cholesterol may stimulate a sensor to increase cholesterol influx. Being substrate-limited, mitochondrial sterol 27-hydroxylase could thereby be stimulated. It was speculated that the 27-hydroxylated product may feed back positively on the sensor, resulting in a further increase in influx of cholesterol through the plasma membrane. The possibility that 27-hydroxycholesterol could be secreted from the cells was never taken into account. It was shown here that exposure of the lung macrophages to cholesterol-containing fetal calf serum, isolated lipoproteins, or albumin caused a decreased intracellular concentration of 27-hydroxycholesterol in parallel with an increased efflux of 27-oxygenated steroids from the cell. Since the exposure of the cells to fetal calf serum is likely to increase the uptake of cholesterol in the cells, it is evident that the above hypothesis can not be valid for the macrophages studied here.

It is interesting that the most efficient elimination of cholesterol via sterol 27-hydroxylase was found in cells that are normally exposed to low concentrations of lipoproteins. Tissue macrophages are operative in extracellular fluid containing lipoproteins in concentrations considerably lower than those in the circulation. The dominant lipoprotein in HDL and LDL, apoA-I and apoB, respectively, is present in extracellular fluids in concentrations about 10-20% of those in plasma (28, 29). These concentrations are similar to those used in the present experiments (Table V).

Since endothelial cells are exposed to high concentrations of lipoproteins under physiological conditions, the present mechanism is probably less important for these cells. Since the total mass of endothelial cells is large, it is still possible that a significant part of the 27-hydroxycholesterol present in the circulation originates from endothelial cells.

The very high sterol 27-hydroxylase activity in human alveolar lung macrophages in relation to all other cells tested is notable and consistent with a specific function of this hydroxylase in the lung. There is one report in which granulomatous lesions were found in the lung of a patient with CTX (12). These lesions contained multinucleated giant cells and large foam cells.

The most important new finding here is that the present oxidative mechanism for elimination of cholesterol is an alternative and/or a complement to HDL-mediated reverse cholesterol transport in some mammalian cells. The relative importance of the two different strategies for elimination of cholesterol seems to be different in different species. The concentration of 27-oxygenated metabolites in the circulation of rabbits is only a small percentage of that in humans (19). However, sterol 27-hydroxylase is not the only oxidative enzyme involved in elimination of cholesterol in extrahepatic cells. Thus, it was recently shown that the human brain can eliminate part of its cholesterol through the blood-brain barrier by a mechanism involving a microsomal sterol 24-hydroxylase (30). The sterol 27- and 24-hydroxylases have in common that their primary products are side chain-hydroxylated. As judged from experiments with 25-hydroxycholesterol, side chain-hydroxylated cholesterol can be transported in lipid membranes orders of magnitude faster than cholesterol itself (31). A lipopilic acceptor is likely to be required for the transport of both 27-hydroxycholesterol and 24-hydroxycholesterol from extrahepatic cells. The most distinctive feature of the sterol 27-hydroxylase is thus that the end product of this enzyme, 3-hydroxy-5-cholestenoic acid, can be efficiently excreted from the cells also in the absence of a lipoprotein.