Effect of Coffee Lipids (Cafestol and Kahweol) on Regulation of Cholesterol Metabolism in HepG2 Cells
Abstract We studied the effect of the coffee diterpene alcohols, cafestol and kahweol, on cholesterol metabolism in HepG2 cells. Uptake of 125I-tyramine cellobiose–labeled LDL was decreased by 15% to 20% (P<.05) after 18 hours of preincubation with cafestol (20 μg/mL), whereas 25-hydroxycholesterol reduced uptake by 55% to 65% (P<.05). Degradation of LDL in the presence of cafestol was decreased by 20% to 30% (P<.05) under the same conditions. The effect of cafestol (20 μg/mL) on uptake and degradation of LDL was greatest (35% to 40%, P<.05) after 6 and 10 hours of preincubation, respectively. Furthermore, the effect of cafestol was also dependent on its concentration, and a significant decrease in the LDL uptake (19%) was observed at 10 μg/mL (P<.05). Specific binding of LDL was reduced by 17% (P<.05) and 60% (P<.05) after preincubation with cafestol (20 μg/mL) and 25-hydroxycholesterol (5 μg/mL) for 6 hours, respectively, compared with control cells. Analysis of LDL binding showed that cafestol reduced the number of binding sites for LDL on the cell surface (capacity) by 35% (P<.05). In contrast, no significant effect on the level of mRNA for the LDL receptor was observed after incubation with cafestol, whereas 25-hydroxycholesterol reduced the mRNA level for the LDL receptor by 40% to 50% (P<.05). A fusion gene construct consisting of a synthetic sterol regulatory element-1 (SRE-1) promoter for the human LDL receptor coupled to the reporter gene for chloramphenicol acetyltransferase (CAT) was transfected into HepG2 cells. No change was observed in CAT activity in SRE-1–transfected cells after incubation with cafestol, whereas 25-hydroxycholesterol reduced CAT activity by 30% to 40% (P<.05). Incorporation of [14C]acetate into unesterified cholesterol and 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase activity were unaffected in cells incubated with cafestol as well as the cafestol-kahweol mixture compared with control cells. Moreover, cafestol and the cafestol-kahweol mixture did not promote increased incorporation of radiolabeled [14C]oleic acid into cholesteryl esters after short-term incubation compared with control cells. On the other hand, 25-hydroxycholesterol caused a 70% to 90% reduction of cholesterol synthesis (P<.05) and HMG-CoA reductase activity (P<.05), decreased HMG-CoA reductase mRNA level by 70% to 80% (P<.05), and promoted a twofold increase in cholesterol esterification (P<.05). Finally, no effect of the coffee diterpenes on bile acid formation was observed. These results suggest that cafestol (and kahweol) may reduce the activity of hepatic LDL receptors and thereby cause extracellular accumulation of LDL.
- coffee lipids (cafestol and kahweol)
- LDL receptor
- low density lipoprotein
- cholesterol metabolism
- Received April 25, 1996.
- Accepted January 17, 1997.
In 1983, Thelle et al1 reported a significant, positive correlation between the amount of coffee ingested and serum cholesterol levels in a population-based study in Northern Norway. These findings have been confirmed in several subsequent studies.2 3 4 5 6 It was later shown that the method of brewing (boiling) was crucial for the hypercholesterolemic effect of coffee. In contrast, consumption of filtered coffee had little or no effect on serum cholesterol level.2 3 4 5 6 7 Zock et al8 prepared a lipid-rich fraction of boiled coffee by centrifugation and administered the coffee lipids to 10 volunteers for 6 weeks. This treatment resulted in raised serum LDL cholesterol and triacylglycerol levels. Recently, Weusten-Van der Wouw et al9 showed that a mixture of cafestol and kahweol, two coffee-specific diterpenes present in the lipid fraction of boiled coffee, raised serum concentration of cholesterol and triacylglycerol in healthy subjects. A daily intake of 73 mg cafestol and 58 mg kahweol for 6 weeks gave an increase of ≈35% in total plasma cholesterol, of which 75% was due to a rise in LDL cholesterol. Coffee oil without cafestol and kahweol caused no effects on serum cholesterol.9
The mechanisms by which these coffee lipids raise serum cholesterol are unknown, but a concurrent alteration in liver function enzymes is observed.9 Moreover, the cholesterol-raising effect of the diterpenes from coffee oil seems to be specific for human primates.10 A potential site of action for cafestol and kahweol may be the LDL receptor, which is involved in the endocytic process of apoB- and apoE-containing lipoproteins. One important way to regulate the cholesterol content of cells is via feedback repression of the gene for the LDL receptor.11 12 When cells are depleted of cholesterol, the LDL receptor gene is transcribed actively and LDL is cleared from plasma. However, when cholesterol accumulates within cells, the number of LDL receptors is downregulated. By regulating the number of cell-surface LDL receptors, cells are able to control the rate of entry of cholesterol, thereby assuring an optimal supply of the sterol.11 About two thirds of LDL removal from plasma is mediated by the LDL receptor. Furthermore, the liver is responsible for 70% to 90% of the LDL clearance from plasma.13
The aim of the present study was to examine the effect of cafestol and kahweol on cholesterol metabolism in HepG2 cells. These human hepatoma cells express several of the normal biochemical functions of liver parenchymal cells, including high-affinity receptors for the uptake and degradation of LDL.14 HepG2 cells are considered to be a reliable and useful model for studies on regulation of hepatic LDL catabolism.15 16 17 18 19 Furthermore, the effects of cafestol and kahweol on cholesterol metabolism have been compared with 25-hydroxycholesterol. This oxysterol, which suppresses both endogenous cholesterol synthesis and the number of LDL receptors in HepG2 cells, exerts similar effects on cellular cholesterol metabolism as LDL cholesterol.17 18
Cafestol (purity 99%) and a mixture of cafestol and kahweol (48% cafestol, 47% kahweol, and 5% isokahweol; purity 98%) were prepared as free alcohols from coffee oil by R. Bertholet at Nestlé Ltd Research Centre (preparation of cafestol, US patent 4692534, 1987). Simvastatin was supplied by Merck Sharp & Dohme. 25-Hydroxycholesterol, BSA (essentially fatty acid free), l-glutamine, EDTA, sodium acetate, acetyl CoA, diaphorase, resazurin, and 3α-hydroxysteroid dehydrogenase were purchased from Sigma Chemical Co. Na125I, [2-14C]acetic acid sodium salt (5 Ci/mol), [1-14C]oleic acid (58 Ci/mol), l-[3.4-3H]valine (62.6 Ci/mmol), d-threo-dichloroacetyl-[1.2-14C]CAT assay grade (56.8 Ci/mol), [3-32P]dCTP, and [3−14C]HMG-CoA (59 mCi/mmol) were obtained from DuPont NEN Products.
[1-14C]Acetic acid sodium salt (55 mCi/mmol) was obtained from Amersham International. A Ready-to-Go DNA labeling kit was delivered by Pharmacia. Dynabeads Direct kits and oligo-(dT)25 Dynabeads were purchased from Dynal. The cDNA probe for human HMG-CoA reductase (pHRed-102) was obtained from American Type Culture Collection, and the probe for human β-actin was provided by Clontech. The RNA ladder, DMEM, FCS, and penicillin/streptomycin were obtained from Gibco Life Technologies. The medium supplement ITS+ was provided by Collaborative Research. Trypsin was from Difco Laboratories. Plastic culture dishes, wells, and plastic flasks were delivered from either Costar or Falcon Labware. TLC plates (silica gel F 1500) were supplied by Schleicher & Schuell.
The compounds tested were dissolved either in DMSO or ethanol before dilution with cell-culture medium. The final concentration of DMSO or ethanol did not exceed 0.05% and 0.5% (vol/vol), respectively. Under these conditions, neither DMSO nor ethanol had a measurable effect on LDL metabolism, cholesterol synthesis, cholesterol esterification, or HMG-CoA reductase activity in the cells.
The established HepG2 cell line was purchased from American Type Culture Collection. Cell stocks were grown in 75-cm2 tissue-culture flasks containing DMEM with 10% (vol/vol) FCS (FCS-DMEM), supplemented with l-glutamine (2 mmol/L) and penicillin (50 IU/mL)/streptomycin (50 μg/mL). The cells were incubated in a humidified incubator (5% CO2) at 37°C. For subculture, the medium was removed and the cells were detached from the culture flasks with 0.25% trypsin (Difco Laboratories) in Ca2+-, Mg2+-free PBS containing 0.2 g/L EDTA. Culture medium with FCS was added to stop trypsinization. The cells were split at a 1:3 to 4 ratio every 5 to 6 days. During experiments, the cells were normally cultured in 60-mm plastic dishes (Costar) at a concentration of 5 to 6×105/mL in FCS-DMEM. Medium was exchanged every other day. When about 80% confluence was reached (after 4 to 5 days), the cells were used in experiments. The cells were cultured in DMEM alone or medium containing either FCS (10%), ITS+ (1%), or human LPDS (5 mg/mL), penicillin (50 IU/mL)/streptomycin (50 μg/mL), or gentamicin (see legends to figures and tables). For all the experiments described, cells were used between passages 76 and 91.
Cytotoxicity of the coffee diterpenes was studied by measurement of LDH leakage and measurement of incorporation of labeled valine into cell-associated acid-precipitable material (protein synthesis).
HepG2 cells were seeded out in 96-well microtiter plates at ≈7×103 per well in FCS-DMEM for 1 day. The cells were then incubated in FCS-DMEM for 24 hours with various concentrations of cafestol or the cafestol-kahweol mixture (cafestol/kahweol/isokahweol, 48%:47%:5%), dissolved in 0.5% ethanol, up to 50 μg/mL. LDH activity released from the cells was measured by a cytotoxicity detection kit (Boehringer Mannheim) using a positive control of 1% Triton X-100 added to the cells. Control cells were exposed to 0.5% ethanol.
HepG2 cells were cultured in 60-mm plastic dishes (Costar) at a concentration of 5 to 6×105/mL in FCS-DMEM for 4 to 5 days before they were used in experiments. The cells were then preincubated for 22 hours with FCS-DMEM and various concentrations of cafestol or the cafestol-kahweol mixture, dissolved in 0.5% ethanol, up to 50 μg/mL. Control cells were added to 0.5% ethanol. Thereafter, the cells were incubated for 2 hours in identical media supplemented with l-[3.4-3H]valine (5 μCi/mL, 0.8 mmol/L). After incubation, the cells were scraped off the culture dish by a rubber policeman, washed twice with PBS, and resuspended in 500 μL distilled water. Homogenized cell suspension (400 μL) was precipitated by 2 mL 20% (wt/vol) trichloroacetic acid, centrifuged (3000 rpm for 10 minutes), and washed twice during sonication with ice-cold trichloroacetic acid (10%, wt/vol). The acid precipitate was solubilized in 1.0 mL of 0.15 mol/L NaOH containing 70 mmol/L SDS and 10% Triton X-100, and an aliquot was counted by liquid scintillation (Packard Tri-Carb 1900 TR).
Samples (30 μL) were also taken for protein determination20 with BSA as a reference protein, using a microplate reader (Bio-Rad, model 3550).
Isolation and Labeling of LDL
LDL was isolated from freshly prepared human plasma by sequential ultracentrifugation (Centrikon T-2060/Sorwall Pro 80) at 10°C in the density range 1.019 to 1.063 g/mL in a TFT 70.38 rotor for 24 hours at 40 000 rpm.21 The final preparations were extensively dialyzed (Spectra/Por molecular-porous dialysis membrane) against PBS (0.15 mol/L NaCl; 20 mmol/L NaH2PO4) and 1 mmol/L EDTA, pH 7.4. One aliquot of LDL was labeled with 125I-tyramine cellobiose (125I-TC-LDL).22 More than 97% of the radioactivity was precipitated by 10% (wt/vol) trichloroacetic acid. The advantage of labeling a protein with radioiodinated tyramine cellobiose is that the degradation products are trapped in the organelles where degradation takes place. The final specific activity of 125I-TC-LDL was in the range 50 to 550 cpm/ng LDL protein (see legends to figures and tables). LDL was stored in the presence of EDTA under nitrogen at 4°C and used within 1 to 2 weeks.
The protein content was measured by the Bradford20 method with BSA as a reference protein.
Preparation of LPDS
LPDS was isolated from freshly prepared human serum by adding NaBr to a density of 1.25 g/mL. The samples were centrifuged at 40 000 rpm at 10°C for 48 hours. The newly isolated LPDS (bottom fraction) was extensively dialyzed against PBS (pH 7.4).
Uptake and Degradation of LDL
HepG2 cells were cultured in 60-mm plastic dishes (Costar) at a concentration of 5 to 6×105/mL in FCS-DMEM for 4 to 5 days. Twenty-four hours before commencing the experiments, the medium was replaced with DMEM supplemented with LPDS (LPDS-DMEM, 5 mg/mL) to maximize LDL receptor activity.23 The cells were then washed and preincubated for up to 24 hours with LPDS-DMEM containing either control (0.5% ethanol), cafestol (5 to 30 μg/mL), a cafestol-kahweol mixture (20 μg/mL), or 25-hydroxycholesterol (5 μg/mL). Thereafter, the cells were incubated for up to 24 hours with the respective coffee diterpenes or the oxysterol in the presence of 125I-TC-LDL (5 μg LDL protein per milliliter, 50 to 70 cpm/ng protein). Experiments using 40 times excess of unlabeled LDL were also performed to determine specific (saturable) uptake and degradation of LDL. After incubation, the cells were placed on ice, extensively washed with PBS (2×2 mL), and scraped off the culture dish by a rubber policeman. The cell suspension was homogenized with an ultrasound homogenizer and an aliquot was taken out for measurement of protein as previously described for LDL. The remaining homogenized cell fraction was added to 80 μL BSA (6%) and precipitated by an equal volume of ice-cold trichloroacetic acid (20%, wt/vol). After centrifugation (3000 rpm for 10 minutes), the samples were counted in a gamma counter (Packard Cobra II). Uptake and degradation of labeled LDL were calculated from the trichloroacetic acid insoluble (undegraded products) and soluble fractions, respectively.
Samples were taken for protein determination, using BSA as a reference protein.20
Binding Studies With LDL
LDL binding assays at 4°C were carried out according to Brown and Goldstein.24 HepG2 cells were cultured in DMEM containing 10% FCS for 4 to 5 days before they were used in experiments. The cells were pretreated with DMEM supplemented with LPDS (5 mg/mL) for 24 hours before fresh LPDS-DMEM containing either control (0.5% ethanol), cafestol (20 μg/mL), or 25-hydroxycholesterol (5 μg/mL) was provided for 18 hours. The cells were then chilled at 4°C and washed with cold DMEM. Fresh, cold LPDS-DMEM including 125I-TC-LDL (5 μg LDL protein per milliliter, 150 to 200 cpm/ng) was added, and the dishes were then incubated at 4°C for 2 hours. To determine nonspecific binding, identical dishes were incubated with the same concentration of 125I-TC-LDL in the presence of 200 μg/mL unlabeled LDL for 2 hours at 4°C. Specific binding represents the difference between the total and nonspecific binding. In addition, analysis of the saturable LDL binding was performed using a nonlinear regression data analysis program (Enzfitter), assuming one binding site per LDL particle. The cells were pretreated with LPDS-DMEM for 24 hours before fresh LPDS-DMEM containing either control (0.5% ethanol) or cafestol (20 μg/mL) was provided for 8 hours. The cells were then chilled at 4°C and washed with cold DMEM. Fresh, cold LPDS-DMEM with 125I-TC-LDL (450 to 550 cpm/ng) and increasing concentrations of unlabeled LDL was added, and the dishes were incubated at 4°C for 2 hours. To determine nonspecific binding, identical dishes were incubated with the same concentration of 125I-TC-LDL in the presence of 750 μg/mL unlabeled LDL for 2 hours at 4°C.
Northern Blot Analysis
HepG2 cells were grown in 75 cm2 flasks (Costar) at a concentration of 2×105/mL in DMEM containing 10% FCS for 2 days. The cells were then pretreated with DMEM supplemented with LPDS (5 mg/mL) for 24 hours before fresh LPDS-DMEM containing either control (0.5% ethanol), cafestol (20 μg/mL), or 25-hydroxycholesterol (5 μg/mL) was given for 11 and 24 hours, respectively. After incubation, the cells were placed on ice, washed twice with ice-cold PBS, scraped off the culture dish by a rubber policeman, and homogenized with an ultrasound homogenizer. mRNA [poly (A+) RNA] was isolated from cells using Dynabeads mRNA Direct kit. Six micrograms poly (A+) RNA was separated on a 1% agarose gel containing 6.7% formaldehyde and transferred to a nylon membrane. cDNA probes for the human LDL receptor (pSP 15),25 human HMG-CoA reductase (pHRed-102),26 and human β-actin were labeled with [32P]dCTP, using the Ready-to-Go labeling kit. The prehybridization, hybridization, and washing was carried out at 65°C according to Church and Gilbert.27 The size of the mRNA was determined with references to RNA standards visualized by methylene blue staining after they were cut off the nylon membrane before hybridization. Finally, the signals were analyzed by PhosphorImager (Molecular Dynamics) and calibrated against β-actin as an internal standard. The sizes of the mRNA were 5.3 kb, 4.2 kb, and 2.0 kb for the LDL receptor, HMG-CoA reductase, and β-actin, respectively.
HepG2 cells were plated at 7.5×103/cm2 in 100-mm dishes and transfected 24 hours postplating with 15 μg of DNA (a synthetic, active SRE-1 promoter) kindly provided by Dr J.L. Goldstein and Dr M.S. Brown, University of Texas Health Science Center, Dallas, Tex.28 On the day of transfection, the cells were fed complete medium 2 hours before precipitation. Fifteen micrograms of test plasmid DNA, 3 μg of SR α-lacZ, 3 μg of Bluescript KS+, and 50 μL of CaCl2 (2.5 mol/L) in 430 μL of sterile water were gently mixed with 500 μL 2× HEPES-buffered saline (50 mmol/L HEPES, 280 mmol/L NaCl, 1.5 mmol/L Na2HPO4×2H2O) at pH 6.95 to 6.98. The precipitate was allowed to form over 20 minutes at room temperature, and thereafter the precipitated solution was added dropwise to each monolayer. The cells were incubated for 20 hours with the DNA and then washed twice with 10 mL of warm PBS (37°C) and refed with 10 mL of DMEM containing calf LPDS (5 mg protein per milliliter) in the presence of control (0.1% ethanol), cafestol (10 to 20 μg/mL), or 25-hydroxycholesterol (5 μg/mL). After incubation for 11 and 24 hours, the cells were harvested for measurement of β-galactosidase29 and CAT activity.30 The transfected cells were washed twice with PBS, scraped into 180 μL of 0.25 mol/L Tris-HCl/0.05 mmol/L EDTA at pH 7.5, lysed by freezing and thawing six times, and centrifuged at 12 000 rpm for 5 minutes at 4°C.
Measurement of β-Galactosidase
The supernatant (40 μL) was incubated at 37°C for 30 to 60 minutes with 140 μL of o-nitrophenyl-β-d-galactopyranoside (4 mg/mL) and 700 μL of 60 mmol/L NaHPO4, 40 mmol/L NaH2PO4, 10 mmol/L KCl, 1 mmol/L MgSO4×7H2O. The reactions were stopped with 400 μL of 1 mol/L Na2CO3, and the amount of o-nitrophenol formed was measured spectrophotometrically at 420 nm. The β-galactosidase activity was a measurement for the transfection efficiency.
Measurement of CAT Activity
An aliquot of the supernatant was incubated for 1 hour at 37°C in a final volume of 200 μL containing 4 μL of [14C]chloramphenicol and 25 μL of acetyl-CoA (2.6 mg/mL). The aliquot was calculated on the basis of β-galactosidase activity to calibrate for unequal transfection efficiency. Ice-cold ethyl acetate (1 mL) was then added and the mixture centrifuged at 12 000 rpm for 5 minutes at 4°C. The organic phase was dried and the residual extract redissolved in 15 μL ethyl acetate. The reaction products were separated by TLC using chloroform/methanol (95:5, vol/vol) as developing solvent. After autoradiography (Fuji medical X-ray film), the radioactive spots were cut out and counted in a scintillation counter.
Measurements of Cholesterol Synthesis and Cholesteryl Ester Formation
Cholesterol synthesis from [14C]acetate was measured as previously described by Rustan et al.31 The cells were preincubated with LPDS-DMEM containing either control (0.5% ethanol), cafestol (20 μg/mL), or 25-hydroxycholesterol (5 μg/mL) for the indicated time. The medium was then replaced with new, identical medium containing [2-14C]acetate (1 μCi/mL, 100 μmol/L) and further incubated for various time periods at 37°C. After incubation, the cells were scraped off the culture dishes into PBS, and lipids from cells and media were extracted with chloroform/methanol (2:1, vol/vol). The homogenized cell fraction was mixed with 20 vol chloroform/methanol (2:1, vol/vol).32 Four volumes of a 0.9% NaCl solution (pH 2) was added and the mixture allowed to separate into two phases. The organic phase was dried under nitrogen at 40°C. Medium without cellular debris was added to 4 vol chloroform/methanol (2:1, vol/vol) and 2% FCS as unlabeled carrier for the lipids. The water phase of the medium extract was reextracted once with 4 vol chloroform/methanol (2:1, vol/vol), and the combined organic phases were further treated in the same way as for the cells. The residual lipid extract was redissolved in 200 μL hexane and separated by TLC using hexane/diethyl ether/acetic acid (65:35:1, vol/vol/vol) as developing solvent. The various lipids were identified by iodine, the TLC foils cut, and the samples counted by liquid scintillation.
Cholesteryl Ester Formation
The cells were preincubated with FCS-DMEM containing either control (0.5% ethanol), cafestol (20 μg/mL), cafestol-kahweol mixture (20 μg/mL), or 25-hydroxycholesterol (5 μg/mL) for 18 hours. The medium was then replaced with new, identical FCS-DMEM containing [1-14C]oleic acid (0.25 μCi/mL, 200 μmol/L) bound to albumin (ratio 2.5:1), and further incubated for 1 to 24 hours at 37°C. The cells were harvested before protein content was measured and the lipids extracted, separated by TLC, and counted by liquid scintillation.
HMG-CoA Reductase Activity
A modification of the method of Nagata et al34 was used for the determination of HMG-CoA reductase activity. HepG2 cells were seeded out in six-well plates at a density of 3×105 cells per well in DMEM with 10% FCS for 24 hours. The cells were washed and incubated for a further 48 hours in DMEM containing 1% ITS+ (ITS-DMEM). The cells were then incubated in ITS-DMEM for 24 hours with cafestol, a cafestol-kahweol mixture, or 25-hydroxycholesterol, dissolved in 0.05% DMSO. The cell extract was prepared by suspending the pellet in 200 μL of buffer containing 0.1 mol/L phosphate buffer (pH 7.5), 0.2 mol/L KCl, 5 mmol/L EDTA, and 0.25% Brij 96 and sonicating for 2 seconds at 0°C. The cell homogenate was centrifuged at 4°C and an aliquot of the supernatant was taken for determination of HMG-CoA reductase activity.
Bile Acid Formation
Bile acid formation in the cells was measured by a fluorometric method obtained from Dr Z.F. Stephan, Ciba-Geigy Corporation, Summit, NJ. HepG2 cells grown for 4 to 5 days in FCS-DMEM were washed twice with DMEM and incubated for 11 hours in DMEM with control (0.5% ethanol), cafestol (10 to 30 μg/mL), or cafestol-kahweol (20 μg/mL). After incubation, an aliquot from the medium was taken for determination of bile acids. Briefly, 50 μL medium was added to 1.0 mL 0.05 mol/L Tris-HCl buffer (pH 9.5) and heated for 30 minutes at 67°C. The samples were cooled at room temperature and added to 0.5 mL of a reagent mixture containing 0.1 mL 3α-hydroxysteroid dehydrogenase (0.1 U/mL in distilled water), 0.3 mL NAD/diaphorase (9 mmol/L NAD/0.5 U/mL diaphorase, dissolved in 0.065 mol/L NaH2PO4 buffer, pH 7.4), and 0.1 mL resazurin (100 μmol/L in distilled water). The samples were left for 1 hour and fluorescence was measured (Perkin-Elmer LS 50) at 580 nm with excitation wavelength set at 560 nm (slit-width 10 nm). Bile acid concentration was measured against sodium taurocholate (0.5 to 5 μg/mL) dissolved in methanol. Methanol was used as blank. Cell samples were also taken for determination of protein.
Presentation of Data
All values are reported as mean±SD from several separate experiments (with duplicate or triplicate cell-culture dishes for each value determined) unless otherwise indicated. Comparisons of different treatments were statistically evaluated by Student’s t test (two-tailed). A value of P<.05 was taken to be statistically significant.
LDH Leakage and Protein Synthesis
The cytotoxicity of cafestol and the cafestol-kahweol mixture was determined by measuring LDH release from cells and by synthesis of protein after incubation for 24 hours (Fig 1⇓). Cafestol or the cafestol-kahweol mixture did not markedly affect LDH leakage and protein synthesis at concentrations up to 30 μg/mL in medium with serum (FCS) (Fig 1⇓), although significant changes indicating reduced cytotoxicity were observed for some concentrations. LDH leakage was increased by 18% in the presence of 50 μg/mL cafestol-kahweol mixture (Fig 1⇓). When the cells were incubated in DMEM without serum, a 10% to 20% (n=3; P<.05) increase in LDH leakage at concentrations from 5 to 30 μg/mL of the coffee lipids was seen (data not shown). It should be noted that in all our experiments except for HMG-CoA reductase activity and bile acid formation, serum or LPDS was present during the incubations. Moreover, no effect of cafestol and the cafestol-kahweol mixture was observed on the amount of cellular protein at incubations up to 30 μg/mL for 24 hours in culture medium with and without serum (data not shown).
Uptake and Degradation of LDL
When the cells were incubated with cafestol (20 μg/mL) for 18 hours and further incubated for various times with radiolabeled LDL, the uptake of LDL decreased by 15% to 20% compared with control cells (Fig 2⇓). Incubation of the cells with the cafestol-kahweol mixture (20 μg/mL) for 18 hours and thereafter for 6 hours in the presence of LDL reduced the uptake of LDL by 24±1% (n=3; P<.05). Under the same conditions, 25-hydroxycholesterol (5 μg/mL) decreased the LDL uptake by 55% to 65% (Fig 2⇓). Moreover, simvastatin (2 μmol/L), an HMG-CoA reductase inhibitor, increased uptake of LDL by 54±20% (n=4; P<.05) compared with controls (after 6 hours of preincubation with the drug and 5 hours of incubation with labeled LDL). These findings confirmed the responsiveness of the cells to agents known to regulate hepatic LDL metabolism.14 17 18 Degradation of LDL was reduced by 20% to 30% after incubation with cafestol, whereas 25-hydroxycholesterol decreased degradation of LDL by 60% to 65% (Fig 2⇓).
The uptake of LDL was dependent on the preincubation time with cafestol (20 μg/mL) (Fig 3⇓). The greatest reduction in LDL uptake of 35% was observed after 6 and 10 hours of preincubation. Under the same conditions, degradation was reduced by ≈40% (37±8%, P<.05, and 43±9%, P<.05, after 6 and 10 hours of preincubation) compared with control cells. The uptake of LDL was decreased by 70% after 6 hours’ preincubation of the cells with 25-hydroxycholesterol (Fig 3⇓).
The effect of cafestol on uptake and degradation of radiolabeled LDL in HepG2 cells was also concentration dependent (Fig 4⇓). The uptake of LDL was significantly decreased by 19% compared with control cells after incubation with 10 μg/mL cafestol. In the presence of 20 to 30 μg/mL cafestol, the uptake of LDL was reduced by 30% to 50%. At 5 μg/mL cafestol, the degradation of LDL was significantly decreased by 11%. At 10 to 30 μg/mL cafestol, degradation of LDL was reduced by 30% to 70% compared with control cells. 25-Hydroxycholesterol (5 μg/mL) decreased the uptake and degradation of LDL by 60% to 65% (Fig 4⇓).
The uptake of LDL was also measured with excess unlabeled LDL after 6 hours of preincubation and 5 hours of subsequent incubation in the presence of radiolabeled LDL to determine the effect of cafestol on the specific (saturable) LDL uptake. Under these conditions, cafestol (20 μg/mL) reduced specific uptake by 24% (76±21% of control, n=4; P<.05), whereas 25-hydroxycholesterol (5 μg/mL) decreased specific uptake of LDL by 70% (30±5% of control, n=3; P<.05).
Measurement of LDL Binding
To examine whether the cafestol-mediated decrease in uptake and degradation of radiolabeled LDL was due to a decreased number of LDL receptors on the cell surface, binding experiments at 4°C were performed. Cafestol significantly reduced specific binding of radiolabeled LDL by 17%, whereas 25-hydroxycholesterol reduced specific binding of LDL by 60% (Table 1⇓). We also performed an analysis of specific LDL receptor binding after incubation of HepG2 cells for 8 hours in the absence or presence of cafestol (20 μg/mL) (Fig 5⇓). Cafestol reduced the number of binding sites for LDL on the cell surface (capacity) by 35% (P<.05) compared with control cells. The binding affinity (Kd) was unchanged after exposure of the cells to cafestol. A Scatchard plot of the LDL binding is also shown (Fig 5⇓).
Expression of LDL Receptor mRNA
HepG2 cells were incubated with cafestol to examine whether the regulation of LDL receptor expression may occur at the transcriptional level. Incubation of HepG2 cells with 20 μg/mL cafestol promoted no significant change in LDL receptor mRNA compared with control cells (108±27% and 114±48% of control after 11 and 24 hours, n=5). In contrast, incubation of the cells with 25-hydroxycholesterol (5 μg/mL) decreased LDL receptor mRNA expression markedly (57±33% and 51±33% of control after 11 and 24 hours, n=5; P<.05).
Measurement of CAT Activity in SRE-1 Promoter–Transfected Cells
HepG2 cells were transfected with the active promoter region of the SRE-1 coupled to the gene for CAT to study whether cafestol regulated LDL receptor expression at the transcriptional level by interacting with the SRE-1 promoter. SRE-1 is a conditional positive element that enhances LDL receptor gene transcription in the absence of sterols.28 35 36 37 When the transfected cells were incubated with cafestol (10 to 20 μg/mL) for 11 and 24 hours, no significant changes in CAT activity were observed (Fig 6⇓). On the other hand, 25-hydroxycholesterol (5 μg/mL) reduced CAT activity by ≈30% after 11 and 24 hours of treatment (Fig 6⇓).
Cells preincubated for 30 minutes with cafestol at a concentration of 20 μg/mL did not reduce incorporation of labeled acetate into cholesterol during a 6-hour incubation period (Fig 7⇓). In comparison, 25-hydroxycholesterol (5 μg/mL) decreased incorporation of labeled acetate into cholesterol by about 80% after 2 hours of incubation (Fig 7⇓). In accordance with the maximum effect of cafestol observed in the LDL experiments after 6 to 10 hours of preincubation, a 6-hour preincubation time was also chosen. Under these conditions, cafestol still did not reduce cholesterol synthesis from acetate, which was markedly reduced by 25-hydroxycholesterol (Fig 7⇓). Exposure of HepG2 cells to a mixture of cafestol and kahweol up to a concentration of 20 μg/mL for 18 hours had no effect on cholesterol synthesis (data not shown). Secretion of unesterified cholesterol was low (range, 0.2 to 0.4 nmol/mg cell protein), and there was no difference between control and cafestol-treated cells after 6 hours of preincubation (data not shown).
Measurement of HMG-CoA Reductase Activity
The cafestol-kahweol mixture up to a concentration of 20 μg/mL had no effect on the activity of HMG-CoA reductase (Table 2⇓). Moreover, increasing the exposure time of the cells to 5 μg/mL cafestol-kahweol mixture from 2 to 48 hours did not affect the enzyme activity (data not shown). Similar results were obtained when the cells were incubated with cafestol (5 μg/mL) for 24 hours (Table 2⇓). In contrast, 25-hydroxycholesterol (2 μg/mL) markedly decreased HMG-CoA reductase activity (Table 2⇓).
Expression of HMG-CoA Reductase mRNA
Incubation of HepG2 cells with 20 μg/mL cafestol for 11 and 24 hours promoted a slight increase of 17% to 35% in HMG-CoA reductase mRNA level compared with control cells (135±19% and 117±31% of control after 11 and 24 hours, n=5; P<.05). In contrast, incubation of the cells with 25-hydroxycholesterol (5 μg/mL) caused a 65% to 80% decrease of HMG-CoA reductase mRNA expression (22±18% and 35±20% of control after 11 and 24 hours, n=5; P<.05). Moreover, incubation of the cells with the cafestol-kahweol mixture (2 to 10 μg/mL) did not change the expression of HMG-CoA reductase mRNA (data not shown).
Cholesteryl Ester Formation
Preincubation of the cells with cafestol (20 μg/mL) or the mixture of cafestol and kahweol (20 μg/mL) for 18 hours and thereafter incubation with labeled oleic acid for different time periods did not affect cholesterol ester synthesis compared with controls (Fig 8⇓). However, incorporation of radiolabeled oleic acid into cholesteryl ester was increased 2.1-fold in the presence of 25-hydroxycholesterol (5 μg/mL) after 6 hours of incubation (Fig 8⇓). Secretion of cholesteryl ester was low (range, 0.05 to 0.25 nmol/mg cell protein), and there was no difference between control and cafestol-treated cells (data not shown). Cholesteryl ester formation was also examined after 6 hours of preincubation with cafestol (20 μg/mL), providing similar results (data not shown).
Bile Acid Formation
The effect of the coffee lipids on bile acid formation was also studied. A slight nonsignificant decrease of 5% to 10% in bile acid formation was seen after incubation with various concentrations of cafestol and the cafestol-kahweol mixture (20 μg/mL) for 11 hours (Table 3⇓). Simvastatin (2 μmol/L) decreased bile acid production by 45% after 24 hours of incubation compared with control (data not shown).
The present study shows that pure cafestol and a cafestol-kahweol mixture decreased the binding, uptake, and degradation of labeled LDL in HepG2 cells (Table 1⇑ and Figs 2⇑, 3⇑, 4⇑, and 5⇑). Our data also show that the decrease in uptake and degradation was dependent on concentration (Fig 4⇑) and preincubation time with cafestol (Fig 3⇑). 25-Hydroxycholesterol (5 μg/mL) decreased uptake and degradation of LDL considerably more than cafestol. Moreover, cafestol reduced the number of binding sites without affecting affinity to the LDL receptors (Fig 5⇑). The decrease in LDL binding, uptake, and degradation was probably not due to cytotoxic effects of the coffee diterpenes, since the diterpenes did not markedly affect protein synthesis or LDH leakage from the cells (Fig 1⇑).
No significant effect was noted on either mRNA expression for the LDL receptor or CAT activity in HepG2 cells transfected with SRE-1 (Fig 6⇑). It has been reported that small changes in hepatic LDL receptor gene expression (2% to 4% reconstitution) may cause 30% reduction in serum cholesterol in Watanabe heritable hyperlipidemic rabbits injected with LDL receptor–transfected hepatocytes.38 Thus, it is possible that although we are unable to detect significant differences in LDL receptor mRNA level, there could still be differences in LDL receptor transcription. Another explanation for our observations might be that cafestol (and kahweol) may regulate LDL metabolism through a different mechanism compared with oxysterols such as 25-hydroxycholesterol. The lack of decrease in the expression of mRNA levels for the LDL receptor observed after incubation with cafestol suggests that cafestol did not execute its effect directly on the gene for the LDL receptor. In contrast, 25-hydroxycholesterol treatment decreased the expression of LDL receptor mRNA by 40% to 50%. This is in accordance with results showing that 25-hydroxycholesterol regulates the LDL receptor mRNA by changing the rate of gene transcription in HepG2 cells.39 Our observation is also in agreement with results from human skin fibroblasts, in which no significant effects of cafestol were observed on the level of mRNA for the LDL receptor, whereas a 40% decrease on the uptake and degradation of radiolabeled LDL was observed (1997, unpublished data, 1996).
Two essential proteins regulating cellular cholesterol homeostasis, HMG-CoA synthase and LDL receptor, are known to contain SRE-1 in the promoter region of their genes.28 37 This element is able to bind SREBP-1, which regulates the transcription of these genes.35 SREBP-1 is anchored to the cholesterol-poor endoplasmic reticulum as a precursor. When the cells are depleted of sterols, this membrane-bound precursor is proteolytically cleaved and translocated to the nucleus, where it binds to SRE-1 and thereby promotes transcription of the genes. It is suggested that the cholesterol-mediated changes in the endoplasmic reticulum membrane may induce cleavage of the active SREBP-1.36 37 The SRE-1–transfected HepG2 cells incubated with 25-hydroxycholesterol showed 30% to 40% reduced CAT activity, which is in agreement with former studies,35 whereas SRE-1–transfected cells incubated with cafestol showed no significant reduction in CAT activity compared with control cells (Fig 6⇑). These data, in combination with our data on the mRNA level for the LDL receptor, suggest that cafestol does not directly modulate the transcription of the LDL receptor gene, whereas 25-hydroxycholesterol regulates the transcription of the LDL receptor gene39 by inhibition of the cleavage of the SREBP-1 precursor.
The reduced binding, uptake, and degradation of radiolabeled LDL caused by cafestol could be due to a reduced transport of newly synthesized LDL receptors to the cell surface, an intracellular accumulation of newly synthesized LDL receptors, or decreased recycling of the LDL receptors.40 41 42 It has been observed that HMG-CoA reductase inhibitors enhance hepatic uptake of LDL, despite a constant steady state level of hepatic LDL receptor protein, suggesting increased recycling of LDL receptors.42 The existence of posttranscriptional mechanisms for sterol-mediated reduction of LDL receptor activity has also been reported.43 Another mechanism by which cafestol might regulate LDL receptor activity could be via phosphorylation/dephosphorylation. A phosphorylation/dephosphorylation mechanism for regulation of the scavenger receptor has recently been described.44 45
In contrast to HepG2 cells and human skin fibroblasts, incubation of CaCo-2 cells (a human colon cancer cell line) with the coffee diterpenes enhanced the uptake and degradation of LDL by 50%, whereas 25-hydroxycholesterol showed a 30% decrease in LDL metabolism.46 A 20% increase in the uptake of radiolabeled LDL was also seen in P19 cells (a murine embryonic carcinoma cell line) and J774 cells (a murine macrophage-like cell line), whereas 25-hydroxycholesterol decreased uptake of labeled LDL in both these cell lines.46 This observation indicates that the coffee diterpenes themselves or metabolites thereof may act differently in various cell lines. As far as we know, cafestol and kahweol are the only dietary lipids that have such different potential in the regulation of cholesterol metabolism in various cell types.
Results from this study also demonstrate that the coffee diterpenes affected neither cholesterol synthesis nor the activity or the mRNA level of HMG-CoA reductase, unlike 25-hydroxycholesterol (Fig 7⇑ and Table 2⇑), although a significant increase in HMG-CoA reductase mRNA level was observed after 11 hours of incubation with cafestol. HMG-CoA reductase is a highly regulated enzyme, with its activity reflected both by the amount (synthesis and degradation) and activation (phosphorylation and dephosphorylation) of the enzyme.26 47 48 49 It has been shown that the rate of synthesis of the HMG-CoA reductase protein is controlled by oxysterols acting on transcriptional as well as translational levels.47 48
We observed no effect of cafestol and the cafestol-kahweol mixture on cholesterol esterification, whereas 25-hydroxycholesterol promoted a twofold increase in cholesteryl ester formation (Fig 8⇑). The effect of 25-hydroxycholesterol is most likely due to stimulation of the activity of acyl-CoA:cholesterol acyltransferase or a consequence of increased substrate availability.50 51 52
Our present results are different from what was found in human skin fibroblasts (Halvorsen B, Ranhelm T, Nenseter MS, Huggett AC, and Drevon CA, unpublished data, 1997) and CaCo-2 cells.46 In fibroblasts, cafestol promoted a reduced synthesis of cholesterol and increased cholesterol esterification, whereas in CaCo-2 cells, cafestol caused increased synthesis of cholesterol and reduced esterification. These observed differences in cholesterol metabolism may reflect artificial changes due to cell-culture conditions, different response and/or formation of metabolites from the coffee diterpenes, or the fact that liver cells, mucosal cells, and fibroblasts have different metabolic potentials.
In conclusion, our studies show that the coffee diterpenes cafestol and kahweol significantly decrease binding, uptake, and degradation of LDL in human hepatoma cells. The mRNA level for the LDL receptor was unaffected after incubation of HepG2 cells with cafestol, and cells transfected with a synthetic promoter SRE-1 for the LDL receptor and HMG-CoA synthase were insensitive to cafestol. These findings suggest that the cholesterol-raising effect observed in human studies after intake of cafestol and kahweol may involve the downregulation of the hepatic LDL receptors, possibly by posttranscriptional mechanisms.
Selected Abbreviations and Acronyms
|DMEM||=||Dulbecco’s modified Eagle’s medium|
|FCS||=||fetal calf serum|
|HMG-CoA||=||3-hydroxy-3-methylglutaryl coenzyme A|
|SRE-1||=||sterol regulatory element-1|
|SREBP-1||=||sterol regulatory element binding protein-1|
|125I-TC-LDL||=||125I-tyramine cellobiose–labeled LDL|
This work was supported by grants from Freia Chocolade Fabriks Medisinske fond, Anders Jahre Foundation, Nansen Foundation, Novo Nordisk Foundation, the Norwegian Council on Cardiovascular Disease, the Norwegian Research Council, Norwegian Medicinal Depot, and Nestlé Ltd Research Centre, Lausanne, Switzerland. Drs Joseph L. Goldstein and Michael S. Brown kindly provided the SRE-1 promoter and Dr Helge Tolleshaug provided the tyramine-cellobiose adduct. We thank Dr Zouhair F. Stephan for the bile acid method. We are grateful to Mari-Ann Baltzersen, Anne Randi Alvestad, Joril Lieberg, Kristine Dranger, Linda van den Bosch, Martijn van der Gaag, Claudine Bezencon, Arlette Chiesa, and Marie-Claire Haymoz for excellent technical assistance.
Bønaa K, Arnesen E, Thelle DS, Førde OH. Coffee and cholesterol: is it all in the brewing?: the Tromsø study. BMJ. 1988;297:1103-1104.
van Dusseldorp M, Katan MB, van Vliet T, Demacker PN, Stalenhoef AF. Cholesterol-raising factor from boiled coffee does not pass a paper filter. Arterioscler Thromb. 1991;11:586-593.
Weusten-Van der Wouw MP, Katan MB, Viani R, Huggett AC, Liardon R, Lund-Larsen PG, Thelle DS, Ahola I, Aro A, Meyboom S, Beynen AC. Identity of the cholesterol-raising factor from boiled coffee and its effects on liver function enzymes. J Lipid Res. 1994;35:721-733. Published erratum in J Lipid Res. 1994;35:1510.
Terpstra AH, Katan MB, Weusten-van der Wouw MP, Nicolosi RJ, Beynen AC. Coffee oil consumption does not affect serum cholesterol in rhesus and cebus monkeys. J Nutr. 1995;125:2301-2306.
Javitt NB. Hep G2 cells as a resource for metabolic studies: lipoprotein, cholesterol, and bile acids. FASEB J. 1990;4:161-168.
Carlson TL, Kottke BA. Effect of 25-hydroxycholesterol and bile acids on the regulation of cholesterol metabolism in Hep G2 cells. Biochem J. 1989;264:241-247.
Dashti N. The effect of low density lipoproteins, cholesterol, and 25-hydroxycholesterol on apolipoprotein B gene expression in HepG2 cells. J Biol Chem. 1992;267:7160-7169.
Havel RJ, Eder HA, Bragdon JH. The distribution and chemical composition of ultracentrifugally separated lipoproteins in human serum. J Clin Invest. 1955;34:1345-1353.
Pittman RC, Carew TE, Glass CK, Green SR, Taylor C Jr, Attie AD. A radioiodinated, intracellularly trapped ligand for determining the sites of plasma protein degradation in vivo. Biochem J. 1983;212:791-800.
Goldstein JL, Dana SE, Brown MS. Esterification of low density lipoprotein cholesterol in human fibroblasts and its absence in homozygous familial hypercholesterolemia. Proc Natl Acad Sci U S A. 1974;71:4288-4292.
Peacock SL, Bates MP, Russell DW, Brown MS, Goldstein JL. Human low density lipoprotein receptor expressed in Xenopus oocytes: conserved signals for O-linked glycosylation and receptor-mediated endocytosis. J Biol Chem. 1988;263:7838-7845.
Luskey KL, Stevens B. Human 3-hydroxy-3-methylglutaryl coenzyme A reductase: conserved domains responsible for catalytic activity and sterol-regulated degradation. J Biol Chem. 1985;260:10271-10277.
Church GM, Gilbert W. Genomic sequencing. Proc Natl Acad Sci U S A. 1984;81:1991-1995.
Lee F, Hall CV, Ringold GM, Dobson DE, Luh J, Jacob PE. Functional analysis of the steroid hormone control region of mouse mammary tumor virus. Nucleic Acids Res. 1984;12:4191-4206.
Gorman CM, Moffat LF, Howard BH. Recombinant genomes which express chloramphenicol acetyltransferase in mammalian cells. Mol Cell Biol. 1982;2:1044-1051.
Rustan AC, Nossen JO, Osmundsen H, Drevon CA. Eicosapentaenoic acid inhibits cholesterol esterification in cultured parenchymal cells and isolated microsomes from rat liver. J Biol Chem. 1988;263:8126-8132.
Folch J, Lees M, Sloane Stanley GH. A simple methods for the isolation and purification of total lipids from animal tissue. J Biol Chem. 1957;226:497-509.
Brown MS, Faust JR, Goldstein JL. Induction of 3-hydroxy-3-methylglutaryl coenzyme A reductase activity in human fibroblasts incubated with compactin (ML-236B), a competitive inhibitor of the reductase. J Biol Chem. 1978;253:1121-1128.
Wang X, Briggs MR, Hua X, Yokoyama C, Goldstein JL, Brown MS. Nuclear protein that binds sterol regulatory element of low density lipoprotein receptor promoter, II: purification and characterization. J Biol Chem. 1993;268:14497-14504.
Chowdhury JR, Grossman M, Gupta S, Chowdhury NR, Baker JR, Wilson JM. Long-term improvement of hypercholesterolemia after ex vivo gene therapy in LDLR-deficient rabbits. Science. 1991;254:1802-1805.
Lodish HF, Kong N. Glucose removal from N-linked oligosaccharides is required for efficient maturation of certain secretory glycoproteins from the rough endoplasmic reticulum to the Golgi complex. J Cell Biol. 1984;98:1720-1729.
Sharkey MF, Miyanohara A, Elam RL, Friedmann T, Witztum JL. Post-transcriptional regulation of retroviral vector-transduced low density lipoprotein receptor activity. J Lipid Res. 1990;31:2167-2178.
Fong LG. Posttranslational regulation of macrophage scavenger receptor by protein phosphorylation. Circulation. 1994;90(suppl I):I-189. Abstract.
Fong LG. Modulation of macrophage scavenger receptor transport by protein phosphorylation. J Lipid Res. 1996;37:574-587.
Ranheim T, Halvorsen B, Huggett AC, Blomhoff R, Drevon CA. Effect of a coffee lipid (cafestol) on regulation of lipid metabolism in CaCo-2 cells. J Lipid Res. 1995;36:2079-2089.
Panini SR, Sexton RC, Rudney H. Regulation of 3-hydroxy-3-methylglutaryl coenzyme A reductase by oxysterol by-products of cholesterol biosynthesis: possible mediators of low density lipoprotein action. J Biol Chem. 1984;259:7767-7771.
Faust JR, Luskey KL, Chin DJ, Goldstein JL, Brown MS. Regulation of synthesis and degradation of 3-hydroxy-3-methylglutaryl-coenzyme A reductase by low density lipoprotein and 25-hydroxycholesterol in UT-1 cells. Proc Natl Acad Sci U S A. 1982;79:5205-5209.
Omkumar RV, Rodwell VW. Phosphorylation of Ser871 impairs the function of His865 of Syrian hamster 3-hydroxy-3-methylglutaryl-CoA reductase. J Biol Chem. 1994;269:16862-16866.
Brown MS, Dana SE, Goldstein JL. Cholesterol ester formation in cultured human fibroblasts: stimulation by oxygenated sterols. J Biol Chem. 1975;250:4025-4027.
Drevon CA, Weinstein DB, Steinberg D. Regulation of cholesterol esterification and biosynthesis in monolayer cultures of normal adult rat hepatocytes. J Biol Chem. 1980;255:9128-9137.