INTRODUCTION

SR-12813 (tetra-ethyl 2-(3, 5-di-tert-butyl-4-hydroxyphenyl)ethenyl-1,1-bisphosphonate; see Fig. 1) and related compounds are potent hypocholesterolemic drugs in mouse, rat, hamster, rabbit, dog, and monkey (1). To explain these cholesterol-lowering effects, studies were initiated in the human hepatoma cell line Hep G2. These cells are particularly suitable to study effects on lipid metabolism (2, 3, 4). Our initial studies in Hep G2 cells showed that SR-12813 inhibited cholesterol biosynthesis at the level of 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA)¹ reductase, the key regulatory enzyme in the cholesterol biosynthetic pathway (5). The regulation of cellular HMG-CoA reductase activity is extremely complicated. The basis for this is that mevalonate, the product of the enzymatic reaction, is not only the precursor for the biosynthesis of cholesterol but also for a variety of other important products like dolichols, ubiquinones, and isoprenoids, which are required for growth and survival of the cell. Feedback mechanisms for the cholesterol biosynthetic pathway impinge on HMG-CoA reductase expression. HMG-CoA reductase is regulated on a transcriptional level by (oxy)sterols (6, 7). Post-transcriptionally, feedback regulation on HMG-CoA reductase synthesis and degradation is provided by isoprenoid metabolites (7, 8, 9, 10, 11), by oxylanosterols (12, 13, 14, 15), and by oxysterols like 25-hydroxycholesterol. The specific activity of HMG-CoA reductase is reversibly regulated by phosphorylation by AMP kinase (16) and by oxidation events (17, 18) or possibly by still unknown mechanisms (19, 20). HMG-CoA reductase has a molecular mass of 98 kDa and resides in the endoplasmic reticulum, where the enzyme also is degraded (21, 22).

In the present report we present evidence that SR-12813 decreases cholesterol synthesis by increasing HMG-CoA reductase degradation. Our data further show that none of the other known regulation mechanisms as mentioned above is involved and that SR-12813 does not inhibit HMG-CoA reductase activity directly. Two alternative proposals were tested, that SR-12813 functions as a direct suppressor or that it elicits the formation of regulatory isoprenoids. Our results favor a direct action and further suggest that SR-12813 operates to rapidly inactivate of HMG-CoA reductase activity by a mechanism that primes the enzyme for degradation.

EXPERIMENTAL PROCEDURES

Dulbecco's modified Eagle's medium/HEPES was purchased from Flow laboratories. Cab-osil M-5 was from Fluka. Lovastatin was a generous gift from Merck Sharp and Dohme. Benzamidine, leupeptin, pepstatin, phenylmethylsulfonyl fluoride, soybean trypsin inhibitor, 1-chloro-3-tosylamido-7-amino-2-heptanone, tosylamido-2-phenylethyl chloromethyl ketone, protein A-agarose (high affinity), and N-ethylmaleimide were all from Sigma. BCA protein assay reagent was from Pierce. Fluorotran and Biodyne A membranes were from Pall. HR 58, the catalytic subunit of human HMG-CoA reductase, and rabbit anti-HR58 antibody were from Ruth Mayer (SB, Upper Merion (14)). Rabbit antibody against HMG-CoA reductase catalytic subunit from rat HR 53 (16) was from Dr. G. Hardie, University of Dundee. Hep G2 cells were from the ECACC (deposited by B. Knowles, Wistar Institute). SR-12813 (Fig. 1) was made by Symphar.² cDNA probes for HMG-CoA reductase and LDL receptor were prepared as described (23). [³H₂O], [1,2(n)-³H]cholesterol, [1-¹⁴C]acetate, and [9,10(n)-³H]oleic acid, L-[³⁵S]methionine (in vivo cell labeling grade), anti-rabbit Ig, horseradish peroxidase-linked F(ab)2 from donkey, Enhanced Chemiluminescence Reagents (ECL), and Hyperfilm ECL were from Amersham UK. 25-Hydroxycholesterol was from Steraloids UK Ltd.

Hep G2 cells were grown to confluence on 6-well or 24-well Costar plates in Dulbecco's modified Eagle's medium/20 mM HEPES/10% fetal calf serum/glutamine/0.85 g/liter bicarbonate in 2% CO₂/air. Lipoprotein-deficient serum (LPDS) was prepared from fetal calf serum using Cab-osil M-5 as described previously (14). Unless stated otherwise, before an experiment fetal calf serum was replaced by 5% LPDS overnight.

Tritiated water incorporation into lipids followed basically the protocol of (24) with slight modifications. Briefly, cells were grown to 80% confluency. SR-12813 in Me₂SO (final concentration, 0.1%) was added 1 h before the addition of 1.8 mCi of [³H₂O] to each well. Cells were incubated for a further hour, and tritium incorporation in cholesterol and fatty acids was determined as described (24). For acetate incorporations into cholesterol Hep G2 cells were pre-incubated for 16 h in 5% LPDS medium with or without SR-12813. This was followed by a 1.5-h addition of 16 µCi of [¹⁴C]acetate (13 µCi/µmol)/10-cm² well. Lipids were extracted using chloroform/methanol/water (1:2:0.8, v/v (25)) with the inclusion of a known amount of [³H]cholesterol as a recovery marker for the extractions. The chloroform layer was dried under nitrogen, and the lipid fraction was redissolved in 1 ml of hexane. Lipids were separated by chromatography on Kieselgel 60 (Merck) silica TLC plates developed in hexane/diethyl ether/acetic acid (80:20:1, v/v). The cholesterol band was scraped off, and the radioactivity was determined by liquid scintillation counting.

Mevalonate incorporations were carried out as described (14). Briefly, 10 µM SR-12813 was added to the cells 20 min before the addition of [³H]mevalonic acid lactone (0.2 µCi/nmol; final concentration, 100 µM), and cells were incubated for 2 h. Cells were saponified in 2 M NaOH/65% ethanol, and [¹⁴C]cholesterol (about 2500 dpm) was added to each tube as a recovery marker. Nonsaponifiable lipids were three times extracted from the aqueous phase with hexane. The lipids were redissolved in 150 µl of hexane and analyzed on 20 × 5-cm silica TLC plates developed in hexane/diethylether (1:1). Iodine was used to stain the marker lipids. Bands that aligned with the marker cholesterol were scraped from the plate and counted.

HMG-CoA reductase activity was measured essentially as described (24, 26). Briefly, compounds were added to the cells in Me₂SO (final concentration, 0.1%). After the experiment cells were lysed by the addition of 0.1 ml of 0.25% Brij 96, 0.1 M sucrose, 0.1 M KF, 50 mM KCl, 40 mM potassium dihydrophosphate, 30 mM EDTA, 5 mM dithiothreitol, pH 7.4 at room temperature. In some experiments KF was omitted to measure ``total'' HMG-CoA reductase activity (27). HMG-CoA reductase activity in the cell lysate was further determined as described (24).

Protein levels of HMG-CoA reductase were determined by immunoblot analysis. 6-well Costar plates containing confluent Hep G2 cells were pre-incubated overnight in 5% LPDS medium. For some experiments 0.1 µM lovastatin was included overnight. Cells were incubated with 5 µM SR-12813 in 0.1% Me₂SO. After the period of incubation, the cells were washed twice with ice-cold buffer containing 50 mM HEPES, pH 7.4, 100 mM NaCl, 10 mM EDTA, 10 mM EGTA. Then 300 µl of lysing buffer was added to each well. Lysing buffer consisted of the above buffer with the following added: 1 mM phenylmethylsulfonyl fluoride, 10 mM N-ethyl maleimide, 2.2% Me₂SO, 1% Triton X-100, 0.5 mM leupeptin (28). Lysed cells were homogenized using a 1-ml syringe with and 21 gauge needle. After being on-ice for 10 min, the samples were centrifuged at 11,000 × g for 5 min. To 300 µl of supernatant, 900 µl of sample buffer was added. The sample buffer consisted of 0.5 M Tris-HCl, pH 6.8, 20% glycerol, 2% SDS (w/v), 5% -mercaptoethanol, 0.05% (w/v) bromphenol blue. Samples were immediately loaded onto a 20 × 16-cm casted gel containing 7.5% acrylamide, 0.375 M Tris, pH 8.8, 0.1% SDS (29). For unknown reasons heating of the samples results in the disappearance of the 98-kDa HMG-CoA reductase band and the formation of a band with twice the molecular mass, apparently a dimer. This anomaly has also been observed by other authors (30). BCA (Pierce) protein assay reagent was used to determine protein concentrations. Proteins were blotted onto a fluorotran membrane using 0.025 M Tris-HCl, 0.192 M glycine, 20% methanol as the transfer buffer using a semi-dry blotting apparatus (Biometra). The membrane was blocked with 0.5% casein in PBS overnight at 4 °C while shaking. Rabbit antibody to human HMG-CoA reductase catalytic subunit (HR 58 (14)) was diluted 2500 times in 0.1% casein and incubated for 1 h at room temperature. After four washes each for 5 min in 0.1% Triton X-100 in PBS, membrane was washed in PBS. The secondary antibody (rabbit Ig, horseradish peroxidase-linked F(ab)₂ from donkey) was diluted 10,000 times in 0.1% casein in PBS. This was incubated for 1 h at room temperature before washing four times with 0.1% Triton X-100. The membrane was finally washed in 0.1% Tween 20 in PBS. Proteins were detected using the ECL detection reagents and exposed to Hyperfilm ECL. Bands on the fluorographs were analyzed by densitometry using a video image system that was calibrated with a photographic step tablet (Kodak) and the Optimas image analysis program. Using HR 58 it was determined that the density on the Western blots increased linearly with the protein concentration up to an optical density of 2.

Degradation rates of HMG-CoA reductase in Hep G2 cells were basically determined according to Edwards (31). Experiments were set up in a staggered fashion such that all chases were terminating at the same time. Cells were preincubated in 5% LPDS medium overnight in 6-well plates. Cells were pre-incubated for 15 min with 1.5 ml/well methionine-free 5% LPDS medium made to pH 7.0 before 40 µCi of [³⁵S]methionine was included for 1 h. Under our experimental conditions, methionine incorporations into protein were linear for at least 2 h (not shown). Methionine-free medium was then replaced with 2 mM methionine, 5% LPDS medium, and compound (final concentration, 0.1% Me₂SO). The cells were incubated for a further 0-3 h before all reactions were terminated at the same time. Cells were solubilized in 0.5 ml of ice-cold lysis buffer containing 1% Triton X-100, 0.5% deoxycholate, 0.1% SDS, 50 mM NaCl, 50 mM NaF, 20 mM HEPES, pH 7.2, 5 mM EDTA, 5 mM EGTA (32). The following protease inhibitors were added just before the experiment to the final concentration indicated; 1 mM phenylmethylsulfonyl fluoride, 0.1 mM leupeptin, 0.1 mM 1-chloro-3-tosylamido-7-amino-2-heptanone, 0.1 mM benzamidine, 0.1 mM soybean trypsin inhibitor, 0.1 mM tosylamido-2-phenylethyl chloromethyl ketone (16).

The samples were homogenized using a syringe and a 21 gauge needle before being left on ice for 10 min and then spun in a microfuge for 5 min. To measure the incorporation of [³⁵S]methionine into the total protein fraction, two times 5 µl of the supernatant was mixed with 100 µl of 10 mg/ml BSA, followed by 1 ml of ice-cold 5% (w/v) trichloroacetic acid. The precipitated protein was collected by centrifugation (table centrifuge, 10 min, 4 °C). The pellet was dissolved in 0.1 ml of distilled water and 2 ml of Soluene 350; 10 ml of Hionic Fluor (Packard) was then added for counting.

The remaining supernatant containing the solubilized protein was incubated with 10 µl of rabbit anti-HR53 or with 10 µl of rabbit nonimmune serum on ice for 1 h. 40 µl of protein A-agarose were then added to the samples and incubated overnight at 4 °C on an rotary shaker. After a 20-s centrifugation, the immune complexes were washed three times with 1 ml of lysis buffer (no protease inhibitors). Samples were incubated at room temperature for 30 min in the presence of 150 µl of 2% SDS, 10% glycerol, 5% -mercaptoethanol, 0.001% bromphenol blue, and 62.5 mM Tris, pH 6.8. The protein A beads were pelleted by a 20-s spin in a Eppendorf microfuge. Samples were immediately separated by electrophoresis in the presence of SDS (29) on 16 × 20-cm casted gels containing 7.5% acrylamide. Gels were stained and processed for fluorography using Amplify (Amersham Corp.). Bands on the fluorographs were analyzed as described in the HMG-CoA reductase assays section.

Measurement of LDL receptor-dependent association and degradation of LDL was basically as described previously (24). Briefly, Hep G2 cells were incubated with 5% LPDS medium in the presence of SR-12813, lovastatin (free acid), or 0.1% (v/v) dimethylsulfoxide (control). The cells were incubated for 16 h. The medium was removed, 300 µl of ¹²⁵I-LDL (10 µg/ml, 53 cpm/ng) was added, and incubation continued for 2.5 h. At the end of the incubation, 250 µl of the medium was collected, and the cells were washed at 0 °C six times with 1 ml of ice-cold PBS containing 1% (w/v) bovine serum albumin and 1 mM CaCl₂ and once with PBS only. Cells were dissolved into 0.2 N NaOH, and cellular protein was measured. The radioactive content of the samples was determined and used as a measure for association of LDL. For degradation, bovine serum albumin, trichloroacetic acid, and AgNO₃ (final concentrations, 0.2%, 0.5 M, and 0.115 M, respectively) were added to the medium samples to precipitate ¹²⁵I-LDL and free iodide (33). The degradation product from ¹²⁵I-LDL, iodotyrosine, remained in the supernatant and was counted. Nonspecific ¹²⁵I-LDL association and degradation were determined by including 300 µg/ml of unlabeled LDL in the experiment (1 well/dose).

Hep G2 cells on a 6-well plate were incubated with either 1 µM lovastatin or 3 µM SR-12813 in 5% LPDS medium. mRNA was isolated using a Fast-track kit (Invitrogen). mRNAs were analyzed as described (23). Briefly, mRNAs were separated by electrophoresis on a 1.2% agarose/formaldehyde gel. RNA was transferred to Biodyne A nylon membrane by capillary blotting using 20 × SSC and fixed to the membrane by UV light. Blots were prehybridized for 15 min at 68 °C in 200 mM sodium phosphate, pH 7.2, 1 mM EDTA, 1% w/v BSA, 7% SDS, and 15% deionized formamide (v/v) and then hybridized for 18-20 h at 68 °C in the same buffer containing the ³²P-labeled cDNA probes. Blots were washed at 68 °C in 40 mM sodium phosphate, pH 7.2, 1 mM EDTA, 1% SDS (w/v). Autoradiographs were scanned with a Joyce-Loebl chromoscan III densitometer.

RESULTS

To measure the effects of SR-12813 on sterol synthesis Hep G2 cells were incubated with different radiolabeled precursors, and incorporation of radioactivity into cholesterol was determined. SR-12813 inhibited the incorporation of ³H₂O into cholesterol in Hep G2 cells (Fig. 2A) with an IC₅₀ of 1.2 µM. A similar inhibition with an IC₅₀ of 0.6 µM was found when [¹⁴C]acetate was used instead of ³H₂O (Fig. 2A). There were no effects of SR-12813 on the incorporation of ³H₂O into fatty acids. Incorporation of [³H]mevalonic acid lactone into cholesterol was not influenced in the presence of SR-12813 (Fig. 2B) nor into squalene and farnesol.³ In good agreement with this finding, the sterol profile of the nonsaponifiable lipid fraction from Hep G2 cells treated for 16 h with SR-12813 showed no increase in any of the sterol intermediates (not shown, but similar to sterol profile of control in Ref. 34). These data indicate that SR-12813 interferes with the cholesterol biosynthetic pathway at the level of one of the enzymatic conversions between acetyl-CoA and mevalonate but not beyond HMG-CoA reductase.

Exposure of Hep G2 cells to SR-12813 inhibited cellular HMG-CoA reductase activity with an apparent IC₅₀ of 0.85 µM (Fig. 3A). This agreed well with the IC₅₀ of 1.2 µM found for the inhibition of ³H₂O incorporation into cholesterol shown in Fig. 2A. Inhibition of 50% of the activity was achieved in 10 min (Fig. 3B). At concentrations of 5 µM SR-12813, the maximum inhibition was 80% of the initial activity, and inhibition lasted for periods of 24 h or more (see also Fig. 2A). No direct effect of SR-12813 on the activity of HR 58, the cloned catalytic subunit of human HMG-CoA reductase (14), at concentrations of SR-12813 of up to 0.5 mM was observed nor on the activity of Hep G2 cellular extracts (10 µM). These data indicate that the activity of HMG-CoA reductase is strongly suppressed in the presence of SR-12813 in cells but that there is no direct inhibition. HMG-CoA reductase can be inactivated via an AMP kinase-mediated phosphorylation (16). We investigated if SR-12813 inactivated HMG-CoA reductase via an increased phosphorylation of HMG-CoA reductase. The HMG-CoA reductase activity assay is routinely performed in the presence of fluoride to give the ``expressed activity'' (27), which represents the activity of HMG-CoA reductase in the cell. When fluoride is omitted, phosphorylated, mostly inactive, HMG-CoA reductase is dephosphorylated by phosphatases to yield total activity. Under these conditions HMG-CoA reductase activity was still decreased by SR-12813 (Fig. 3B). HMG-CoA reductase activity in controls increased about 3-fold from 48 to 132 pmol/min/mg cellular protein, showing that reductase was indeed reactivated in our experimental conditions. This finding indicated that SR-12813 did not increase phosphorylation of HMG-CoA reductase.

HMG-CoA reductase has also been reported to be inactivated by oxidation (17, 18). We hypothesized SR-12813 might oxidize HMG-CoA reductase and that dithiothreitol, which is a standard constituent of the HMG-CoA reductase activity assay, might protect the reductase from oxidation. No inactivation of HR 58 reductase activity by SR-12813 was observed under nonreducing conditions when dithiothreitol was omitted from the assay.³

LDL receptor activity and HMG-CoA reductase activity are co-ordinately regulated in Hep G2 cells (24, 35). We therefore investigated the effect of SR-12813 on LDL receptor activity by measuring the receptor-mediated association of ¹²⁵I-LDL with the cells over a period of 2.5 h. Lovastatin was used as a positive control (35, 36). In overnight incubations both lovastatin and SR-12813 enhanced the receptor-mediated association of iodinated LDL with Hep G2 cells (Table I). This was not the result of a decreased catabolism of LDL because LDL receptor-mediated degradation of ¹²⁵I-LDL was also increased (Table I) in the presence of lovastatin or SR-12813. This result shows that the suppression of HMG-CoA reductase activity by SR-12813 results in an increase of LDL receptor activity.

Table I. Effect of SR-12813 on LDL receptor activity in Hep G2 cells Hep G2 cells were incubated overnight with 5% LPDS medium including SR-12813 or lovastatin (free acid) at the indicated concentrations (3 wells/dose). 125I-LDL was added for 2.5 h, after which time LDL receptor-mediated association and degradation were determined as described under ``Experimental Procedures.'' Data of treated cells are all significantly different from controls (t test). Absolute control values were 142 ± 15 and 37 ± 6 ng of 125I-LDL/mg of protein (mean ± S.D.) for association and degradation, respectively. The experiment was repeated once with a similar result. Association ± S.D. Degradation ± S.D. % of control 0.05 µM lovastatin 127  ± 12a 150  ± 28 0.15 µM lovastatin 144  ± 10 158  ± 22 0.5 µM lovastatin 163  ± 15 189  ± 25 3 µM SR-12813 165  ± 7 172  ± 15 10 µM SR-12813 223  ± 5 254  ± 18 a Not significantly different from controls.

The increase in LDL receptor activity might result from regulation at the transcriptional level. Similarly HMG-CoA reductase expression might be regulated by SR-12813. We studied the effect of SR-12813 on mRNA content of HMG-CoA reductase and LDL receptor. The HMG-CoA reductase inhibitor lovastatin that is known to increase mRNA levels of HMG-CoA reductase and LDL receptor (35) was used as a positive control. After a 21-h incubation with SR-12813 or lovastatin, mRNA content of both of these proteins were increased (Fig. 4) in comparison with the housekeeping gene, glyceraldehyde-diphosphate dehydrogenase. There were no short term effects (1-4 h) on expression of mRNA of either protein (not shown). It is concluded that SR-12813 modulates LDL receptor and HMG-CoA reductase at the transcriptional level in a way similar to lovastatin.

The rapid suppression of HMG-CoA reductase activity by SR-12813 suggests the involvement of a post-transcriptional mechanism. We measured the effect of 5 µM SR-12813 on HMG-CoA reductase protein levels in time by Western blot analysis. To facilitate analysis, HMG-CoA reductase levels were induced severalfold by pre-incubation of the cells overnight with 0.1 µM lovastatin (Fig. 5). Under these conditions, SR-12813 was still able to reduce HMG-CoA reductase activity, although suppression of the reductase activity now had a T1/2 of about 30 min compared with 10 min without lovastatin (Fig. 6). Fig. 5 shows the decay of the HMG-CoA reductase protein levels with time as measured by densitometric analysis of the area corresponding to the 98-kDa band. The T1/2 of HMG-CoA reductase protein reduction appears to be slightly longer (1 h) than that found for suppression of enzyme activity in the presence of lovastatin (30 min, Fig. 6). Without lovastatin HMG-CoA reductase protein levels were reduced about 4-fold in the presence of SR-12813 (Fig. 5).

Further experiments were performed in order to establish if this decrease in protein levels by SR-12813 was a result of an enhanced degradation or a decreased synthesis. Pulse-chase experiments were conducted with methionine. The chase was in the presence of 5 µM SR-12813. The anti-HR 53 immune precipitates were analyzed by SDS-polyacrylamide gel electrophoresis. Densitometric analysis of the 98-kDa band on the fluorograph revealed that the rate of HMG-CoA reductase degradation was enhanced more than 4-fold (Fig. 7) in the presence of SR-12813. The apparent half-life of HMG-CoA reductase was reduced from 90 to 20 min. No change was detected in the rate of total protein degradation as was measured by trichloroacetic acid precipitation of the cell lysate (T1/2 of 3.2 and 2.9 h in controls versus treated, respectively). Effects on HMG-CoA reductase synthesis were also investigated. SR-12813 (5 µM) apparently halved the incorporation of [³⁵S]methionine into HMG-CoA reductase over a period of 30 min (not shown). This effect can be fully explained by the rapid degradation of newly synthesized enzyme. The 51% reduction of synthetic rate after 30 min corresponds to a T1/2 for reductase degradation of 29 min following the equation,

(Eq. 1)

where t = the pulse time in h; K_s(app) = ³⁵S cpm incorporated into the reductase protein at time t, K_d (the rate constant for degradation) = 0.693/T1/2 (half-life for the reductase in h), and K_s the rate constant for synthesis in cpm/h (15). This is very close to the T1/2 of 20 min found experimentally. The outcome of these pulse and pulse-chase experiments suggest that SR-12813 mainly influences HMG-CoA reductase degradation.

We compared the properties of SR-12813 with those of known suppressors of HMG-CoA reductase activity, the oxysterol 25-hydroxycholesterol and the oxylanosterol SK&F 104976 (14). The effects on HMG-CoA reductase activity and degradation were studied. Fig. 8 shows that the suppression of HMG-CoA reductase activity by SR-12813 was rapid (T1/2 = 8.5 min) compared with 25-hydroxycholesterol (T1/2 = 30 min) and SK&F 104976 (T1/2 = 141 min). As was predictable from the effects on enzyme activity, the increase in the rate of degradation of the reductase is much greater for the bisphosphonate than for the oxysterols (Table II). For both activity and degradation there was a lag time of 20-30 min before any effects of 25-hydroxycholesterol were observed. The above results suggest that SR-12813 increases HMG-CoA reductase degradation via a mechanism different from both oxysterol classes.

Table II. Effect of oxysterols, SR-12813, and lovastatin on HMG-CoA reductase degradation Hep G2 cells were incubated overnight in 5% LPDS medium. In some experiments 50 µM lovastatin was included during the methionine-free pre-incubation, the pulse, and the chase. The chase periods were 1.5, 3, and 5 h, and HMG-CoA reductase half-life was determined as described in the legend to Fig. 7. 25-Hydroxycholesterol, SK&F 104976, SR-12813, or 0.1% Me2SO (control) were added at the start of the chase. In the final experiment, lovastatin was only present during the chase. Experiment T1/2 Acceleration of degradation Control Treated min -fold 5 µM SR-12813 90 20 4.5 100 nM SK&F 104976 116 77 1.5 5 µM 25-hydroxycholesterol 179 90 2.0 50 µM lovastatin + 5 µM SR-12813 1027 260 4.0 50 µM lovastatin + 5 µM 25-hydroxycholesterol 1600 389 4.1 10 µM lovastatin (chase only) + 5 µM SR-12813 143 25 5.7

It is well known that mevalonate-derived metabolites play an important part in HMG-CoA reductase degradation (7). We tested the hypothesis that SR-12813 might promote the formation of these mevalonate-derived regulatory molecules. Incubation with high concentrations of lovastatin should deplete the cell from mevalonate and thus would inhibit the effects of SR-12813. Inclusion of 50 µM lovastatin in the medium during the methionine-free preincubation, pulse, and chase periods resulted in a 10-fold increase in T1/2 for HMG-CoA reductase degradation (Table II). This suggests that lovastatin strongly inhibited mevalonate-driven HMG-CoA reductase degradation. In contrast, this increase in half-life was not found when lovastatin was added during the chase period only (Table II). Despite the presence of lovastatin, SR-12813 increased the rate of degradation of HMG-CoA reductase, suggesting that this compound acts independently from mevalonate-derived metabolites. The same independence of mevalonate was observed for 25-hydroxycholesterol-mediated reductase degradation.

DISCUSSION

Our results clearly demonstrate that SR-12813 inhibits cholesterol synthesis at the level of HMG-CoA reductase. Pulse-chase experiments with methionine showed that the reduction of cholesterol synthesis resulted from a 4-fold increase in rate of HMG-CoA reductase degradation. The also observed decrease in HMG-CoA reductase synthesis may be fully explained by this enhanced degradation and is supported by the lack of decrease in mRNA levels for up to 4 h in the presence of SR-12813. However, it should be noted that SR-12813-mediated degradation of HMG-CoA reductase was so fast that it was difficult to make an accurate assessment of its effects on synthesis. Therefore we cannot completely exclude that SR-12813 may also reduce HMG-CoA reductase synthesis to some degree.

Overnight exposure of Hep G2 cells with SR-12813 did result in an increase of both LDL receptor and HMG-CoA reductase mRNA. This is similar to effects of lovastatin in Hep G2 cells described by other authors (35). The explanation is that the decrease in cellular cholesterol biosynthesis causes a co-ordinate regulated increase of gene expression of LDL receptor and HMG-CoA reductase. The resulting increase of LDL receptor activity by SR-12813 is likely to be an important contributor to the strong hypocholesterolemic effect of this class of bisphosphonates (1, 37). Collectively, our data support the premise that SR-12813 regulates HMG-CoA reductase mainly through an enhanced degradation.

Degradation of HMG-CoA reductase by SR-12813 (T1/2 = 20 min; Fig. 7) may be preceeded by the rapid inactivation of the protein activity (T1/2 = 10 min, Fig. 3B), as was also suggested for 25-hydroxycholesterol by other authors (19). A rapid inactivation of the reductase might occur by phosphorylation (38) or by oxidation (17, 18), which might lead to an enhanced degradation (39). HMG-CoA reductase and acetyl-CoA carboxylase are both inactivated by phosphorylation via AMP kinase (16). Our data suggest that phosphorylation is not involved in the mode of action of SR-12813. Firstly, the incorporation of tritiated water into fatty acids (Fig. 2A) would have been suppressed by SR-12813 if AMP kinase-mediated phosphorylation was involved. Secondly, dephosphorylation by substitution of chloride (total activity) for fluoride (expressed activity) in the HMG-CoA reductase activity assay did not abolish the effect of SR-12813 (Fig. 3B). It is also unlikely that SR-12813 enhances oxidation of HMG-CoA reductase because firstly SR-12813 does not inactivate the reductase activity of HR 58 when dithiothreitol is omitted from the activity assay,³ and secondly SR-12813 belongs to a compound class that has anti-oxidant activity (40).

The underlying mechanism for triggering rapid degradation of HMG-CoA reductase is still unresolved in the literature. Both mevalonate-derived metabolites and sterols can induce HMG-CoA reductase degradation (7, 8). The requirement of mevalonate in HMG-CoA reductase degradation has been shown by eliminating mevalonate synthesis using a high concentration of lovastatin (7). We assume that in our experiments very low intracellular levels of mevalonate were established by incubating the cells with 50 µM lovastatin during the preincubation and [³⁵S]methionine pulse, because this resulted in a 10-fold increase of the T1/2 for HMG-CoA reductase degradation (Table II). Some authors have implied that lovastatin itself prevents HMG-CoA reductase degradation (41). Against this argument is our finding that when lovastatin is added only during the methionine chase, the T1/2 of reductase degradation was not increased (Table II). One might expect that addition of lovastatin during the chase would decrease levels of mevalonate-derived HMG-CoA reductase suppressors, followed by a consequential increase of the T1/2 for reductase degradation. However, it has been shown that mevalonate-induced degradation of HMG-CoA reductase that is already initiated is not affected by the removal of mevalonate (42), which supports our finding that there is no effect on the T1/2 of reductase degradation when lovastatin is added only during the chase. Therefore we conclude that our data are consistent with an elimination of mevalonate-derived activators of HMG-CoA reductase degradation in the presence of high concentration of lovastatin. Under these conditions SR-12813 was still able to accelerate HMG-CoA reductase degradation (Table II), implying that mevalonate is not an essential requirement for SR-12813-mediated HMG-CoA reductase degradation. We found the same for 25-hydroxycholesterol (Table II). Other authors found that mevalonate is required to allow sterol-mediated HMG-CoA reductase degradation (7, 42). This discrepancy might be due to differences between the CHO and Hep G2 cell lineages similar to those observed between CHO and C100 cells (43).

Another class of compounds that may operate by a related mechanism is formed by the tocotrienols (44). Similarly to SR-12813, -tocotrienol induces LDL receptor activity and both compounds reduce HMG-CoA reductase activity mainly via an enhanced reductase degradation, which is independent of mevalonate-derived isoprenoids. SR-12813 differs from tocotrienols in that it acts more rapidly on HMG-CoA reductase (T1/2 of activity decrease is 10 min versus 2 h, respectively) and that SR-12813 increases mRNA levels of HMG-CoA reductase after 21 h, whereas -tocotrienol suppresses reductase mRNA levels initially and reduces HMG-CoA reductase synthesis. From the independence of mevalonate we propose that similar to tocotrienols, SR-12813 acts as a mimetic of a post-mevalonate regulator of reductase. Farnesol and analogues of farnesol have been proposed as post-mevalonate regulators of HMG-CoA reductase (11, 42, 45, 46), and tocotrienols are thought to act as farnesol analogues. One can speculate that SR-12813 mirrors a farnesol-derived metabolite that triggers HMG-CoA reductase degradation via a more direct action than -tocotrienol.

In preliminary experiments we observed that SR-12813-mediated inhibition of HMG-CoA reductase activity was attenuated by the Calpain inhibitor N-acetyl-Leu-Leu-norleucinal. This suggests that nonlysosomal Ca²⁺-dependent proteases are involved in SR-12813 mediated reductase degradation, similar to mevalonate and sterol-induced reductase degradation (47). Recently novel microsomal cysteine proteases ER 60 and ER 72 have been identified that are possibly involved in the degradation of HMG-CoA reductase (48). The involvement of cysteine proteases needs to be elaborated further in future studies.

To conclude, HMG-CoA reductase is known to be regulated in the cell by a range of molecular mechanisms. In this paper we have shown that SR-12813 can regulate HMG-CoA reductase rapidly on the level of protein via an enhanced degradation.