Squalestatin 1, a Potent Inhibitor of Squalene Synthase, Which Lowers Serum Cholesterol in Vivo*

Squalestatin 1 is a member of a novel family of fermentation products isolated from a previously unknown Phoma species (Coelomycetes). Squalestatin 1 is a potent, selective inhibitor of squalene synthase, a key enzyme in cholesterol biosynthesis; in vitro, 50% inhibition of enzyme activity is observed at a concentration of 12 +/- 5 nM (range of 4-22 nM). Squalestatin 1 inhibits cholesterol biosynthesis from [14C]acetate by isolated rat hepatocytes (50% inhibition at 39 nM) and by rat liver in vivo. In marmosets, a species with a lipoprotein profile similar to that of man, squalestatin 1 lowers serum cholesterol by up to 75%. This compound will allow further investigation of the control of the sterol biosynthesis pathway and could also lead to the development of new therapies for elevated serum cholesterol.

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ To whom correspondence should be addressed. 11 Present address: Rhone-Poulenc-Rorer, Dagenham, Essex, UK. The abbreviations used are: FPP, farnesyl diphosphate; HPLC, high performance liquid chromatography. enzymatic step after the pathway branches to other isoprenederived compounds, it has been proposed that this enzyme is a rate-limiting step because it is regulated to control flux through the pathway both in vitro (Brown and Goldstein, 1980) and in vivo (Bruenger and Rilling, 1986). Other groups have described studies using inhibitors of this enzyme based on substrate analogues (Biller et al., 1988(Biller et al., , 1990, but these compounds have been weak inhibitors and have not been suitable for evaluation in uiuo. Using a high through-put screen based on a novel enzyme assay (Tait, 1992), a fungal isolate was discovered that produces a related group of compounds (the squalestatins) which are potent inhibitors of mammalian and fungal squalene synthase. We have used one of this class of inhibitors (squalestatin 1) to probe the role of squalene synthase in controlling the flux of cholesterol biosynthesis and have shown that compounds of this type can cause profound lowering of serum cholesterol levels in uiuo.
Animals-The rats used in these studies were juvenile males from the AHA strain, except for the studies on cholesterol biosynthesis in vivo where male CD rats (Charles River) were used. All rats were fed on standard laboratory chow and allowed access to water ad libitum. The marmosets used were adults between 15 and 28 months of age, and these were fed a standard primate diet supplemented with fruit.
Measurement of Squalene Synthase Actiuity-Squalene synthase activity was measured using rat liver microsomes as enzyme source and a novel separation of substrates and products. This method is described in detail elsewhere (Tait, 1992), but briefly, after incubation of enzyme with [14C]FPP, the reaction mixture is absorbed onto silica TLC plates followed by elution of reactants by washing with a buffered sodium dodecyl sulfate solution.
[14C]Squalene is retained on the TLC plate and can be quantified by scintillation counting after cutting up the plate. This assay gives results comparable with those obtained with conventional separation of [14C]squalene by TLC.
Isolation of Rat Hepatocytes-Rat hepatocytes were isolated by collagenase perfusion (Seglen, 1973). Viability was assessed by exclusion of trypan blue, and cells were routinely obtained with viabilities

>80%.
Labeling of Cholesterol and Pathway Intermediates in Hepatocytes-Measurement of cholesterol biosynthesis was carried out essentially as described by Tsujita et al. (1986) except synthesized ["C] cholesterol was measured by HPLC. Briefly, isolated hepatocytes were incubated with [14C]acetate (together with squalestatin 1 where appropriate) for 30 min, further reaction was stopped by adding KOH (15% (w/v) in 90% ethanol), and the mixture was saponified overnight at 60 "C. The next day, the mixture was extracted twice with petroleum ether and the organic phases pooled and evaporated under a stream of nitrogen. The residue was dissolved in methanol/isopropanol (l:l, v/v) and ['4C]cholesterol separated by HPLC using a Spherisorb 5-pm ODs2 column eluted with methanol/isopropanol (4:1, v/v). The ['4C]cholesterol was quantified using a Betacord 1208 radioactivity monitor with a 0.5-ml flow cell after mixing the column eluant with Optiphase Safe scintillant (LKB).
To label the intermediates in the pathway, primary rat hepatocytes were plated at a density of 0.5 X lo6 cells/ml in 10-cm2 wells containing 2 ml of Dulbecco's modified Eagle's medium supplemented with 10% (v/v) bovine serum. After 18 h, the medium was changed to Dulbecco's modified Eagle's medium without serum, squalestatin 1 was added after another 2 h, and after 1 h, 2 pCi of [3H]mevalonolactone (27.8 Ci/mmol) was added to each well. At various times after the addition of radiolabel, the medium was rapidly decanted and the cells harvested into 2 ml of ice cold phosphate-buffered saline using a rubber policeman. The harvested cells were extracted as described by Watson et al. (1985) to determine isoprenoid intermediates by HPLC separation. Parallel incubations were saponified and treated for measurement of [3H]squalene and [3H]cholesterol levels by HPLC as described above. Measurement of ATP:Citrate Lyase Activity-A crude preparation of ATP:citrate lyase was prepared as a post-mitochondrial supernatant from rat liver homogenized in 0.25 M sucrose, 10 mM mercaptoethanol, 1 mM MgCl,, and 20 mM Tris-HC1 and centrifuged 10,000 X g for 10 min. Enzyme was incubated with [I4C]citrate (0.08 pCi/ml, 0.7 mM), ATP (10 mM), coenzyme A (0.7 mM), and hydroxylamine (200 mM); the last component converts one of the products (["C] acetyl-coA) to the corresponding hydroxamate. After incubation at 37 "C for 10 min, excess unlabeled citrate was added. The ["C] acetylhydroxamate was separated from unreacted substrate and ["C] oxaloacetate by precipitation of the radiolabeled acids as a complex with barium after incubation with barium hydroxide overnight at 4 "C.
[14C]Acetylhydr~xamate in the supernatant was measured by liquid scintillation counting. This assay gives results comparable with those described by Watson et al. (1969).
Citrate Efflux from Mitochondria-Rat liver mitochondria were prepared as described by Harper and Saggerson (1975) and checked for functional integrity prior to use by determination of their respiratory control ratio using an oxygen electrode.
Citrate efflux from mitochondria was measured using the method of Robinson et al. (1970) where intra-mitochondrial citrate is labeled by incubation with [I4C]bicarbonate, and efflux of ["Clcitrate stimulated by addition of extra-mitochondrial unlabeled citrate. Specific citrate-stimulated efflux was defined by that proportion of total efflux that was inhibited by 1,2,3-benzenetricarboxylic acid.
Measurement of Cholesterol Biosynthesis in Vivo-Cholesterol biosynthesis in vivo was carried out essentially as described by Tsujita et al. (1986), except that [14C]cholesterol was measured by HPLC. Squalestatin 1 was dosed intravenously to rats followed immediately by intraperitoneal administration of ["Clacetate (250 pCi/kg). After 1 h, the rats were killed and the livers removed, and a sample of 0.5 g was saponified in alcoholic KOH at 80 "C for 1.5 h. ['4C]Cholesterol was measured after extraction and HPLC separation as described above.
Effect on Serum Cholesterol Levels in Marmosets-Marmosets of mixed sex were fasted overnight and blood samples (500 pl) taken from the femoral vein and collected in Microtainers (Becton Dickinson). Serum was obtained by centrifugation for 4 min at 10,000 x g in a Beckman Microfuge. Serum cholesterol concentrations were determined using a standard assay kit (Boehringer 237574) on a Kone Progress Plus autoanalyzer. Animals were allocated to treatment groups on the basis of their fasting serum cholesterol levels such that the mean and distribution of serum cholesterol levels were similar for each group.
Effects on Apoprotein Levels-Eight marmosets were dosed orally with squalestatin 1 (10 mg/kg/day in water) for 7 days. Serum samples were collected from the femoral vein prior to the animals receiving the first dose of squalestatin 1, and again 24 h after the last dose. Apolipoprotein levels were measured immunoturbidimetrically using Sigma kits for human apolipoproteins B and A1 (Sigma catalog 357-A and 356-A, respectively).

RESULTS
Biological Effects in Vitro-Squalestatin 1 is a potent inhibitor of squalene synthase from both rat and marmoset liver, with an ICso of 12 nM (range 4-22 nM). Typical results of in vitro inhibition of the rat liver enzyme are shown in Fig. 2. Similar inhibitory activity is seen against squalene synthase present in microsomes isolated from Hep-GZ cells, a human hepatoma line, and from Candida albicans (data not shown). Squalestatin 1 also inhibits the synthesis of cholesterol from [I4C]acetate by isolated rat hepatocytes, with an ICbo of 39 nM (Fig. 3). The tricarboxylic acid structure of squalestatin 1 resembles citrate, and 1,2,3-benzenetricarboxylate has been shown to inhibit cholesterol synthesis through its effect on citrate efflux from mitochondria (Claeys and Ami, 1989). Therefore, the effects of squalestatin 1 on the early citrateutilizing steps in the cholesterol biosynthetic pathway were examined, i.e. citrate efflux from mitochondria and ATP:citrate lyase activity. Squalestatin 1 has no effect on ATP:citrate lyase activity at concentrations up to 100 p~ and no effect on citrate efflux from mitochondria at concentrations up to 1 mM.
When isoprenoid intermediates were measured in isolated hepatocytes, [3H]FPP could be readily detected in these cells after 30 min incubation, and the level decreased slowly over the next hour; other isoprenoid intermediates could not be detected. In the presence of 50 or 500 nM squalestatin 1, the amount of radiolabel in FPP increased approximately 6-fold at all time points, as would be expected from inhibition of squalene synthase (Fig. 4). The reason for the decrease in labeling of FPP over 90 min is under further investigation; in control cells, labeling of cholesterol from [3H]mevalonate increases linearly over this time. Similar experiments in which radiolabeled squalene in cell extracts was measured by HPLC showed that squalene labeling (which is low but measurable in control cells) was not detectable in the presence of squalestatin 1 (data not shown). Effects on Cholesterol Biosynthesis in Vivo-To help investigate the role of squalene synthase activity in determining the flux in the cholesterol biosynthesis pathway in situ, an in vivo method in rat liver was used. The results are presented in Fig. 5 and show that cholesterol biosynthesis in viuo can be inhibited by 50% after dosing with 0.1 mg/kg intravenously.
Effects on Serum Cholesterol Levels in Viuo-Inhibitors of cholesterol biosynthesis which act at hydroxymethylglutaryl-CoA reductase are reported to have no effect on the serum cholesterol levels in rodents, although other mammalian species such as rabbits or primates respond to these agents (Tsujita et al., 1986). Adult marmosets have been shown to possess a serum lipoprotein profile similar to that of man (Crook et al., 1990), and we have shown that this species is sensitive to inhibitors of hydroxymethylglutaryl-CoA reductase.' Fig. 6a shows the effect of squalestatin 1 on serum cholesterol in adult marmosets; a significant effect is apparent at an oral dose of 10 mg/kg/day, and cholesterol lowering of up to 75% can be achieved at an oral dose of 100 mg/kg/day. The cholesterol lowering is apparent within 24 h (data not shown) and can be maintained for at least 8 weeks on prolonged dosing with no attenuation of the response (Fig. 66). Table I shows the effects of squalestatin 1 on apolipoprotein levels in the marmosets. Apo-B, characteristic of low density lipoprotein (and very low density lipoprotein), is reduced by 45%, whereas apo-A1 levels (indicative of the high density lipoprotein fraction) are unchanged.

DISCUSSION
Many studies on the pathway of sterol synthesis have used inhibitors of various steps, including hydroxymethylglutaryl-CoA reductase, hydroxymethylglutaryl-CoA synthase, squalene epoxidase, and lanosterol 14-a-demethylase. With the discovery of squalestatin 1, a potent inhibitor of squalene synthase, we have been able to study inhibition of the first committed step of cholesterol biosynthesis, after the branchpoint in the pathway. The compound does not inhibit hydroxymethylglutaryl-CoA reductase, another important enzyme in A. D. McCarthy and R. J. Williams, manuscript in preparation.

Squalene Synthase Inhibition: in Vitro and in Vivo Effects
controlling cholesterol biosynthesis, at concentrations up to 5 PM in vitro (data not shown). Additional evidence for its selectivity is provided by its lack of effect on the mitochondrial tricarboxylate transporter and its poor activity against ATPxitrate lyase.
We have also shown squalestatin 1 to be a potent inhibitor of cholesterol synthesis in freshly isolated hepatocytes (ICs0 = 39 nM). Similar ICs0 values are obtained if the experiment is performed using hepatocytes that have been allowed to adhere to tissue culture dishes overnight, showing that the effect of squalestatin 1 is not related to the trauma of the collagenase-perfusion isolation method (data not shown). The observation in hepatocytes that the labeling of the precursor (FPP) increases while that of the product (squalene) decreases provides strong evidence that squalene synthase is the site of action of squalestatin 1 and further underlines the selectivity of action of the compound. In addition, the flux through the cholesterol biosynthetic pathway is readily inhibited by squalestatin 1, both in isolated hepatocytes and in rat liver in vivo; this suggests that squalene synthase activity in hepatic cells is similar to the rate of flux through the pathway, consistent with its proposed role as a rate-determining step in the pathway (Brown and Goldstein, 1980;Bruenger and Rilling, 1986).
Marmosets have been shown to have a lipoprotein metabolism similar to that of man (Crook et al., 1990), and therefore we believe that this primate is a good model in which to study the effects of cholesterol lowering drugs. Our observations with squalestatin 1 in marmosets provide the first demonstration that a potent inhibitor of squalene synthase can affect serum cholesterol in vivo. It is important that only the apo-B-containing lipoproteins are affected. As squalene synthase is the first committed step on the cholesterol biosynthetic pathway, a therapeutic agent acting by inhibiting this step should have minimal effects on non-steroidal products of the pathway from mevalonate. Perturbations in the levels of ubiquinone have been implicated in some of the side effects of inhibitors of hydroxymethylglutaryl-CoA reductase (Folkers et aL, 1990). Additional studies to determine the mechanism of inhibition at a molecular level and possible clinical utility of this class of compounds are under way.