Significance of Garlic and Its Constituents in Cancerand Cardiovascular Disease

Inhibition of Sterol 4a-Methyl Oxidase Is the Principal Mechanism byWhich Garlic Decreases Cholesterol Synthesis1–3

Dev K. Singh and Todd D. Porter4

Department of Pharmaceutical Sciences, College of Pharmacy, University of Kentucky, Lexington,KY 40536-0082

ABSTRACT Clinical and experimental evidence indicates that garlic ingestion lowers blood cholesterol levels, andtreatment of cells in culture with garlic and garlic-derived compounds inhibits cholesterol synthesis. To identify theprincipal site of inhibition in the cholesterolgenic pathway and the active components of garlic, cultured hepatomacells were treated with aqueous garlic extract or its chemical derivatives, and radiolabeled cholesterol andintermediates were identified and quantified. Garlic extract reduced cholesterol synthesis by up to 75% withoutevidence of cellular toxicity. Levels of squalene and 2,3-oxidosqualene were not altered by garlic, indicating that thesite of inhibition was downstream of lanosterol synthesis, and identical results were obtained with 14C-acetate and14C-mevalonate, confirming that 3-hydroxy-3-methylglutaryl-CoA reductase activity was not affected in these short-term studies. Several methylsterols that accumulated in the presence of garlic were identified by coupled gaschromatography–mass spectrometry as 4,49-dimethylzymosterol and a possible metabolite of 4-methylzymosterol;both are substrates for sterol 4a-methyl oxidase, pointing to this enzyme as the principal site of inhibition in thecholesterolgenic pathway by garlic. Of 9 garlic-derived compounds tested for their ability to inhibit cholesterolsynthesis, only diallyl disulfide, diallyl trisulfide, and allyl mercaptan proved inhibitory, each yielding a pattern of sterolaccumulation identical with that obtained with garlic extract. These results indicate that compounds containing anallyl-disulfide or allyl-sulfhydryl group are most likely responsible for the inhibition of cholesterol synthesis by garlicand that this inhibition is likely mediated at sterol 4a-methyl oxidase. J. Nutr. 136: 759S–764S, 2006.

KEY WORDS: � garlic � cholesterol synthesis � sterol 4a-methyl oxidase � diallyl disulfide � lanosterol

Garlic is regarded with much interest by the general publicas a means to safely reduce blood cholesterol levels. Indeed,

several clinical trials and meta-analyses support the ability ofgarlic to reduce blood cholesterol, although the decrease istypically modest (1–3). Although the mechanism by whichgarlic reduces cholesterol levels has not been established,studies with garlic extracts have shown that garlic compoundsinhibit cholesterol synthesis in cultured hepatocytes, in liverhomogenates, and in cultured hepatoma cells (4–7) and thatthis inhibition occurs in a dose-dependent manner that isnot related to cellular toxicity. In several studies the use of14C-mevalonate instead of acetate prevented the decreasein cholesterol synthesis (5,8). This suggests that garlic de-creased 3-hydroxy-3-methylglutaryl (HMG)5-CoA reductaseactivity, the second and regulated step in cholesterol synthesis.Other enzymes in the pathway were not examined, althoughGebhardt et al. (5,9,10) reported that higher concentrationsof extract, as well as allicin-derived compounds, led to theaccumulation of lanosterol, dihydrolanosterol, and 7-dehydro-cholesterol, suggesting the inhibition of later steps in choles-terol synthesis.

1 Published in a supplement to The Journal of Nutrition. Presented at thesymposium ‘‘Significance of Garlic and Its Constituents in Cancer and Cardiovas-cular Disease’’ held April 9–11, 2005 at Georgetown University, Washington, DC.The symposium was sponsored by Strang Cancer Prevention Center, affiliatedwith Weill Medical College of Cornell University, and Harbor-UCLA Medical Center,and co-sponsored by American Botanical Council, American Institute for CancerResearch, American Society for Nutrition, Life Extension Foundation, GeneralNutrition Centers, National Nutritional Foods Association, Society of Atheroscle-rosis Imaging, Susan Samueli Center for Integrative Medicine at the University ofCalifornia, Irvine. The symposium was supported by Alan James Group, LLC,Agencias Motta, S.A., Antistress AG, Armal, Birger Ledin AB, Ecolandia Inter-nacional, Essential Sterolin Products (PTY) Ltd., Grand Quality LLC, IC Vietnam,Intervec Ltd., Jenn Health, Kernpharm BV, Laboratori Mizar SAS, Magna Trade,Manavita B.V.B.A., MaxiPharm A/S, Nature’s Farm, Naturkost S. Rui a.s., NicheaCompany Limited, Nutra-Life Health & Fitness Ltd., Oy Valioravinto Ab, Panax, PT.Nutriprima Jayasakti, Purity Life Health Products Limited, Quest Vitamins, Ltd.,Sabinco S.A., The AIM Companies, Valosun Ltd., Wakunaga of America Co. Ltd.,and Wakunaga Pharmaceutical Co., Ltd. Guest editors for the supplementpublication were Richard Rivlin, Matthew Budoff, and Harunobu Amagase. GuestEditor Disclosure: R. Rivlin has been awarded research grants from Wakunaga ofAmerica, Ltd. and received an honorarium for serving as co-chair of theconference; M. Budoff has been awarded research grants from Wakunaga ofAmerica, Ltd. and received an honorarium for serving as co-chair of theconference; and Harunobu Amagase is employed by Wakunaga of America, Ltd.

2 Author disclosure: No relationships to disclose.3 Supported by a grant from the American Heart Association.4 To whom correspondence should be addressed. E-mail: tporter@email.

uky.edu.

5 Abbreviations used: DMEM, Dulbecco’s modified medium; HMG, 3-hydroxy-3-methylglutaryl.

0022-3166/06 $8.00 � 2006 American Society for Nutrition.

759S

by on October 19, 2010

jn.nutrition.orgD

ownloaded from

Garlic is rich in sulfur-containing compounds, principallyS-allylcysteine and alliin, the latter of which is rapidly metabo-lized when garlic is crushed and alliinase is released. The highlyreactive sulfenic acid that is formed from alliin condenses toallicin, which then rapidly recombines to various di- and tri-sulfides, depending on conditions. Ultimately these compoundsare believed to yield allyl mercaptan and allyl methyl sulfide,which can react with cellular components or be eliminated onthe breath. The organosulfur compounds formed in garlic arehighly reactive with other sulfhydryl compounds, includingcysteines found in proteins, and it is likely that the chemicalmodification of enzyme-sulfhydryls is responsible for thepurported therapeutic effects of garlic. The question of whichcompounds are most important to the therapeutic effects ofgarlic remains unresolved, although several studies have shownthat the diallyl disulfides, allyl mercaptan, and S-alk(en)ylcysteines are effective inhibitors of cholesterol synthesis in cells(6–8,10). Similarly, the enzyme targets that mediate the effectsof garlic have not been identified.

The present studies were undertaken to identify thecholesterolgenic enzyme or enzymes inhibited by garlic andthe active principles therein. Our studies with hepatoma cellsin which cholesterol and intermediates are radiolabeled andidentified by coupled gas chromatography–mass spectrometryreveal that garlic causes the accumulation of sterol 4a-methyloxidase substrates and that an allyl disulfide or allyl sulfhydrylgroup is necessary for inhibition by garlic-derived compounds.

MATERIALS AND METHODS

Chemicals. Dulbecco’s modified medium (DMEM), penicillin-streptomycin-glutamine (3 100), fetal bovine serum, and trypsinwere purchased from Invitrogen. Diallyl disulfide, diallyl trisulfide, allylmercaptan, allyl methyl sulfide, lactate dehydrogenase, pyruvate,NADH, Triton X100, cholesterol, ketoconazole, squalene, andlanosterol were purchased from Sigma Chemical Co. Zymosterol(8,24(5a)-cholestadien-3b-ol), desmosterol (5,24-cholestadien-3b-ol),24-dihydrolanosterol, lathosterol (7, 5a-cholesten-3b-ol), and 7-dehydrocholesterol were purchased from Steraloids, Inc. Terbinafinewas from TCI America, Inc.; AMO 1618 was bought from Calbiochem(EMD Biosciences). 14C-Acetate, sodium salt (56 mCi/mmol; 2.07GBq/mmol) and 14C-mevalonate, DBED salt (65 mCi/mmol; 2.41GBq/mmol) were purchased from Amersham, Inc. Alliin, S-allylcysteine,S-methylcysteine, S-ethylcysteine, and S-propylcysteine were donatedby Wakunaga of America.

Preparation of garlic extract. Fresh garlic cloves obtained from acommercial grocer were peeled, chopped, and homogenized with amortar and pestle and a small amount of quartz sand in 20 mmol/L Trisbuffer, pH 7.4. The debris was removed by centrifugation at 10,000 3g for 30 min at 48C, and the supernatants were divided into equal partsand stored at 2808C.

Cytotoxicity assays. McARH7777 rat hepatoma cells (ATCC),used between passages 1 and 40, were grown in DMEM supplementedwith 10% fetal bovine serum and 1.0% penicillin-streptomycin-glutamine in 6-well plates at 378C under a humidified atmosphere of5% CO2 for 48 h, after which the medium was replaced with freshmedium containing garlic extract or a garlic derivative. Stocks of diallyldisulfide, diallyl trisulfide, allyl mercaptan, and allyl methyl sulfidewere prepared in 95% ethanol and added dropwise with stirring toDMEM just before use; ethanol constituted ,5% of the final volume.Controls received ethanol alone. After 3 h the cells were detachedwith trypsin, suspended in DMEM containing 0.2% trypan blue, andcounted in a hemocytometer. Leakage of lactate dehydrogenase fromcells was determined by measurement of NADH oxidation from addedpyruvate spectrophotometrically at 340 nm and compared with totallactate dehydrogenase activity from cells lysed with 0.2% Triton X100.

Determination of sterol synthesis. Hepatoma cells were culturedfor 48 h in 6-well plates, at which time the medium was replaced and

appropriate concentrations of the test substances (garlic extract orderivative) were added along with 1 mCi (37 kBq) of 14C-acetate or14C-mevalonate. Incubations were carried out for 3 h, after which timethe cells were washed twice with phosphate-buffered saline; harvestedby trypsinization or scraping; resuspended in 20 mmol/L Tris buffer, pH7.4, containing 0.1% Triton X100; and lysed by sonication (SonicDismembrator, Fisher Scientific) at medium setting on ice with 10 8-spulses, separated by 30 s each. Lipids were extracted into 5 mL ofchloroform:methanol (2:1), the solvent was removed by evaporativecentrifugation, and the lipids were resuspended in 50 ml of chloroform/methanol and spotted onto silica thin layer plates (Whatman). Chro-matography was carried out in petroleum ether:ethyl ether:acetic acid(60:40:1). Cholesterol, 7-dehydrocholesterol, lanosterol, dihydrola-nosterol, lathosterol, desmosterol, and zymosterol were identified bycochromatography of authentic standards visualized by iodine vaporand quantified by electronic autoradiography (Packard Instant Imager).Further confirmation of the identity of these and unknown sterols wasobtained by scraping the corresponding region of nonradiolabeledsamples into chloroform:methanol (2:1), derivatizing the samples withtrimethylsilane, and submitting them to gas-chromatographic separationon a Trace gas chromatograph with a DB-5ms column with heliumcarrier gas, followed by ion-trap mass spectrometry on a ThermoFinniganPolarisQ at the University of Kentucky Mass Spectrometry Facility.

Determination of squalene, 2,3-oxidosqualene, and lanosterolsynthesis. For the determination of squalene, 2,3-oxidosqualene,and lanosterol synthesis, cells were incubated as described above forcholesterol synthesis with the inclusion of 60 mmol/L terbinafine,an inhibitor of squalene monooxygenase (for the determination ofsqualene), or 0.3 mmol/L AMO 1618, an inhibitor of oxidosqualenecyclase (for the determination of 2,3-oxidosqualene), or 10 mmol/Lketoconazole, an inhibitor of lanosterol demethylase (for the determi-nation of lanosterol). Lipids were saponified by the addition of 0.5 mLof 10% methanolic potassium hydroxide and incubated at 808C for 1 h.For the determination of squalene and 2,3-oxidosqualene, the neutrallipids were extracted into 5 mL of petroleum ether; the solvent wasremoved by centrifugal evaporation, and the samples were resus-pended in 50 mL of petroleum ether and resolved by silica thin-layerchromatography in 5% ethyl acetate in hexane. Lanosterol wasdetermined as described for cholesterol. Authentic standards forsqualene and lanosterol were visualized by iodine-vapor staining; 2,3-oxidosqualene was confirmed by cochromatography of the product of14C-squalene conversion to 2,3-oxidosqualene by purified recombi-nant squalene monooxygenase (11). Further confirmation of theseproducts was obtained by scraping the corresponding region of non-radiolabeled samples into chloroform:methanol (2:1) and submittingthem to mass spectrometric analysis as described above.

RESULTS

Treatment of McARH7777 rat hepatoma cells with anaqueous garlic extract (#5 g/L) reduced the incorporation of14C-acetate into cholesterol over a 3-h time period by #75%without evidence of cellular toxicity (Fig. 1). At 7.5 g/L, garlicextract caused a marked elevation in lactate dehydrogenaseactivity in the medium, indicating the release of this enzymefrom cells, and a significant loss of cell viability as measured bytrypan blue exclusion. The ability of the extract to inhibitcholesterol synthesis at lower concentrations without toxicitysuggested that 1 or more enzymes in the cholesterolgenic path-way were inhibited by garlic components.

Earlier studies had suggested that garlic inhibits choles-terol synthesis by reducing HMG-CoA reductase activity(5,8–10,12). To evaluate this possibility, squalene and 2,3-oxidosqualene synthesis was monitored in the presence of garlicextract, with the use of both 14C-acetate and 14C-mevalonate assubstrates. Squalene and 2,3-oxidosqualene are the last 2nonsterol intermediates in the cholesterolgenic pathway; asshown in Figure 2, the labeling of these intermediates was notaffected by treatment with garlic extract. Moreover, cholesterol

760S SUPPLEMENT

by on October 19, 2010

jn.nutrition.orgD

ownloaded from

labeling from 14C-mevalonate was decreased in the presence ofgarlic extract (Fig. 2B), indicating a site of inhibition down-stream of HMG-CoA reductase. Together, these results indicatethat the inhibition of cholesterol synthesis by garlic in theseshort-term studies is mediated at or beyond lanosterol demeth-ylation.

Thin-layer chromatographic analysis of radiolabeled sterolsfrom cells treated with garlic extract revealed the presence ofseveral bands with a mobility characteristic of methylsterols(Fig. 3A). Treatment of cells with ketoconazole, an inhibitor oflanosterol demethylase, caused the accumulation of a prom-inent band (labeled ‘‘a’’ in Fig. 3A), which comigrated withauthentic lanosterol. The identity of this sterol as lanosterolwas confirmed by gas chromatographic analysis of the eluted

sample (Fig. 3B) with comparison with authentic lanosterol,and by subsequent mass spectrometric analysis (data notshown). Although band ‘‘a’’ from the garlic-treated cells has amobility apparently identical with that of lanosterol, gaschromatographic analysis of this product did not reveal thepresence of lanosterol (Fig. 3B, center panel). In this sample, 2prominent peaks were evident, with elution times of 23.23 minand 24.22 min. Mass spectrometric analysis revealed that theearlier peak corresponds most closely to 4-methyllathosterol(Fig. 3C), whereas the latter peak could not be identified.Despite repeated assays, lanosterol could not be shown to bepresent in garlic-treated samples. Gas chromatographic analysisof band ‘‘b’’ from the garlic-treated cells revealed 2 peaks elut-ing at 23.23 min and 24.88 min. The later peak was identifiedby mass spectrometry as 4,49-dimethylzymosterol (Fig. 3D), thefirst substrate for sterol 4a-methyl oxidase. The presence of 24-methyl, 14-demethylated sterols in the garlic-treated cells,and the absence of lanosterol, strongly suggests that garlicextract acts downstream of lanosterol demethylase to inhibitsterol 4a-methyl oxidase.

To identify the active principle(s) of garlic extract, 9 garlic-derived compounds were tested for their ability to inhibitcholesterol synthesis in hepatoma cells. Of these compounds,only diallyl disulfide, diallyl trisulfide, and allyl mercaptan wereeffective inhibitors of cholesterol synthesis at the micromolarlevel, and each yielded an inhibitory pattern identical with thatof garlic extract (Fig. 4). Cholesterol synthesis was similarlyreduced with 14C-mevalonate, and 2,3-oxidosqualene synthesisremained unaffected by these compounds, indicating that allcompounds acted downstream of lanosterol synthesis (data notshown). Gas chromatographic–mass spectrometric analysis ofthe methylsterols that accumulated in the presence of diallyldisulfide yielded results identical with those obtained withgarlic extract. Alliin and allyl methylsulfide were without effecton cholesterol synthesis, whereas S-allylcysteine and S-ethyl-cysteine reduced cholesterol synthesis by 10–20% but requiredconsiderably higher concentrations (4 mmol/L; data not shown).S-methylcysteine and S-propylcysteine were ineffective at con-centrations #4 mmol/L. The estimated IC50 for diallyl disulfidewas 15 mmol/L, with maximal inhibition of ;80% at 200 mmol/L, where toxicity became evident (Fig. 5). The estimated IC50

for diallyl trisulfide was somewhat higher, 40 mmol/L, andmaximal inhibition reached only ;55% because of the ap-pearance of toxicity at 100 mmol/L. Allyl mercaptan was theleast potent inhibitor, with an IC50 of ;200 mmol/L, buttoxicity was not evident at concentrations ;750 mmol/L. Al-though these compounds appeared to fully replicate the in-hibition obtained with garlic extract, the possibility that lessabundant garlic derivatives, including allicin, contribute to theinhibition cannot be excluded. The similarity of the inhibitorypattern with all 3 agents to that of garlic extract argues thatinhibition of sterol 4a-methyl oxidase is the principal mech-anism by which garlic reduces cholesterol synthesis.

DISCUSSION

Early studies on the inhibition of cholesterol synthesis bygarlic indicated that inhibition of HMG-CoA reductase was thelikely mechanism by which garlic acted, on the basis of theobservation that feeding garlic extract to chickens loweredserum cholesterol and reduced hepatic HMG-CoA reductaseactivity (13) and that HMG-CoA reductase could be inhibitedin vitro by garlic or garlic-derived compounds (12). In sub-sequent studies by Gebhardt and et al (5,9), incubation of rathepatocytes and human hepatoma cells with a reconstituted

FIGURE 1 Aqueous garlic extract inhibits cholesterol synthesis.The effect of garlic extract on the incorporation of 14C-acetate intocholesterol in cultured hepatoma cells was monitored by thin-layerchromatography coupled with quantitative radiography. Lactate dehy-drogenase (Lactate DH) activity in the medium and trypan blue exclusionwere used as measures of cellular toxicity. Each value represents themean and standard error of 3 determinations carried out in duplicate.

FIGURE 2 Aqueous garlic extract does not inhibit the synthesis ofnonsterol cholesterol intermediates. The effect of garlic extract on (A) theincorporation of 14C-acetate into squalene and cholesterol and (B) theincorporation of 14C-mevalonate into 2,3-oxidosqualene and cholesterolin cultured hepatoma cells was monitored by thin-layer chromatographycoupled with quantitative radiography. Each value represents the meanand SEM of 3 determinations carried out in duplicate. The labeling of 2,3-oxidosqualene from 14C-acetate was similarly unaffected by garlic extract(data not shown).

761SINHIBITION OF CHOLESTEROL SYNTHESIS BY GARLIC

by on October 19, 2010

jn.nutrition.orgD

ownloaded from

garlic extract inhibited HMG-CoA reductase activity by#23%, although a much greater reduction was seen in cho-lesterol synthesis and was attributed to the inhibition ofadditional downstream enzymes. At very low garlic concentra-tions, only HMG-CoA reductase inhibition appeared relevant,given that the substitution of 14C-mevalonate for 14C-acetatereduced the inhibition of cholesterol synthesis. Similarly, Liuand Yeh (7,8) showed that the inhibition of cholesterol syn-thesis obtained with S-alk(en)yl cysteine compounds, includingS-allylcysteine, could be prevented by substituting 14C-meva-lonate for 14C-acetate and that these compounds loweredHMG-CoA reductase activity by promoting its inactivation byphosphorylation.

Our studies do not reveal an effect of garlic extract onHMG-CoA reductase, although a small reduction in activitycannot be excluded. The lack of change in squalene and 2,3-oxidosqualene labeling over a range of garlic concentrationsthat reduces cholesterol synthesis by #75% argues that in-hibition of 1 or more enzymes downstream of HMG-CoAreductase must predominate at higher concentrations of garlic,a conclusion also reached by Gebhardt (5,9,10). In the studiesof Liu and Yeh (7,8) relatively high concentrations of theS-alk(en)yl cysteines were needed to decrease cholesterolsynthesis; the maximum inhibition achieved with S-allylcys-teine was only 50% at a concentration of 4 mmol/L, yielding anIC50 of 0.61 mmol/L.

FIGURE 3 Radiolabeled methylsterols accumulate in garlic-treated cells. (A) Lipids isolated from hepatoma cells incubated in duplicate with14C-acetate were fractionated by thin-layer chromatography and visualized by autoradiography. Co, lipids from untreated cells; the band labeled ‘‘c’’corresponds to cholesterol, as determined by cochromatography of authentic cholesterol and by coupled gas chromatographic–mass spectrometricanalysis. Ketoconazole, cells incubated in the presence of 10 mmol/L ketoconazole; the band labeled ‘‘a’’ was eluted and further fractionated by gaschromatography, as shown in the upper tracing labeled ‘‘Ketoconazole’’ in panel B. Bands ‘‘a’’ and ‘‘b’’ were eluted from the cells incubated in the presenceof garlic (2.5 g/L) and submitted to gas chromatographic separation, as shown in the middle and lower tracings (labeled ‘‘a’’ and ‘‘b’’) in panel B. (B) Gaschromatographic profiles of bands ‘‘a’’ and ‘‘b’’; sterols were identified by mass spectrometry for the peaks at 24.40 (lanosterol), 23.23 (4-methyllathosterol,shown in panel C), and 24.88 (4,4-dimethylzymosterol, shown in panel D). Other peaks did not correspond to sterols or could not be identified. Theautoradiogram in (A) was merged from 2 images for clarity of presentation.

762S SUPPLEMENT

by on October 19, 2010

jn.nutrition.orgD

ownloaded from

At this concentration we obtained a maximum of 10%inhibition with both 14C-acetate and 14C-mevalonate, arguingagainst a specific effect on HMG-CoA reductase. It should benoted, however, that Liu and Yeh used freshly isolated rat

hepatocytes, whereas we used cultured rat hepatoma cells;differences between hepatocytes and hepatoma cells, as well asthe medium and culture conditions may explain the differentresults.

Gebhardt concluded that higher concentrations of garlicand garlic-derived compounds inhibit lanosterol demethylase(5,9,10) on the basis of the accumulation of a radiolabeledband on thin-layer chromatography with the mobility oflanosterol. Our mass spectrometric analysis of this productand a second radiolabeled band with lower mobility was unableto demonstrate the presence of lanosterol but instead identified4,4-dimethylzymosterol and 4-methyllathosterol, a putativemetabolite of 4-methylzymosterol. 4,4-Dimethylzymosterol and4-methylzymosterol are substrates for sterol 4a-methyl oxidase,an enzyme downstream of lanosterol demethylase that is knownto be sensitive to sulfhydryl reagents (14). Moreover, ketocon-azole, an inhibitor of lanosterol demethylase, yielded a differentpattern of sterol intermediates both in our study (Fig. 3) and inGebhardt’s report (5), lending support to our conclusion thatsterol 4a-methyl oxidase, rather than lanosterol demethylase, isthe most sensitive target of garlic inhibition.

Of 9 garlic-derived organosulfur compounds examined inthe present study, only diallyl disulfide, diallyl trisulfide, andallyl mercaptan were inhibitory to cholesterol synthesis, eachyielding a pattern of sterol accumulation identical with thatobtained with garlic extract. Diallyl disulfide has previouslybeen shown to inhibit HMG-CoA reductase in microsomes(12) and cholesterol synthesis in liver homogenates (4), pri-mary hepatocytes (7,10,15), and hepatoma cells (6); diallyltrisulfide was similarly found to be effective in primary hepa-tocytes (7) and hepatoma cells (6), although cell toxicity wasgenerally greater with the trisulfide, as found in the presentstudy. Allyl mercaptan, the least potent inhibitor in our study,was similarly found to be 10–15% as effective as diallyl disulfidein hepatocyte culture (10,15) and hepatoma cells (6,16). Othergarlic-derived compounds found to be effective inhibitors ofcholesterol synthesis without overt toxicity include ajoene andmethyl ajoene, allicin, 1,3-vinyl dithiin (4,9,10), and someS-alk(en)yl cysteines (7,8,17). Excluding the alk(en)yl cyste-ines and the cyclic 1,3-vinyl dithiin, all the inhibitory com-pounds share an allyl (or vinyl) group adjacent to a disulfide orsulfhydryl group. Garlic compounds found not to be effectiveinhibitors lack this allyl-disulfide or allyl-sulfhydryl group andinclude alliin, S-methylcysteine, methylcysteine sulfoxide,propylcysteine sulfoxide, diallyl sulfide, dipropyl sulfide, andallyl methyl sulfide (4,5,7,9,15). Alliin, allyl methyl sulfide, andseveral alk(en)yl cysteines (S-allylcysteine, S-methylcysteine,S-ethylcysteine, and S-propylcysteine) were shown to be inef-fective in the present study.

The alk(en)yl cysteines S-allylcysteine, S-ethylcysteine, andS-propylcysteine appear to be unique in that they do notconform to the allyl disulfide/sulfhydryl rule. Liu and Yeh (8)have shown that incubation of hepatocytes with S-allyl-,S-ethyl-, and S-propyl-cysteine lowers microsomal HMG-CoAreductase activity by 30–40% without changing enzyme mRNAor protein levels. This lower activity was attributed to anincrease in the amount of phosphorylated (inactivated) enzymein cells incubated with the alk(en)yl cysteines and additionallyto an increase in sulfhydryl oxidation in HMG-CoA reductasein the presence of S-allylcysteine. Although the S-alk(en)ylcysteines did not inhibit cholesterol synthesis in the presentstudy, this laboratory has previously found S-allylcysteine toinhibit squalene monooxygenase, a downstream enzyme in thecholesterolgenic pathway, with an IC50 of 110 mmol/L (18);S-methyl-, S-ethyl-, and S-propyl-cysteine, each of which lacksthe allyl moiety, were not inhibitory to this enzyme.

FIGURE 4 Thin-layer radiochromatographic analysis of cholesterolsynthesis in the presence of garlic-derived compounds. Hepatoma cellswere incubated in duplicate with 14C-acetate in the presence of theindicated garlic compound, and radiolabeled lipids were separated bythin-layer chromatography and visualized and quantified by autoradiog-raphy. Co, untreated cells; DADS, diallyl disulfide; AM, allyl mercaptan;DATS, diallyl trisulfide. The asterisk indicates the position of cholesterol.

FIGURE 5 Inhibition of cholesterol synthesis by garlic derivatives.Hepatoma cells were incubated with the indicated garlic compound, andthe incorporation of 14C-acetate into cholesterol was quantified by thin-layer radiochromatography. Toxicity was monitored by release of lactatedehydrogenase (LDH) into the medium and by trypan blue exclusion(Viability). Each value represents the mean and SEM of 3 determinationscarried out in duplicate.

763SINHIBITION OF CHOLESTEROL SYNTHESIS BY GARLIC

by on October 19, 2010

jn.nutrition.orgD

ownloaded from

Conjugation of the alk(en)yl cysteines to glutamate or acetateto form the g-glutamyl and N-acetyl conjugates reduces theirinhibitory potency (7), suggesting that there is somethingunique about the alk(en)yl cysteines that enhances their abilityto downregulate HMG-CoA reductase activity; nonetheless, itshould be noted that the concentrations of these organosulfurcompounds needed to reduce HMG-CoA reductase activity inhepatocytes by 50% approaches the millimolar range (0.58–0.72 mmol/L).

Inhibition of cholesterol synthesis is thought to be aprincipal mechanism by which garlic lowers blood cholesterol,although other mechanisms may also be important. Indeed,there are very few studies on the effect of garlic on cholesterolsynthesis in whole animals, and those early studies were limitedto documenting a decrease in HMG-CoA reductase activity(13,19). Given that HMG-CoA reductase is down-regulated byisoprenoid and sterol intermediates (20–22), it can be expectedthat inhibition of a downstream cholesterolgenic enzyme willresult in the accumulation of 1 or more intermediates that may

feed back to decrease HMG-CoA reductase activity. Ourconclusion that garlic inhibits sterol 4a-methyl oxidase is inaccord with this view, given that 4-dimethylated sterols,including lanosterol and dimethylzymosterol, have been shownto strongly promote the degradation of HMG-CoA reductasevia an Insig-mediated pathway (22) (Fig. 6). Further studies areneeded to determine whether garlic-derived organosulfur com-pounds inhibit purified sterol 4a-methyl oxidase and whethergarlic effectively inhibits this enzyme in vivo.

LITERATURE CITED

1. Alder R, Lookinland S, Berry JA, Williams M. A systematic review of theeffectiveness of garlic as an anti-hyperlipidemic agent. J Am Acad Nurse Pract.2003;15:120–9.

2. Ackermann RT, Mulrow CD, Ramirez G, Gardner CD, Morbidoni L,Lawrence VA. Garlic shows promise for improving some cardiovascular risk fac-tors. Arch Intern Med. 2001;161:813–24.

3. Stevinson C, Pittler MH, Ernst E. Garlic for treating hypercholesterolemia:a meta-analysis of randomized clinical trials. Ann Intern Med. 2000;133:420–9.

4. Sendl A, Schliack M, Loser R, Stanislaus F, Wagner H. Inhibition ofcholesterol synthesis in vitro by extracts and isolated compounds prepared fromgarlic and wild garlic. Atherosclerosis. 1992;94:79–85.

5. Gebhardt R. Multiple inhibitory effects of garlic extracts on cholesterol bio-synthesis in hepatocytes. Lipids. 1993;28:613–9.

6. Cho BH, Xu S. Effects of allyl mercaptan and various allium-derivedcompounds on cholesterol synthesis and secretion in Hep-G2 cells. Comp Bio-chem Physiol C Toxicol Pharmacol. 2000;126:195–201.

7. Liu L, Yeh YY. Inhibition of cholesterol biosynthesis by organosulfurcompounds derived from garlic. Lipids. 2000;35:197–203.

8. Liu L, Yeh YY. S-alk(en)yl cysteines of garlic inhibit cholesterol synthesisby deactivating HMG-CoA reductase in cultured rat hepatocytes. J Nutr. 2002;132:1129–34.

9. Gebhardt R, Beck H, Wagner KG. Inhibition of cholesterol biosynthesis byallicin and ajoene in rat hepatocytes and HepG2 cells. Biochim Biophys Acta.1994;1213:57–62.

10. Gebhardt R, Beck H. Differential inhibitory effects of garlic-derivedorganosulfur compounds on cholesterol biosynthesis in primary rat hepatocytecultures. Lipids. 1996;31:1269–76.

11. Laden BP, Tang Y, Porter TD. Cloning, heterologous expression, andenzymological characterization of human squalene monooxygenase. Arch Bio-chem Biophys. 2000;374:381–8.

12. Kumar RV, Banerji A, Kurup CK, Ramasarma T. The nature of inhibition of3-hydroxy-3-methylglutaryl CoA reductase by garlic-derived diallyl disulfide. BiochimBiophys Acta. 1991;1078:219–25.

13. Qureshi AA, Abuirmeileh N, Din ZZ, Elson CE, Burger WC. Inhibition ofcholesterol and fatty acid biosynthesis in liver enzymes and chicken hepatocytesby polar fractions of garlic. Lipids. 1983;18:343–8.

14. Miller WL, Gaylor JL. Investigation of the component reactions of oxidativesterol demethylation: oxidation of a 4,4-dimethyl sterol to a 4 beta-methyl-4 alpha-carboxylic acid during cholesterol biosynthesis. J Biol Chem. 1970;245:5375–81.

15. Gebhardt R. Amplification of palmitate-induced inhibition of cholesterolbiosynthesis in cultured rat hepatocytes by garlic-derived organosulfur com-pounds. Phytomedicine. 1995;2:29–34.

16. Xu S, Cho BH. Allyl mercaptan, a major metabolite of garlic compounds,reduces cellular cholesterol synthesis and its secretion in Hep-G2 cells. J Nutr Bio-chem. 1999;10:654–9.

17. Yeh YY, Yeh SM. Garlic reduces plasma lipids by inhibiting hepaticcholesterol and triacylglycerol synthesis. Lipids. 1994;29:189–93.

18. Gupta N, Porter TD. Garlic and garlic-derived compounds inhibit humansqualene monooxygenase. J Nutr. 2001;131:1662–7.

19. Qureshi AA, Crenshaw TD, Abuirmeileh N, Peterson DM, Elson CE.Influence of minor plant constituents on porcine hepatic lipid metabolism: impacton serum lipids. Atherosclerosis. 1987;64:109–15.

20. Roitelman J, Simoni RD. Distinct sterol and nonsterol signals for theregulated degradation of 3-hydroxy-3-methylglutaryl-CoA reductase. J Biol Chem.1992;267:25264–73.

21. Sever N, Song BL, Yabe D, Goldstein JL, Brown MS, DeBose-Boyd RA.Insig-dependent ubiquitination and degradation of mammalian 3-hydroxy-3-methylglutaryl-CoA reductase stimulated by sterols and geranylgeraniol. J BiolChem. 2003;278:52479–90.

22. Song BL, Javitt NB, DeBose-Boyd RA. Insig-mediated degradation ofHMG CoA reductase stimulated by lanosterol, an intermediate in the synthesis ofcholesterol. Cell Metab. 2005;1:179–89.

FIGURE 6 Feedback inhibition in the cholesterolgenic pathway bygarlic derivatives. A mechanism by which garlic compounds may down-regulate HMG-CoA reductase through inhibition of sterol 4a-methyl oxidaseis illustrated. Only selected cholesterol intermediates are shown.

764S SUPPLEMENT

by on October 19, 2010

jn.nutrition.orgD

ownloaded from