Journal of Clinical InvestigationVol. 41, No. 5, 1962

THE METABOLISMOF DESMOSTEROLIN HUMANSUBJECTSDURINGTRIPARANOLADMINISTRATION *

BY DEWITT S. GOODMAN,JOEL AVIGAN AND HILDEGARDWILSON

(From the Section on Metabolism, National Heart Institute, and the National Institute ofArthritis and Metabolic Diseases, Bethesda, Md.)

(Submitted for publication October 25, 1961; accepted January 25, 1962)

Recent studies with triparanol (1-[p-,3-diethyl-aminoethoxyphenyl ]-1- (p-tolyl) -2- (p-chloro-phenyl)ethanol) have demonstrated that this com-pound inhibits cholesterol biosynthesis by blockingthe reduction of 24-dehydrocholesterol (desmos-terol) to cholesterol (2-4). Administration oftriparanol to laboratory animals and to man re-sults in the accumulation of desmosterol in theplasma and tissues, usually with some concom-itant lowering of the plasma total sterol concen-tration (2, 5). In studies with human subjectsit was observed that after about 2 weeks the des-mosterol concentration tended to plateau at anaverage level of 27 per cent of the total circulatingsterols (5). The ratios of desmosterol to cho-lesterol in animal tissues (2) differed somewhatfrom the ratios in plasma. In addition, injection ofthe cholesterol precursor 2-C14-mevalonic acid dur-ing triparanol administration resulted in the rapidappearance of labeled desmosterol in blood, whileno significant radioactivity appeared in cholesterolin the first few days (5).

The object of the present experiments was tostudy the metabolism of desmosterol in patientsreceiving triparanol. Since cholesterol serves asthe biosynthetic precursor for several physiologi-cally important compounds, including bile acidsand steroid hormones, the question arose whetherdesmosterol could also function as a precursor forthese compounds without first being transformedinto cholesterol. The studies reported herein dem-onstrate the direct conversion of desmosterol toboth bile acids and to steroid hormones, and alsoprovide information about the over-all metabolismof this sterol in the triparanol-treated man.

EXPERIMENTAL

Clinical material. Both of the patients studied werehospitalized on a metabolic ward of the Clinical Center.

* Presented in part at the First International Pharma-cology Meeting, Stockholm, Sweden, August, 1961 (1).

Patient G.B. was a 55 year old man with known arterio-sclerotic heart disease and mild hypercholesterolemia;since 1957 he had maintained a satisfactory and stablecardiac status. At the time of the present study he hadbeen taking 250 mg triparanol daily for 4 weeks, and hada total serum sterol level in the high normal range.Patient F.A. was a 40 year old man with a 4- to 5-yearhistory of gout and essential hyperlipemia. At the timeof this study he had been on an isocaloric low purine dietfor several weeks, and both the gout and hyperlipemiawere in remission. He received 250 mg triparanol dailyfor 2 weeks prior to the start of this study.

Design of study. Each study was begun in the fastingpatient by the rapid injection of 20 /Ac of 2-C14-mevalonicacid, in isotonic saline, into an antecubital vein. PatientF.A. also received, in the same injection, 20 jsc of 7-a H'-cholesterol, previously incorporated into his own serumlipoproteins by the method of Avigan (6). Blood sam-ples were collected 3 hours later from the opposite ante-cubital vein, and at varying intervals thereafter. Weob-tained bile in the early mornings by duodenal intubation,employing intravenous cholecystokinin when necessary,to ensure an adequate sample, and also collected 24-hoururine samples, using small amounts of toluene as pre-servative.

Serum sterols. Sera were extracted with 25 vol ofalcohol: acetone, 1: 1 (vol/vol), and the free sterols wereprecipitated from these extracts as their digitonide com-plexes. After collection of the digitonide precipitates, thesupernatants were saponified with ethanolic KOH, neu-tralized with acetic acid, and the hydrolyzed ester sterolswere precipitated as digitonides. The sterol digitonidesthus derived from the free and ester sterol fractionswere then washed, cleaved with pyridine, and the sterolsextracted, washed, and dried as described by Steinberg,Avigan and Feigelson (5). The sterol fractions so ob-tained, consisting of mixtures of desmosterol and choles-terol, will subsequently be referred to as unfractionatedfree or ester sterols. Small portions of the unfractionatedsterols were used for the measurement of cholesterol anddesmosterol in the sample, by the differential colori-metric method of Avigan and co-workers (2). Otherportions of the sterols were directly assayed for radio-activity, and the specific radioactivities were determined.The remaining unfractionated sterols of each samplewere separated into cholesterol and desmosterol by chro-matography of their p-phenylazobenzoyl esters on col-umns of silicic acid: Celite, 2: 1 (2). The purified cho-lesterol and desmosterol fractions were then individually

962

DESMOSTEROLMETABOLISMDURING TRIPARANOL

assayed for radioactivity and were also colorimetricallyanalyzed for cholesterol and desmosterol. Cholesterolwas usually obtained free of desmosterol, but the des-mosterol samples were generally contaminated with smallamounts of cholesterol. As discussed elsewhere, how-ever (5), by knowing the analytical values for the puri-fied sterols in the samples counted, it was possible tocalculate with precision the individual specific radioac-tivities of cholesterol and desmosterol, despite the con-

tamination of the desmosterol samples with small amountsof cholesterol. An internal check on the accuracy ofthese determinations was provided by the specific radio-activity measurements of the unfractionated sterols. Thecalculated and measured values for specific radioactivityof the unfractionated sterols agreed closely in all cases.

Bile. Bile samples were extracted with 25 vol ofethanol. Each ethanol extract was evaporated nearly todryness and saponified with NaOHunder pressure. Theunsaponifiable fraction was then extracted with ethylether, followed by acidification of the sample and extrac-tion of the acidic fraction with ethyl ether. Cholic acidwas purified from the acidic fraction of the bile of Pa-tient G.B. by reversed-phase partition chromatography,with methanol-water as moving phase and isooctanol-chloroform as stationary phase, as described by Norman(7) and Sjovall (8). A mixture of chenodeoxy- anddeoxycholic acids was purified from the acidic fractionsof the bile samples of Patient F.A. by partition chroma-tography as described by Mosbach, Zomzely and Ken-dall (9), with acetic acid on Celite as stationary phaseand isopropyl ether-petroleum ether mixtures as mobilephase. The amount of purified bile acid obtained was de-termined by weighing the material after crystallizationand drying; verification of these values was secured byspectrophotometric analyses of aliquots of the samples bythe method of Mosbach, Kalinsky, Halpern and Kendall(10). The specific radioactivity of the bile acids was

obtained by radioassay of a measured aliquot of thepurified sample.

In some samples sterols were precipitated from theunsaponifiable fraction of bile as their digitonide com-

plexes, analyzed colorimetrically, and counted to deter-mine specific radioactivity.

Urinary steroids.' Ether extracts of the neutral ster-

1 Trivial names of the steroids used in this paper fol-low. Etiocholanetrione: etiocholane-3,11,17-trione. THF:3a,llp,17a,21-tetrahydroxypregnane-20-one. THE: 3a,-17a,21-trihydroxypregnane-1 1,20-dione. I1-Ketoetiocho-lanolone: 3a-hydroxyetiocholane-11,17-dione. 11- Hy-droxyetiocholanolone: 3a,11,8-dihydroxyetiocholane-17-one.11-Hydroxyandrosterone: 3a,11,B-dihydroxyandrostane-17-one. Etiocholanolone: 3a - hydroxyetiocholane - 17-one .

Androsterone: 3a-hydroxyandrostane-1 7-one. Adrenoster-one: A' - androstene - 3,11,17 - trione. Androstanediol: andro-stane-3a,17a-diol, dihydroandrosterone. Cortisol: 1 lp,17a,-21-trihydroxy-4-pregnene-3,20-dione. Corticosterone: 118,-21 -dihydroxy-4 -pregnene-3,20-dione. Etiocholanedione:etiocholane-3,17-dione. Androstanedione: androstane-3,17-dione.

oids were obtained from the 24-hour urine samples byincubation of the urine with p-glucuronidase, followed byacidification, continuous ether extraction for 72 hours,and appropriate washings of the ether extract to re-move acids and phenolic compounds (11, 12). In thefirst study (Patient G.B.), the neutral extract was thendirectly oxidized with chromic acid (CrO3) solution bya slightly modified procedure of Wotiz, Lemon andMarcus (13). The CrO3 oxidation converts most corti-sol metabolites to etiocholanetrione; the other metabo-lites giving significant amounts of this product are theC1903 steroids (e.g., 1 1-hydroxyetiocholanolone). Thisprocedure, consequently, provided enough of a singlecompound for accurate determination of specific radio-activity. The excess CrO3 was reduced with 0.2 percent NaHSO3and the steroids extracted with methylenechloride. This extract was washed with 0.1 N NaOHand with water, and then prepared for paper chromatog-raphy.

In the second study (Patient F.A.) the entire extractof each urine was first subjected to partition columnchromatography as described elsewhere (11), with acolumn of 15 g aluminum silicate and 50 per centethanol as stationary phase. The elution sequence wassimplied so that only two fractions were collected: 1)200 ml 15 per cent chloroform in hexane; and 2) 260ml 30 per cent chloroform in hexane, followed by 120 ml50 per cent chloroform in hexane. The first fractioncontained all C,,O2 and C,903 (and any C2,04) steroids;the second fraction contained all C2105 and C210A metabo-lites present (i.e., all the cortisol metabolites and littleor nothing else). Paper chromatograms run on aliquotsof both fractions showed that in all cases the column hadeffected the desired separation. These spot tests alsoproved that the patient was excreting the normal ma-jor steroid metabolites in his urine. Thus the cortisolmetabolite fraction included both THF and THE; theC390, steroids present included 11-ketoetiocholanolone, 11-hydroxyetiocholanolone (in lesser amounts), and 11-hy-droxyandrosterone; the C1902 steroids were representedby a band corresponding to etiocholanolone and andros-terone (which would not separate in the system used).

Each column fraction from the urine samples of Pa-tient F.A. was then subjected to chromic acid oxidation,as described above, followed by paper chromatography ofthe oxidized steroids. The oxidized steroid fractions, ex-cept the C,,03 + C1903 fractions of F.A., were chromato-graphed on paper with a Bush "A2" system (ligroin:methanol: water, 100: 70: 30) for 6 hours (14). Teststrips were analyzed by the Zimmermann reaction for 17-ketosteroids. The oxidized extracts showed a strongband corresponding to etiocholanetrione, and a weakerband corresponding to adrenosterone. The etiocholane-trione bands were cut out and eluted with ethanol by de-scending capillary flow. An aliquot was assayed by themicro-Zimmermann reaction (15), and the remainderwas used for radioassay.

The oxidized C190, + C,,03 fractions of F.A. werechromatographed on paper with the Bush "Al" system

963

DEWITT S. GOODMAN,JOEL AVIGAN AND HILDEGARDWILSON

TABLE I

Specific radioactivity of serum sterols and of sterol metabolitesafter C14-mevalonate, during treatment with triparanol,

in Patient G.B.

SpecificSample radio

Compound analyzed obtained at: activity

cpm/,JmoleSerum free desmosterol 3 hrs 452Serum free cholesterol 3 hrs 0Serum unfract. free sterols 3 hrs 154

Serum free desmosterol 3 days 99.5Serum free cholesterol 3 days 5.1Serum unfract. free sterols 3 days 33.6

Red cells, unfract. freesterols 3 days 36.4

Cholic acid from bile 3 days 30.1Cortisol derivs. from urine 3-4 days 37.7

(ligroin: methanol: water, 100: 90:10). These chromato-grams showed bands corresponding to etiocholanetrione(from C,,O, metabolites) and also strong, less polarbands containing mixtures of androstanedione (oxidationproduct of androsterone and androstanediol) and etiocho-lanedione (oxidation product of etiocholanolone and etio-cholanediol); both "diones" (to be referred to as "an-drogen derivatives") are derived from urinary C,,O, ster-oids and do not separate in the chromatographic systemused. These bands were eluted, assayed against an an-

drostanedione reference standard (the chromogenic valueof etiocholanedione is very close to that of androstane-dione), and radioassayed.

Radioassay was performed with a Packard liquid scintil-lation spectrometer. Sterols and oxidized steroids were

dissolved in 15 ml 0.5 per cent diphenyloxazole (DPO)in toluene for counting; bile acids were first dissolved in1 ml p-dioxane, followed by the addition of 15 ml DPOin toluene. All samples from F.A., together with ab-solute C1' and H' standards, were simultaneously countedfor C" and H' by using appropriate settings for the dis-criminators of the two channels. The quenching of eachisotope was individually corrected by counting the samplesagain after the addition of a H' internal standard, andthen after addition of a C" internal standard. Countingefficiencies for the double-isotope procedure were 35 per

cent for C" and 18 per cent for H'; when C" was

counted alone, the efficiency was 52 per cent.Materials. DL-2-C"-mevalonic acid (specific activity,

2.7 mc per mmole) was obtained from the Volk Radio-chemical Corp. 7-a-H'-cholesterol (specific activity, 5.4mc per mmole) was a product of the New England Nu-clear Corp. Pure desmosterol, for colorimetric stand-ard, was isolated from livers of rats fed triparanol, as

previously described (3). Cholic and deoxycholic acids,for use as analytic and chromatographic standards, were

obtained from Nutritional Biochemicals Corp. The refer-ence steroids were obtained from USP reference stand-ards. Cholecystokinin (Cecekin) was obtained from the

Vitrum Pharmaceutical Co. of Stockholm. The isotope-containing solutions for intravenous injection were pre-pared by the Pharmacy Department of the Clinical Cen-ter of the National Institutes of Health.

Correction of results. Cholesterol molecules synthe-sized from 2-C"-mevalonate contain five labeled carbonatoms: carbon atoms 1, 7, and 15 in the sterol nucleus andcarbon atoms 22 and 27 in the side chain (16, 17). Des-mosterol synthesized from 2-C"-mevalonate should havethe same pattern of labeling. On conversion of this kindof labeled cholesterol, or desmosterol, to bile acids, theone labeled carbon atom in the terminal isopropyl group(carbon atom 27) is lost. Similarly, two of the five la-beled atoms (nos. 22 and 27) are lost during the removalof the terminal six carbons of the side chain, en route tosteroid hormones (18). The results that follow havebeen corrected for the above losses by multiplying themeasured bile acid C1' specific radioactivities by 5/4, andthe corresponding steroid C1' specific activities by 5/3.

RESULTS

For Patient G.B. the specific activities of the se-rum sterols were determined after 3 hours and after3 days. Table I demonstrates that at 3 hours therewas considerable radioactivity in serum free des-mosterol (452 cpm per pmole) but no measurableradioactivity in cholesterol. After 3 days thespecific activity of desmosterol had decreased to99.5 cpm per Atmole and a little radioactivity (5.1cpm per umole) was detectable in cholesterol. Thespecific activity of the unfractionated sterols at both3 hours and 3 days reflects the fact that in this pa-tient about one-third of the serum sterols was des-mosterol and two-thirds cholesterol.

The ratio of desmosterol to cholesterol in theunfractionated red cell sterols obtained after 3 dayswas very similar to that in plasma. This fact,plus the closeness of the specific radioactivities ofthe unfractionated free sterols of serum and redcells (see Table I), suggests that desmosterol andcholesterol of serum and red cells were in isotopicequilibrium.

After 3 days, bile and urine were collected, fromwhich cholic acid and the cortisol derivatives (plusC190, derivatives, as etiocholanetrione: vide supra)were respectively isolated. Both cholic acid andthe cortisol derivatives had specific radioactivitymuch greater than that of serum cholesterol (seeTable I), demonstrating that a significant part ofthese materials must have been synthesized di-rectly from desmosterol.

The second study, with Patient F.A. as subject,

964

DESMOSTEROLMETABOLISMDURING TRIPARANOL

was more elaborate in that samples were collectedat 1- or 2-day intervals for a week. In addition,as already mentioned, H3-labeled cholesterol wasinjected simultaneously with the C14-mevalonate.For the sake of clarity, however, the C"A data fromthis study will initially be presented alone.

Figure 1 shows the C14 specific radioactivity ofthe serum free and esterified desmosterol and un-fractionated sterols. The C14 specific radioactivityof all cholesterol samples was below 10, and henceis not shown in this semilogarithmic plot. The C"Aspecific radioactivity of cholesterol ranged from6.2 at 3 hours to 7.7 cpm per umole at 7 days forfree cholesterol, and from 2.3 to 7.3 cpm per umolefor esterified cholesterol. These low values con-stituted, however, a significant portion of the totalradioactivity by 7 days. The ratio of desmosterolto cholesterol was remarkably constant during thestudy period; desmosterol comprised 32 to 34per cent of the total serum sterols in all sixsamples.

Figure 2 shows the corrected C14 specific activityof the bile acids from Patient F.A., compared withthe specific activity curve of serum free desmos-

600 _500

400

300 0 0 FREE0 A~~~t A ESTER

200

a:~~~'100 U DESMOSTEROL

70 NO60-

50'AA4040

a. 30( OuNFRAcTN TED

. v 1 2 3 4 5 6 7

DAYS

FIG. 1. C' SPECIFIC RADIOACTIVITY OF FREE ANDESTERI-

FIED STEROLS, DURING TRIPARANOL ADMINISTRATION, IN

THE SERUMOF PATIENT F.A. AFTER INJECTION OF 2-C -

MEVALONATE.

w

J

0

CL

i-

nw

a-

co)

700600500

400

300

200 _

100

706050

40

30

1 2 3 4 5 6DAYS

FIG. 2. C' SPECIFIC RADIOACTIVITY OF THE BILE ACIDS

OF PATIENT F.A. DURING TRIPARANOL ADMINISTRATION,

AFTER INJECTION OF 2-C -MEVALONATE.

terol. Because of a laboratory accident the cholicacid samples could not be obtained, and mixturesof chenodeoxy- and deoxycholic acids were puri-fied and analyzed instead. It is evident that,throughout the period of study, the bile acid C14specific radioactivity was much greater than that ofserum cholesterol. This indicates that at least partof these acids must have been made directly fromdesmosterol.

Figure 3 shows similar data for the cortisolderivatives, analyzed as etiocholanetrione, and forthe androgen derivatives, analyzed as "androstane-dione," isolated from the urine samples of Pa-tient F.A. The C14 specific radioactivity of bothclasses of steroids was much higher than that ofcholesterol throughout the study, demonstratingthat part of these steroids must have been madedirectly from desmosterol.

Let us now consider the double-isotope dataobtained from Patient F.A. Figure 4 againshows the C14 specific radiactivity curves for se-

rum free and esterified desmosterol, and alsoshows the H3 specific radioactivity curves for freeand esterified cholesterol isolated from the same

serum samples. The disappearance rates of cho-lesterol and desmosterol are fairly similar, desmos-terol disappearance being a little more rapid thanthat of cholesterol.

*-0 BILE ACIDS (MIXTURECHENODEOXY-ANDDEOXY-CHOLIC ACIDS)

-- - REFERENCECURVE:SERUMFREEDESMOSTEROL

(SE RUM CHOLESTEROL:ALL<10)I

965

DEWITT S. GOODMAN,JOEL AVIGAN AND HILDEGARDWILSON

t00_loo

00 _

00 -x *-* CORT/SOL DERIV'S

00O _ \ ^ANDROGENDER/V'S--- REFERENCECURVE:

SERUM FREEDESMOSTEROL00

00

7060_-50 -...

40- z

30(SERUM CHOLESTEROL:ALL<10)

201 2 3 4 5 6

DAYS

FIG. 3. C' SPECIFIC RADIOACTIVITY OF THE URINARY

STEROID DERIVATIVES OF PATIENT F.A. DURING TRAPARANOL

ADMINISTRATION, AFTER INJECTION OF 2-C'-MEVALONATE.

From these data information can be derivedabout the total metabolism of desmosterol in thispatient. For this purpose it is necessary to as-

sume that the mass ratio of cholesterol to desmos-

700600500 DESMOSTEROL(C'4)

400 \ O-O FREEA ---- ESTER

J 300 - CHOLESTEROL(H3)0

0*---- FREE200 A-A ESTER

RU -.ATETFADRN RPRNO DIITA

z~~~~~~~~~~z

70-

60-

50-

iC. 40

5w 30Q0.

2012 3 4 5 6

DAYS

FIG. 4. SPECIFIC RADIOACTIVITY OF FREE ANDESTERIFIED

DESMOSTEROL(C") AND CHOLESTEROL(H') IN THE SE-

RUMOF PATIENT F.A. DURING TRIPARANOL ADMINISTRA-

TION, AFTER INJECTION OF 2-C'-MEVALONATE AND 7-H'-CHOLESTEROL. In cpm of either C" or H' per usmole sterol.

terol in the entire metabolically active pool (i.e.,the plasma-red cell-liver pool) was similar tothat in plasma. This assumption is reasonable,since this ratio was very similar in red cells andplasma, and since even a moderate difference inthe liver should have only a small effect on thesterol ratio of the entire pool. By then using thethree sets of measurements-1) the rate of disap-pearance of C14 in desmosterol, 2) the rate of dis-appearance of H3 in cholesterol, and 3) the rateof appearance of C14 in cholesterol-a calculationwas made of the fraction of the disappearing OC-desmosterol that reappeared as C14-cholesterol.2The calculation showed that in this patient 21 ± 4per cent (SE of mean) of the desmosterol whichdisappeared from the metabolically active pool re-

TABLE II

Specific radioactivity ratios of H3 to C'4 (cpm/,umole), afterC14-mevalonate and H3-cholesterol during triparanol

therapy in Patient F.A.

Serum unfrac- Urinary steroidstionated sterols

Sample Cortisol Androgencollected at: Free Ester derivs. derivs.

3 hrs 2.64 2.981 day 1.74 1.75 0.79 1.092 days 2.27 1.74 0.67 0.783 days 2.12 2.125 days 2.41 1.71 0.97 0.737 days 2.40 2.07 0.91 0.87

appeared therein as cholesterol. More than three-fourths of desmosterol disappearance thereforeoccurred via other metabolic pathways.3

The use of H3-cholesterol together with C14-mevalonate also provided additional informationabout the relationship of the serum sterols to the

2 The difference between the disappearance rates of C'and H' in serum cholesterol represented the rate of syn-thesis of C"-cholesterol from C"4-desmosterol. The dis-appearance rate of C"4-desmosterol (in cpm per mg), mul-tiplied by the desmosterol: cholesterol mass ratio, gavethe maximum possible increment in C"4-cholesterol spe-cific radioactivity, should cholesterol be the only endproduct of desmosterol metabolism. Comparison of theobserved and maximum possible C"-cholesterol incre-ments then yielded the fraction of disappearing desmosterolthat reappeared as cholesterol in the metabolically activepool.

SThis is the maximum possible value for desmosteroldisappearance via noncholesterol pathways, since some ofthe disappearing desmosterol might have been convertedto cholesterol in tissue pools that are not in equilibriumwith the serum cholesterol. Such conversion probablyoccurs only to a minor extent.

966

716

4

3

21

Id

w-J0

a.*0It

0-

U

UE

wa.V,

DESMOSTEROLMETABOLISMDURING TRIPARANOL

urinary steroid metabolites in Patient F.A. TableII presents the ratios of H3 to C'4, in counts perminute per micromole for each isotope, observedin the unfractionated serum sterols and in the uri-nary steroid derivatives. Both the cortisol andthe androgen derivatives contained significantamounts of both HM and C14 throughout the study.This indicates that both the serum cholesterol anddesmosterol pools served as precursors for the twoclasses of steroids. The ratio of H3 to C14 wasdecidedly lower in the steroid derivatives than inthe unfractionated serum sterols. As expected,only minute amounts of H3 were found in the bileacids from F.A. This agrees with the finding ofBergstrom and Lindstedt and their associates that7-a-hydroxylation is an integral part of bile acidbiosynthesis and that the formation of deoxycholicacid is due to subsequent bacterial action on cholicacid in the intestinal lumen (19, 20). The min-ute amounts of H3 (less than 15 per cent of theC14 cpm per ,umole) observed in the bile acid sam-ples might reflect some slight isotopic nonhomo-geneity in the labeling of the 7-a-H3-cholesterolused.

DISCUSSION

The present study has demonstrated the directconversion of desmosterol into bile acids and intosteroid hormones during triparanol therapy. Thisconclusion can be drawn with a high degree ofconfidence from the finding that the C14 specificradioactivity of these two classes of metabolites,after C'4-mevalonate injection, was considerablyhigher than that of the serum cholesterol.

Two possible criticisms may be raised here.First, was the radioactivity observed in the bileacids and urinary steroids solely derived fromsterol precursors? Under normal conditions thenatural D-isomer of mevalonic acid is rapidly andefficiently converted into cholesterol (21), and thelatter serves as the normal precursor of the me-tabolites considered here. The products of theother quantitatively important pathway of mevalo-nate metabolism, prenoic acids, do not give riseto the steroidal ring structure (22). There is nobasis for an assumption that treatment with tri-paranol should introduce new metabolic pathwaysfor mevalonic acid, particularly since in these pa-tients labeled mevalonate was rapidly convertedinto a sterol (desmosterol). Second, to what ex-

tent does the specific radioactivity of the serumsterols reflect that of the precursor sterols in theorgans synthesizing the metabolites under study?For the case of bile acid synthesis the serum freecholesterol has been shown to closely reflect theliver precursor sterol pool. Thus it is known thatthe liver free cholesterol pool is in rapid isotopicequilibrium with serum free cholesterol, and thatthe bile acids derived from the former are in rapidisotopic equilibrium with both (23). Further-more, in the studies with Patient F.A. it was foundthat the specific radioactivities of both desmosteroland cholesterol isolated from bile were fairly closeto the corresponding specific radioactivities of theserum sterols. In these samples desmosterol rep-resented an average of 41 per cent of the totalbiliary sterols. It is thus certain that most of theradioactivity observed in the bile acids was di-rectly derived from desmosterol. It is also pos-sible that during the early hours of the studies la-beled bile acids were formed from sterol precur-sors of desmosterol; no evidence is available onthis subject. It is exceedingly unlikely, however,that any significant part of the radioactivity of thebile acids was derived from cholesterol, since thiswould require a precursor cholesterol pool ofmuch higher specific radioactivity than the en-tire liver cholesterol pool, contrary to the studiesquoted above. This would, moreover, require theexistence of a special fraction of liver desmosterolreductase which was much less inhibited by tri-paranol than was the main body of this enzyme.Studies with animals and with liver homogenatesdo not support this possibility (2, 3).

Comparable detailed information about theadrenal (and testicular) sterol pools is not avail-able for man. Animal studies have indicated thatguinea pig adrenal gland cholesterol is never inisotopic equilibrium with plasma cholesterol (24),probably because of considerable endogenousadrenal sterol biosynthesis, coupled with slowequilibration. Urinary cortisol was, however, inisotopic equilibrium with adrenal cholesterol (24).In contrast, adrenal gland and plasma cholesteroldid reach isotopic equilibrium in the rat (25). Inthe present studies, therefore, the plasma sterolspecific radioactivities cannot be assumed to reflectclosely those of the precursor adrenal sterol pools.The data in Table II do, in fact, suggest that theplasma and adrenal precursor sterol pools might

967

DEWITT S. GOODMAN,JOEL AVIGAN AND HILDEGARDWILSON

have been different during the period of study withPatient F.A. This derives from the persistentlylarge difference in the H3 to C14 ratio of the uri-nary steroids and that of the plasma unfractionatedsterols. One explanation for this difference is thepossibility that there was significant endogenousadrenal sterol biosynthesis from C14-mevalonateduring the early hours of the study. If there werealso a preference for endogenously synthesizedsterol for the synthesis of steroid hormones, orvery poor mixing of the endogenous and plasmasterol pools, or both, the observed lower H3 to C14ratio in the urinary steroids would result. Thereare, of course, other possible explanations. Thus,if desmosterol were a better steroid precursor thanis cholesterol, the lower ratio would result. Inaddition, a variety of other factors might be in-volved, such as different ratios of desmosterol tocholesterol in the different pools, variable rates ofequilibration between the sterol pools, or a pos-sible loss of H3 occurring during the conversion of7-H3-cholesterol to steroids because of labilizationof the hydrogens at carbon 7 during the doublebond shift from 5,6 to 4,5.

Despite the absence of detailed informationabout these factors, the data presented here pro-vide strong evidence that desmosterol (perhapstogether with its precursor sterols) can effectivelyserve as a direct precursor for steroid hormonesynthesis. This derives from the fact that theurinary steroid C14 specific radioactivity was verymuch higher than that of plasma cholesterolthroughout the study. In order for C14 to appearin the steroids via cholesterol it would be neces-sary for the adrenal precursor cholesterol pool tohave had a much higher C14 specific radioactivitythan the corresponding plasma pool had, particu-larly early in the study. This in turn would haverequired a lack of inhibition of adrenal desmosterolreductase by triparanol, in order to permit theearly formation of C14-cholesterol from mevalo-nate. Animal studies (unpublished) have shownthat desmosterol accumulation in adrenal glandsis comparable with its accumulation in other tis-sues, indicating an inhibition of adrenal des-mosterol reductase comparable with that foundelsewhere. It is hence highly unlikely that theC14-labeled steroids were synthesized via choles-terol. Since desmosterol is the major labeledadrenal sterol after C'4-mevalonate, it is probable

that most of the C14 steroids were synthesized di-rectly from desmosterol.

An approximate comparison of the abilities ofdesmosterol and cholesterol to serve as precursorsfor bile acids can be obtained from the specificradioactivity data. In the case of Patient G.B.the specific radioactivity of cholic acid was closeto that of the serum (and red cell) unfractionatedfree sterols. This suggests that desmosterol andcholesterol were fairly similar in their ability toserve as precursors. In the case of Patient F.A.,the C14 specific radioactivity of the bile acids was,in fact, greater than that of the serum unfraction-ated sterols throughout most of the study. It isthus clear that desmosterol is at least comparablewith cholesterol in its ability to serve as a precursorof bile acids.

This conclusion is also consonant with the dataon the over-all metabolism of desmosterol in Pa-tient F.A., indicating that approximately three-fourths of desmosterol disappeared by routes otherthan conversion to cholesterol. It is almost cer-tain that these routes were the same ones normallyinvolved in cholesterol metabolism, with conver-sion of sterol to bile acids and excretion of sterolin the feces being the quantitatively importantprocesses.

Animal studies support this conclusion. Blohm,Kariga and Mackenzie (26) found that, after theinjection of C14-mevalonate into triparanol-treatedrats, a considerable part of the radioactivity ad-ministered appeared in the feces as bile acids.Furthermore, in vitro studies from this laboratory(to be published in detail elsewhere) have directlydemonstrated that desmosterol and cholesterol arecomparable substrates for oxidation by liver mito-chondria from triparanol-treated mice.

Because of the uncertainty about the relationsof the various pools involved in steroid biosyn-thesis, only a less satisfactory estimate can be madeof the comparative abilities of the two sterols toserve as precursors of steroid hormones. Thedata do suggest, however, that the two sterols areprobably comparable, since in both patients thesteroid specific activities were similar to those ofthe bile acids and the unfractionated plasma sterols.At any rate, it is reasonably clear that desmosterolcan serve as an effective precursor for steroidbiosynthesis.

Since desmosterol is an effective precursor for

968

DESMOSTEROLMETABOLISMDURING TRIPARANOL

these sterol metabolites, it is apparent that the en-

zymes involved in the splitting of the cholesterolside chain in the biosynthesis of bile acids andsteroids do not require a great deal of structuralspecificity in the sterol side chain. In addition,these results raise the interesting possibility thateven under normal conditions some bile acid andsteroid biosynthesis takes place from sterol pre-

cursors other than cholesterol. In other words itis possible, during the bioysnthetic sequence fromlanosterol to cholesterol, that some of the inter-mediate sterols could be used directly for the syn-

thesis of bile acids or steroids. Whether this nor-

mally occurs to any extent is, of course, notknown.

Several recent reports indicate that adrenalfunction seems to be reduced during triparanoladministration. Melby, St. Cyr and Dale foundthat the production of cortisol and aldosterone was

significantly reduced in normal subjects at a dos-age of 1,000 mg per day (27). Hollander, Cho-banian and Wilkins noted a slight but statisticallysignificant decrease in urinary 17-hydroxycorti-costeroid excretion during treatment of patientswith 250 mg triparanol daily; 17-ketosteroid ex-

cretion was, however, normal (28). Reducedadrenal vein corticosterone has also been ob-served during the administration of relativelymuch larger doses of triparanol to rats (29). Inview of the present finding that desmosterol is a

suitable precursor for steroid hormones of a quali-tatively normal composition, it is quite possiblethat this decline in adrenal activity during tri-paranol therapy is not due to the blocking of theconversion of desmosterol to cholesterol per se,

but to some other factor. This factor might wellbe relatively nonspecific, possibly related to a va-

riety of toxic reactions noted in this laboratory(unpublished) during administration of relativelylarge doses of triparanol to animals. These effectsinclude reduced weight gain and listlessness andinduction of abortion in rats.

Somewhat opposing therapeutic implicationscan be drawn from the results of these studies.On the one hand, the fact that desmosterol can

serve as a satisfactory precursor for bile acidsand steroid hormones suggests that the partial re-

placement of cholesterol by desmosterol during tri-paranol administration is not likely in itself to

disrupt seriously the over-all physiology of the

body. On the other hand, these results add newinformation on the great similarity that exists be-tween desmosterol and cholesterol. Avigan andSteinberg and their associates have already com-mented upon the similarities of these two sterolsin their physical and chemical properties, and havenoted the widespread distribution of desmosterolin many different tissues, and the comparable ratesof esterification of the sterols in vivo (2, 5). Itis now apparent that the two sterols also enterother biosynthetic reaction sequences in a similarfashion. It is hence likely that desmosterol willresemble cholesterol in its relationship to the de-velopment of atheromatous lesions or clinicalatherosclerosis. Recent observations in this lab-oratory (Avigan, unpublished) have in fact indi-cated that desmosterol can be recovered from bothnormal and atheromatous aortae of animals beingtreated with triparanol. It is therefore probablethat, as previously suggested (5), the best short-term indicator of the potential value of triparanoltherapy resides in its effect on the total sterollevel, rather than on the cholesterol level per se.An average depression of total circulating sterolsof 15 per cent has been reported (5). In addi-tion, the long-term biological effects of this therapystill require further consideration.

SUMMARY

The conversion of 24-dehydrocholesterol (des-mosterol) to bile acids and to steroid hormonesduring triparanol administration has been studiedin two male subjects. In each study, DL-2-C14-mevalonic acid was injected intravenously and theC14 specific radioactivity of serum desmosteroland cholesterol was determined after 3 hours andat 1- to 3-day intervals thereafter. Radioactivityrapidly appeared in desmosterol with peak specificactivity obtained within a few hours; cholesterolC14 specific activity was very much lower duringthe first few days of each study. Bile sampleswere collected by duodenal intubation and the bileacids extracted and purified by column chroma-tography. The urinary metabolites of cortisolwere extracted from 24-hour urine samples andpurified by chromatography in one or two systems.In one study, derivatives of the urinary C,,O0steroids were separately isolated and purified. Inboth patients the C'4 specific radioactivity of the

969

DEWITT S. GOODMAN,JOEL AVIGAN AND HILDEGARDWILSON

bile acids and steroids was much higher than thatof cholesterol throughout the study, demonstrat-ing that a significant part of these compoundsmust have been. made directly from desmosterol.

In one study 7-a-H3-cholesterol, previously in-corporated into the patient's serum lipoproteins,was injected together with the C14-mevalonate, re-sulting in the labeling of cholesterol and desmos-terol with different isotopes. Comparison of theserum specific activity curves for the two sterols(C14-desmosterol and H3-cholesterol) showedthat their disappearance rates were fairly similar,desmosterol disappearance being a little more rapidthan that of cholesterol. Calculation of the over-all metabolism of desmosterol from the double-isotope data indicated that in this patient only21 ± 4 per cent of the desmosterol that disap-peared from the metabolically active pool reap-peared therein as cholesterol. Approximatelythree-fourths of desmosterol disappearance, there-fore, occurred via noncholesterol pathways.

The specific activity data and other considera-tions suggest that desmosterol is similar to cho-lesterol in its ability to serve as a direct precursorfor several sterol metabolites-namely, bile acidsand steroid hormones.

ACNOWLEDGMENT

The authors are greatly indebted to Dr. Daniel Stein-berg for advice in the design and conduct of these stud-ies, and to Mr. Hugh Vroman and Mr. David W. Ryanfor expert technical assistance.

REFERENCES

1. Goodman, D. S., Avigan, J., and Wilson, H. N. Themetabolism of desmosterol in human subjects dur-ing triparanol administration. Biochem. Pharma-col. 1961, 8, 87.

2. Avigan, J., Steinberg, D., Vroman, H. E., Thompson,M. J., and Mosettig, E. Studies of cholesterol bio-synthesis. I. The identification of desmosterol inserum and tissues of animals and man treated withMER-29. J. biol. Chem. 1960, 235, 3123.

3. Steinberg, D., and Avigan, J. Studies of cholesterolbiosynthesis. II. The role of desmosterol in thebiosynthesis of cholesterol. J. biol. Chem. 1960,235, 3127.

4. Frantz, I. D., Mobberley, M. L., and Schroepfer,G. J., Jr. Effects of MER-29 on the intermediarymetabolism of cholesterol. Progr. cardiovasc.Dis. 1960, 2, 511.

5. Steinberg, D., Avigan, J., and Feigelson, E. B. Ef-fects of triparanol (MER-29) on cholesterol bio-

synthesis and on blood sterol levels in man. J.clin. Invest. 1961, 40, 884.

6. Avigan, J. A method for incorporating cholesteroland other lipides into serum lipoproteins in vitro.J. biol. Chem. 1959, 234, 787.

7. Norman, A. Separation of conjugated bile acids bypartition chromatography. Bile acids and steroids6. Acta chem. scand. 1953, 7, 1413.

8. Sjovall, J. On the separation of bile acids by par-tition chromatography; bile acids and steroids.Acta physiol. scand. 1953, 29, 232.

9. Mosbach, E. H., Zomzely, C., and Kendall, F. E.Separation of bile acids by column-partition chro-matography. Arch. Biochem. 1954, 48, 95.

10. Mosbach, E. H., Kalinsky, H. J., Halpern, E., andKendall, F. E. Determination of deoxycholic andcholic acids in bile. Arch. Biochem. 1954, 51, 402.

11. Wilson, H., Borris, J. J., and Garrison, M. M.Chromatographic procedure for the determinationof urinary corticosteroids and C,9 steroids. J. clin.Endocr. 1958, 18, 643.

12. Wilson, H., Lipsett, M. B., and Butler, L. C. Steroidexcretion in hypophysectomized women, and theinitial effects of corticotropin: A study in urinarysteroid patterns. J. clin. Endocr. 1960, 20, 534.

13. Wotiz, H. H., Lemon, H. M., and Marcus, P. De-termination of urinary 11-oxygenated 17-ketogenicsteroids. J. clin. Endocr. 1957, 17, 116.

14. Bush, I. E. Methods of paper chromatography ofsteroids applicable to the study of steroids in mam-malian blood and tissues. Biochem. J. 1952, 50,370.

15. Wilson, H. Chromogenic values of various keto-steroids in a micro modification of the Zimmer-mann reaction: Comparison with the macro pro-cedure. Arch. Biochem. 1954, 52, 217.

16. Cornforth, J. W., Cornforth, R. H., Popjak, G., andGore, I. Y. Studies on the biosynthesis of cho-lesterol. 5. Biosynthesis of squalene from DL-3-hydroxy-3-methyl- (2-14C) pentano-5-lactone. Bio-chem. J. 1958, 69, 146.

17. Isler, O., Ruegg, R., Wuirsch, J., Gey, K. F., andPletscher, A. Zur Biosynthese des Cholesterins ausp,8- Dihydroxy-,8-Methyl -Valeriansaure. Helv.chim. Acta 1957, 40, 2369.

18. Constantopoulos, G., and Tchen, T. T. Cleavage ofcholesterol sidechain by adrenal cortex. I. Cofac-tor requirement and product of cleavage. J. biol.Chem. 1961, 236, 65.

19. Bergstrom, S., Lindstedt, S., Samuelson, B., Corey,E. J., and Gregoriou, G. A. The stereochemistryof 7-a-hydroxylation in the biosynthesis of cholicacid from cholesterol. J. Amer. chem. Soc. 1958,80, 2337.

20. Lindstedt, S., and Sjovall, J. The formation of de-oxycholic acid from cholic acid in the rabbit. Actachem. scand. 1957, 11, 421.

21. Popjak, G., and Cornforth, J. W. The biosynthesisof cholesterol. Advanc. Enzymol. 1960, 22, 281.

970

DESMOSTEROLMETABOLISMDURINGTRIPARANOL

22. Christophe, J., and Popjak, G. Studies on the bio-synthesis of cholesterol. XIV. The origin of pre-noic acids from allyl pyrophosphates in liver en-zyme systems. J. Lip. Res. 1961, 2, 244.

23. Rosenfeld, R. S., and Hellman, L. The relation ofplasma and biliary cholesterol to bile acid syn-thesis in man. J. clin. Invest. 1959, 38, 1334.

24. Werbin, H., and Chaikoff, I. L. Utilization ofadrenal gland cholesterol for synthesis of cortisolby the intact normal and the ACTH-treated guineapig. Arch. Biochem. 1961, 93, 476.

25. Morris, M. D., and Chaikoff, I. L. The origin ofcholesterol in liver, small intestine, adrenal gland,and testis in the rat: Dietary versus endogenouscontributions. J. biol. Chem. 1959, 234, 1095.

26. Blohm, T. R., Kariya, T., and Mackenzie, R. D. Ex-cretory products of the sterol biosynthetic path-way in the MER-29 treated rat. Excretion ofMER-29. Progr. cardiovasc. Dis. 1960, 2, 519.

27. Melby, J. C., St. Cyr, M., and Dale, S. L. Reductionof adrenal-steroid production by an inhibitor ofcholesterol biosynthesis. New Engl. J. Med. 1961,264, 583.

28. Hollander, W., Chobanian, A. V., and Wilkins,R. W. The effects of triparanol (MER-29) insubjects with and without coronary artery dis-ease. J. Amer. med. Ass. 1960, 174, 5.

29. Holloszy, J., and Eisenstein, A. Effect of MER/29(triparanol) on corticosterone secretion by rat

adrenals. Fed. Proc. 1961, 20, 176.

SPECIAL NOTICE TO SUBSCRIBERS

Post Offices will no longer forward the Journal when you move.

Please notify The Journal of Clinical Investigation, BusinessOffice, 10 Stoughton Street, Boston 18, Mass., at once when youhave a change of address, and do not omit the zone number ifthere is one.

971