Abnormal Induction of 3-Hydroxy-3-methylglutaryl Coenzyme A Reductase in Leukocytes from Subjects with Heterozygous Familial Hypercholesterolemia*

Human leukocytes isolated from fresh defibrinated blood were shown to utilize acetate and mevalonate for sterol synthesis. The capacity of the leukocytes to synthesize sterols is limited severely as compared to their ability to convert mevalonate into farnesyl pyrophosphate (which they hydrolyze rapidly to free famesol) and into squalene. When leukocytes are incubated in a medium containing lipid-free serum, synthesis of sterols from acetate, but not from mevalonate, is much enhanced. It was shown that this increased synthesis resulted from increased levels of 3-hy-droxy-3-methylglutaryl-CoA reductase activity in the cells. A comparison was made of the activation of sterol synthesis from acetate in leukocytes of normal individuals and of heterozygous familial hypercholesterolemics. The latter group to incubation in lipid-free with a sig-nificantly higher the


SUMMARY
Human leukocytes isolated from fresh defibrinated blood were shown to utilize acetate and mevalonate for sterol synthesis.
The capacity of the leukocytes to synthesize sterols is limited severely as compared to their ability to convert mevalonate into farnesyl pyrophosphate (which they hydrolyze rapidly to free famesol) and into squalene.
When leukocytes are incubated in a medium containing lipid-free serum, synthesis of sterols from acetate, but not from mevalonate, is much enhanced. It was shown that this increased synthesis resulted from increased levels of 3-hydroxy-3-methylglutaryl-CoA reductase activity in the cells. A comparison was made of the activation of sterol synthesis from acetate in leukocytes of normal individuals and of heterozygous familial hypercholesterolemics. The latter group responded to incubation in lipid-free sera with a significantly higher activation than the cells of normocholesterolemics.
This activation was shown to be well correlated with a higher induction of 3-hydroxy3-methylglutaryl-CoA reductase in the heterozygous cells than in the normals. The leukocytes of a heterozygous familial hypercholesterolemic individual were found to release, into a lipid-free incubation medium, more endogenously synthesized [3H]sterol (but not [3H]squalene) than the cells of a normal person. It is suggested that the genetic abnormality in heterozygous familial hypercholesterolemia could be accounted for by a mutation resulting in a weaker binding of a sterol repressor by heterozygous cells than by normal cells.
We have reported briefly on the use of leukocytes for studying sterol biosynthesis in man and on an abnormality of the control of sterol biosynthesis in cells of heterozygous familial hyper-* This work was supported in part by United States Public Health Service Research Grants HI,-12745 and RR 865, and by the Edna and George Castera and the Occidental Life of California funds at UCLA. $ Veterans Administration Research andEducation Associate, 1972 to 1974. cholesterolemics (1). We have observed that sterol biosynthesis from acetate in the leukocytes was stimulated when the cells were incubated for 6 hours in a lipid-free serum as compared to cells kept in a medium containing full serum. Utilization of mevalonate was not affected by the nature of the incubation medium.
The stimulation of the conversion of [14C]acetate into sterols by cells kept in the lipid-free serum was inhibited completely by 10 PM cycloheximide which had no effect either on the utilization of [14C]acetate by cells kept in full serum or on the utilization of [2-YJmevalonate by cells maintained in full or lipid-free serum. The data were interpreted to mean that in cells incubated in lipid-free sera an enzyme preceding those acting on mevalonate was induced and that the induction of this enzyme, presumed to be 3-hydroxy-3-methylglutaryl-CoA reductase, was repressed in media containing full serum. We have noted further that while the activation of acetate utilization by the lipid-free serum in the cells of normocholesterolemics was 2-fold in B-hour incubations, in the cells of heterozygous familial hypercholesterolemics it was 3-fold. This was the first enzymic abnormality noted in heterozygous familial hypercholesterolemics.
The purpose of this paper is to document fully the sterol-synthesizing properties of human leukocytes and to present evidence that the activation phenomenon was indeed associated with the induction of HMG-CoAl reductase and that the induction of this enzyme in leukocytes of heterozygous familial hypercholesterolemics was more rapid than in cells of normal individuals.
Evidence will be presented also that the induction of HMG-CoA reductase may be a compensatory mechanism for the loss of sterols from the cells into the medium.
We will discuss the correlation of our observations with those of ) made on cultured skin fihroblasts of normocholesterolemic, of homozygous and heterozygous familial hypercholesterolemic individuals.
The heterozygotes were a random sample of new referrals to the cardiac and lipid clinics of the UCLA Medical Center. The controls were new referrals to the cardiac clinic who proved to be normal in respect of their serum lipoproteins; others were members of the Medical Center staff. No one in either group received drugs that might have affected serum cholesterol levels or lipid metabolism. All had normal hematocrits, reticulocyte counts, white blood cell, and differential counts. Informed consent was obtained in writing from each person.
Collection, Separation, and Incubation of Leukocytes-Leukocytes were prepared by a modification of the method of Coulson and Chalmers (9). After a 12-to 14-hour fast, blood was drawn aseptically from a cubital vein through a 19-gauge scalp vein needle into 50 ml sterile plastic syringes. Their contents were emptied immediately into a 250-ml flask containing glass beads -and the flask was shaken at 140 strokes per min on an Eberbach datform shaker (Eberbach Corn.. Ann Arbor. Mich.1 at an anrde bf 10" for 15 min in order to def?brinate the blood. The defibr:nated blood was then decanted from the clot into a wide mouthed test tube and 30 ml was drawn through an l&gauge needle into a 50-ml plastic syringe containing 15 ml of Plasmagel (product of Laboratoire Rdger Bellon, Neuiily, France; suppliers I%TI Corp., Buffalo. N.Y.1. The contents were mixed and the svrinees were rested on the& plungers for 45 min in order to sediient-the red cells. After the sedimentation was completed, a 19-gauge scalp vein needle was attached to the syringe and the barrel depressed over the plunger to express the supernatant.
The first few drops and any subsequent particulate matter, e.g. small fibrin clots, seen through the plastic tubing of the needle were discarded.
The supernatant from all syringes was collected into a single vessel which was constantly swirled in order to ensure a homogenous dispersion of cells. Usually 20 ml of this cell suspension, containing about lo* leukocytes, were dispensed with a sterile plastic pipette into 40-ml wide mouthed centrifuge tubes which were spun at 250 X g for 12 min; the supernatant was discarded.
The cells were resuspended in 10 ml of Krebs-Ringer phosphate buffer, pH 7.4, containing 0.03 M glucose and were sedimented again by centrifuging at 250 X g for 12 min; the supernatant was decanted. The cells (about 0.5 ml) were then suspended in 1.0 ml of the Krebs-Ringer buffer and were transferred to 25-ml flasks packed in melting ice. The quantitative *ransfer of the leukocytes to the incubation flasks was made by the rinsing of the centrifuge tubes with 3.3 ml of either full or lipid-free serum (see further on). The mevalonate were also used; the amounts of these substrates and volume of the incubation wai made up to 5 ml by the addition of 0. Before the start of the incubation, the flasks, removed from the ice bath, were flushed for 4 min with a stream of 95yo 02-5% CO, vigorous enough to saturate the cell suspension without causing splattering.
The flasks were tightly stoppered and sealed with Parafilm and were then incubated at 37" in a New Brunswick gyratory water bath shaker at 150 rpm.
The concentration of the serum (full or lipid-free) was 44% in supernatant was filtered through Whatman No. 42 paper and was frozen at -20". At least 4 units of plasma thus treated were subsequently pooled and titrated with human thrombin (Fibrindix, Ortho Diagnostics, Raritan, N.J.) and CaClz at 37" in a shaking water bath in order to convert fibrinogen to fibrin.
The serum thus produced was heated at 56" for 30 min in order to inactivate complement.
The serum was then centrifuged at 10,000 X B for 1 hour and was filtered through a Whatman-No. 42 paper.2 -The filtrate was diluted with one-third of its volume of glass-distilled water and was then passed through a Millipore filter ypore size 0.22 pm; Millipore Corp., Bedford, Mass.) and was frozen in sterile containers at -20'. Preparation of Lipid-free Serum-Lipid-free serum was made according to the method of McFarlane (10) from the pooled ABnegative serum, before dilution with water, by the use df peroxidefree diethyl ether. The extracted serum. after removal of the ether dissolved in it with a Nz-stream, was diluted with distilled water and was sterilized and stored as described above for the full serum.
McFarlane's method was shown to remove all free and esterified cholesterol and triglycerides from the serum and about onehalf of all of the phospholipids (11,12). We confirm this observation as none of the extracted sera contained detectable amounts of cholesterol or triglycerides.
It is pertinent also that this method of lipid extraction causes no loss or denaturation of serum proteins as judged by determination of albumin to globulin ratios, moving boundary electrophoretic analysis and osmotic pressure measurements of the extracted sera (11, 12).

Determination
of Integrity and Viability of Leukocytes-The leukocytes were observed microscopically before and after incubation for their ability to exclude supravital stains (eosin Y and trypan blue) and their morphology was assessed in fixed and stained preparations.
The average composition of the leukocyte population was 65yo granulocytes, 30% lymphocytes, and 5% monocytes.
No platelets were present. The abilitv of the granulocytes to phagocitize latex particles was assessed bkfore ani after incubation in full and lipid-free sera for varying times up to 24 hours.
The-composition of digiton&-precipitable sterols was examined after decomposition of the digitonides with pyridine and precipi- desmosterol, were first separated into the CZT and C30 compounds The unsaponifiable petroleumsoluble radioactive products were then analyzed for farnesol, by gas-liquid radiochromatography (14), for squalene, and for digiby thin layer chromatography on plates of Silica Gel G with 2,2,4-tonin-precipitable sterols (13).
trimethylpentane-ethyl acetate-acetic acid (60:30:0.6, by volume) and were then acetylated by refluxing with acetic anhydride in dry benzene (15). The acetates of the Cz, and Caa sterols were the incubations.
The cholesterol and triglyceride content of the incubations made with full serum was 0.81 and 0.59 mg per ml, respectively.
Neither cholesterol nor triglycerides could be detected in the lipid-free sera used. The final concentration of Ca2+ in all incubations was 12 mg/lOO ml.
Cell counts, by standard techniques, were made in duplicate before and after the incubations.
The mean leukocyte content in the incubations was (1.15 =t 0.05) X 108 per flask unless otherwise specified. All results were calculated to 10s cells.
Excepting the defibrination flask (glass), and the glass beads it contained, all equipment was plastic or glasss iliconized with Siliclad (Clay Adams, Parsippany, N.J.); also, all transfers of leukocytes from one vessel to another were made with siliconized Pasteur pipettes in order to avoid damage to the leukocytes and adhesion of cells to uncovered glass surfaces.
All of the equipment used in handling the leukocytes and all solutions used were bacteriologically sterile, and aseptic conditions were observed in all manipulations.
Preparation of Human AB-ne(lative Serum-Units (about 1 pint each) of AB-negative human plasma, obtained from the UCLA blood bank, were centrifuged at 10,000 X g for 30 min, then the further resolvkd by chromatography on silver nitrate-impregnated Silica Gel H nlates (12.5% AnNO. with benzene-hexane (3:5. by volume) (16). Thk plates v&e scanned for radioactivity ai& a Packard radiochromatogram scanner, model 7201. The substances on the plates were displayed with a spray of rhodamine G and were eluted from the scraped off gel with diethyl ether.
The yields of products were calculated from the specific activities of the substrates and on the assumption that squalene and the C30 sterols contained 18 and 6 labeled aositions and the C21 sterols 15 and 5 labeled positions derived from [2X?]acetate and [2-"Clmevalonate, respectively, and that the %-substrates were not diluted in the cells with unlabeled material.
Farnesol was taken to have acquired three labeled positions from [2-14C] and colorless preparation of the anhydride. 4 An efficient method of elution from paper strips is the packing of the strips in a roll into conical centrifuge tubes so that the lower edge of the roll remains 2 to 3 cm above the tip of the centrifuge tube.
The roll is moistened with the eluting solvent and the tube is then centrifuged.
The solvent, spun off the paper, is collected with a Pasteur pipette, and the elution repeated.
stopped by the addition of 0.02 ml of 6 N HCl.

Properties of Isolated Human
Leukocytes-Human leukocytes incubated in either the full or lipid-free sera up to 24 hours retained their morphological integrity as judged by their microscopic appearance in fixed and stained preparations and could not be distinguished from freshly prepared cells. They all excluded supravital stains, even after 24.hour incubations. There was, however, a substantial deterioration in the phagocytic ability of the cells with prolonged incubations, whether in full or lipid-free sera: up to 15.hours of incubation all granulocytes phagocytized latex particles, after 18 hours 80% of the cells still phagocytized latex particles, but after 24 hours of incubation, only 30% of the cells retained this property.
Hence it would seem that one is limited to approximately 18 hours of work, about 20 hours after their isolation, with human leukocytes. All our experiments were carried out within this period. blood, kept liquid with 900 units of sodium heparin, under 95% 02-5s CO2 with concentrations of glucose and [2J4C]acetate identical with those used in the standard 5-ml incubations of isolated leukocytes in full serum. The isolated leukocytes converted 300 pmol and the leukocytes in the whole blood converted 244 pmol of acetate per 10s cells into digitonin-precipitable sterols (to be referred to subsequently as sterols for short).
Thus the isolated leukocytes retained fully their synthetic power, which may even have been stimulated slightly by the isolation or incubation in the serum.
The incorporation of [14C]acetate into sterols was linearly proportional to the number of leukocytes incubated (Fig. 1). Thus leukocyte numbers provide a standard measure for intra-and interexperimental comparisons. As will be shown later (cf. Fig. 2), the incorporation of [14C]acetate into sterols was linearly proportional in respect of time also, up to 18 hours, but only in incubations made in full serum.
The system is highly reproducible, as the results of duplicate incubations differed by less than 5%. Repeated studies of the cells of the same individual gave identical results.
Nature of Sterols and Sterol Intermediates Synthesized by Leukocytes-Depending on the starting substrate, various intermediates of cholesterol biosynthesis or substances derived from these can be recognized in the incubations of leukocytes.
The data of Table I show that leukocytes incubated for 6 hours in a lipid-free serum synthesized from acetate substantially more sterol than in full serum, but the nature of the incubation medium had no effect on sterol synthesis from mevalonate. Since both acetate and mevalonate were always in excess in these experiments, this is the first demonstration that, as in other species, the rate-limiting step (or steps) of cholesterol synthesis in man precedes the formation of mevalonate.
Moreover, this is also the first demonstration of limiting reactions even beyond mevalonate in fresh, intact human cells. The data of Table I show that in contrast to the great powers of the cells to synthesize farnesyl pyrophosphate (the presumed source of the free farnesolj 1. Proportionality between leukocyte numbers and conversion of acetate into sterols. The leukocytes of a 25-year-old male were isolated and incubated for 6 hours with 44.05 &i of [2J4C]acetate (56 Ci/mol) in the medium containing full serum in the standard 5-ml mixture. and squalene, the ability of the cells to convert squalene into sterols is limited.
In the standard incubation the concentration of RX-[2-14C]mevalonate was 0.92/5 = 0.184 pmol/ml or 92 PM in respect of the utilizable R-enantiomer.
This was clearly in excess of the amount of mevalonate lo* cells could utilize, as only about 6 nmol of products accumulated of which only 71 to 80 pmol were accounted for in sterols (cf. shown refer to that of the utilizable R-enantiomer.   The differential activities of the enzymes catalyzing reactions beyond mevalonate were apparent also in preparations of disrupted leukocytes and were very similar to those seen in whole cells (Table III).
in the two media are that cholesterol formed a higher proportion of sterols synthesized at the lower synthetic rate observed when [14C]acetate was the substrate as compared to the rates seen with [2-r4C]mevalonate (cf. Table 1) and that in the lipid-free medium more desmosterol was synthesized irrespective of the starting substrate.
Analysis of the digitoninprecipitable sterols synthesized by the leukocytes revealed intermediates between lanosterol and cholesterol (Table IV).
The notable features of the distribution of the various sterols synthesized from the two substrates and  (1) the cells were incubated for 6 hours. We became particularly interested to find out the time course of the apparent activation of sterol synthesis from acetate in leukocytes incubated in the lipid-free sera. The data of Table V show that the synthesis of sterols in cells incubated in lipid-free sera was not linearly proportional to time, but that a rapid rise ensued beyond the 6th hour of incubation. This is in contrast to the data obtained from incubation in full serum (cf. Fig. 2). The data of Table V also show that squalene did not accumulate in the cells proportionately to the increased synthesis of sterols. As will be shown later, HMG-CoA reductase levels rise markedly in cells incubated for more than 6 hours in lipid-free sera.
Subsequent to the observations reported by US previously (l), we have taken leukocytes from further five heterozygous In these experiments, as in the one shown in Table V, the concentration of acetate was increased 3-fold over that used in the earlier experiments in order to exclude the possibility of acetate becoming limiting in the long incubations.
Also, in these long incubations, penicillin, 100 units/ml, and streptomycin, 100 pg/ml, were added to the media.
The data of the individual experiments are set out in Table  VI. All were made with the same batch of full serum from which the lipid-free serum was subsequently prepared.
The blood for the isolation of the leukocytes was taken from the familial hypercholesterolemics and from their controls in the morning after an overnight fast. There was no significant difference in full serum between the cells of normals and heterozygotes, even though the values for the heterozygotes were slightly higher than for the controls (Table VI and Fig. 2). In full serum, the incorporation of ['*C]acetate into sterols was linear with time up to 18 hours. At 6 hours the cells of these five hypercholesterolemics behaved exactly as was found for the previous six heterozygotes (l), i.e. they incorporated 3 times more acetate into sterols in the m x10 -3/10rJ cells 67.6 1 37. lipid-free serum than in the full serum, whereas the control cells showed only a 2-fold rise. The amounts (picomoles) of [2-14C]acetate incorporated into sterols were also in the same range in this set of experiments at 6 hours as in the earlier ones (1) even though the concentration of acetate was 3 times higher, showing that acetate was not limiting, nor could the differences noted between normal and heterozygous cells be ascribed to differences in endogenous pools of acetyl-CoA.
Moreover, we found that the results were identical when the same batch of cells were incubated in lipid-free sera for 18 hours with either 0.44 or 0.88 mM acetate.
Since acetate was not limiting in any of the experiments, and since we have shown that the effect of lipid-free serum was to induce an enzyme, or enzymes, preceding the formation of mevalonate (I), we can consider that the acetate incorporated into sterols in lipid-free serum minus that incorporated into sterols in full serum was the amount attributable to the activity of the "induced" enzyme. The calculated values for the cells of normals and heterozygous hypercholesterolemics are set out in Table VII and show a nearly twice as large activation of acetate utilization in the hypercholesterolemic cells as in the normal cells both in the early and in the later period of incubation. Our terolemic man (serum cholesterol 360 mg/lOO ml; serum triglycerides 60 mg/lOO ml; cornea1 arcus and tendon xanthomata present; no coronary disease) was divided into two lots. One lot, divided into two flasks, was preincubated in full and lipid-free serum, respectively, for 8 hours with 0.29 mM unlabeled acetate.
After 8 hours, [2-i4C] that the enhanced incorporation of ['"Clacetate in leukocytes incubated in lipid-free sera resulted probably from an induction of HMG-CoA reductase (1). To our knowledge, this enzyme has not been assayed previously in extracts of leukocytes.
We have tested the method of Hulcher and Oleson (18) for the assay of this enzyme in leukocyte extracts, but found it unsatisfactory on account of the very high nonspecific deacylase activities of the extracts.
The methods of Goldfarb and Pitot (17) and of Shapiro et al. (22), depending on the extraction from the reaction mixture of [14C]mevalonolactone, formed from [r4C]HMG-CoA, with [3H]mevalonolactone added as internal standard followed by purification of [I%, 3H]mevalonolactone by thin layer chroma,tography, and determination of the r4C:3H ratio in the mevalonolactone could, however, be readily adapted to the assay of HMG-CoA reductase in extracts of leukocytes.
We have experimented with sonically disrupted cells and with cells lysed with detergent.
We have abandoned the use of sonically disrupted preparations of leukocytes for assay for HMG-CoA reductase because such preparation converted, in the presence of even traces of ATP and NADH and NADPH, the minute amounts of mevalonate formed from HMG-CoA into farnesyl pyrophosphate, squalene, and sterols, thus making the assays difficult.
In cell extracts made by the lysis of cells with the KYRO EOR detergent (19) we have never seen the conversion of mevalonate into other products under our assay conditions.
We have found that extracts of leukocytes, prepared as described under "Materials and Methods," converted HMG-COA to mevalonate at a rate that was linear with respect to protein concentration up to 2.4 mg/ml and up to 60 min of incubation. Duplicate analyses differed by less than 5%.
It remained to be established whether the higher utilization of [14C]acetate by cells incubated in lipid-free sera than in full sera was indeed associated with the induction of HMG-CoA reductase or not.
The experiment shown in Table VIII shows that there was a strict proportionality between HMG-CoA reductase levels in the leukocytes and the amount of [2-l*C]acetate incorporated into sterols.
In Table VIII  by the HMG-CoA reductase from HMG-CoA and the amounts of mevalonate that must have been formed in the whole cells from [r4C]acetate to account for the incorporation of r4C into sterols. The data show unequivocally a parallel between induction of HMG-CoA reductase in cells incubated in lipid-free serum (12.6.fold rise) and the induction of increased utilization of [r4C]acetate for sterol synthesis (11.5-fold rise). The calculations of mevalonate synthesis from [14C]acetate by the whole cells assumed no dilution of the labeled precursor with unlabeled acetate generated endogenously from the metabolism of, e.g. glucose present in the media. The lower value (50%) found for mevalonate synthesis from acetate as compared to synthesis from HMG-CoA is understandable since, even after penetration of acetate into the cells, three reactions are needed for the synthesis of HMG-CoA from acetate. It must he borne in mind also that the endogenous synthesis of mevalonate was calculated only from the incorporation of acetate into sterols, the squalene accumulating in the cells not being taken into account. The correlation between HMG-CoA reductase activity and acetate incorporation into sterols is so good in the leukocytes that the synthesis of [14C]sterols from [i4C]acetate may be taken as a measure of HMG-CoA reductase activity. It still remained to be shown that the time course of the induction of increased acetate incorporation into sterols of leukocytes of normal individuals and hypercholesterolemics (cf. Fig. 2) was matched by a sirnilar induction of HMG-CoA reductase. To this end the time course of the induction of HIMG-CoA reductase was determined by measuring enzyme activity in cells taken freshly from the blood (zero time) and then after 3, 6, 9, and 12 hours of incubation in lipid-free sera. The experiments were done with the cells of one normal individual and of one heterozy- The heterozygote was a 40-year-old male with cornea1 arcus, xanthelasma, tendon xanthomata, and coronary artery occlusion (serum cholesterol 378 mg/ 100 ml and triglycerides 94 mg/lOO ml). The normal male control was 40 years old; his serum cholesterol and triglyceride concentrations were 182 mg/lOO ml and 134 mg/lOO ml, respectively. The leukocytes were incubated, their extracts prepared, and assayed for HMG-CoA reductase as described in the text. [3H]Sterols synthesized by leukocytes of a normal and a heterozygous familial hypercholesterolemic incubated first for 3 hours in full serum with [5-3Hlmevalonate and then transferred to lipid-free serum for the times shown. The values refer to the sum of the sterols found in the cell pellets and the medium. The zero time values are the sum of those found in the cells and the medium after the preliminary a-hour incubation in full serum. Experimental details are given in the text. The heterozygote was a 2% year-old male with tendon xanthomata, but no evidence of coronary artery occlusion (serum cholesterol 432 mg/lOO ml; serum triglycerides 80 mg/lOO ml). The normal control was a 2%yearold male with a serum cholesterol level of 233 mg/lOO ml and triglyceride level of 62 mg/lOO ml. taken and 4-ml volumes of the mixture were dispensed to eight 40.ml tubes which were centrifuged at 250 x 0 for 12 min; the gous familial hypercholesterolemic.
The data obtained are full serum was decanted, the cell pellets were washed and transshown in Fig. 3. ferred to 25.ml flasks to which lipid-free sera were added in a The similar reductase activities in the leukocytes of the hetero-final concentration of 44% as already described and were incuzygote and normal taken freshly from the blood (zero time, be-bated for 2, 3, 4 and 6 hours. Duplicate flasks were taken at fore incubation) confirms the data obtained with [Wlacetate in each time point and their contents centrifuged at 10,000 X g full serum.
for 30 min. The [3H]sterol content of the medium' and the cell The time course for the induction of the reductase in normal pellet was determined, and the former expressed as a per cent of and heteroxygous cells in lipid-free sera corresponds to the time the sum of the two. The cell pellets and the media were also course for the increased incorporation of acetate into sterols in analyzed for [3H]squalene at each step of the incubations. The normal and heterozygous cells in lipid-free sera (cf. Fig. 2). This cells were intact and excluded supravital stains even after 6 is further proof of the excellent correlation between the incorpora-hours in the second incubation medium. Not unexpectedly, the tion of acetate into sterols and HMG-CoA reductase activity, cell pellets contained much more [3H]squalene than [3H]sterol at and thus supports the validity of the conclusions drawn from the the end of the preliminary a-hour incubation: 144,000 dpm/lO* acetate data. normal cells and 112,000 dpm/l04 heterozygous cells in squalene, Release of Sterols from Leukocytes to Medium-The induction of as compared to 4200 dpm/l08 normal cells and 4000 dpm/lO* HMG-CoA reductase in leukocytes incubated in lipid-free sera heterozygous cells in sterols (cf. Fig. 4). In spite of the heavy raised the questions, what was the underlying cause of the induc-labeling of cellular squalene, there were only 2080 dpm/lO* cells tion, and why was a higher activity induced in the heterozygous (normal or heterozygous) associated with squalene in the first cells than in the normals?
The lipid-free sera used contained medium.
After transfer of the cells to lipid-free sera (without the sterol-and triglyceride-free apoproteins of the lipoproteins exogenous substrate), both normal and heterozygous cells con-(10-12).
This fact raised the possibility that the induction of tinued to accumulate [3H]sterols, presumably by the conversion HMG-CoA reductase was a compensatory response to a normal of endogenous labeled substrates (e.g. of [3H]squalene) for 3 loss of sterols from cells to the environment and that the cells of hours (Fig. 4). The heterozygous cells accumulated, however, heterozygotes lost cellular sterols faster than the cells of un-less [3H]sterol than the normal cells. After the first 3 hours in affected individuals. full serum, the normal cells had lost 13% of the total [3H]sterols For the study of these questions we have first incubated a large into the medium, while the heterozygous cells had lost 21%. batch of leukocytes from a 2%year-old normal individual (14 X Then upon transfer to the lipid-free serum, both normal and lOa leukocytes) and from a 28.year-old heterozygous familial heterozygous cells rapidly lost sterol into the medium, but by 2 hypercholesterolemic (11 X lo* leukocytes) in 45 ml of medium hours the normal cells had nearly reached a new steady state, of the standard composition with full serum and with 250 &i of while the heterozygous cells continued to lose an ever increasing RS-[5-3H]mevalonate (6.74 Ci per mmol) for 3 hours in order to 7 Unlabeled cholesterol (1 mg/ml) was added for the isolation of label the sterols. At the end of the 3 hours, cell counts were [3H]sterols, from the lipid-free media, as the digitonides.
I I hour and in the heterozygous cells to 32 to 38 pmol per hour per 108 cells upon prolonged incubation in a lipid-free medium. The lower maximum rate of synthesis calculated from mevalonate utilization may be attributed to possible inhibitory effects of intermediates accumulating in the cells (cf . Tables I and II) when mevalonate is freely available.
The capacity of the leukocytes to synthesize farnesyl pyrophosphate and squalene is very large as compared to their ability to generate mevalonate or to convert squalene into sterols. Our results in respect of the synthesis of squalene relative to the conversion of squalene to sterols are very similar to those of Edgren and Hellstrom (25) obtained by in uiuo experiments on rats.
The evidence we have presented here proves firmly that the increased sterol synthesis from acetate seen in cells incubated in a lipid-free medium was indeed associated with the induction of HMG-CoA reductase as we have inferred previously (1)  5. Release of [3H]sterols by leukocytes of a normal and a confirmed our earlier data (1) that the cells of heterozygous heterozygous familial hypercholesterolemic into lipid-free medium. The data were obtained from the same experiments as familial hypercholesterolemics responded to incubation in a described in Fig. 4. For experimental details, see text. lipid-free medium not only with a greater utilization of acetate, but also with a greater induction of HMG-CoA reductase than per cent of their cellular sterol to the medium, reaching 40% of seen in normal cells. We suggest that this abnormally high inthe total [aH]sterol after 6 hours as compared to only 26% loss duction results from a more rapid dissociation of a repressor of by the normal cells (Fig. 5). The continued transfer of sterols the system coding for HMG-CoA reductase from the heterozyfrom the heterozygous cells to the medium over 6 hours is con-gous than from normal cells. This suggestion seems to be borne 135,000 f 5,000 dpm/l08 normal cells and at 109,000 f 2,000 normal and heterozygous cells, in contrast to the 26 and 40% dpm/lOs heterozygous cells. Thus it appears, (a) that there is loss of labeled sterols from the normal and heterozygous cells, rea differential and specific loss of [3H]sterol from leukocytes to the spectively. Since we measured only radioactive squalene and medium, and (b) that this loss is much larger from the heterozy-sterol in these experiments, we cannot tell exactly how much gous cells than from the normal ones. sterol is represented by a certain number of disintegrations per min. But as the defect in familial hypercholesterolemia has DISCUSSION The data presented confirm and extend our observations reported earlier (1) on the usefulness of leukocytes for the study of sterol biosynthesis in man. We have provided evidence that the most severely limiting reaction of sterol biosynthesis in man, as in other species, is that catalyzed by HMG-CoA reductase. Since in cells incubated in a medium containing full serum, sterol synthesis was linearly proportional to time (cf. Fig. 2), we may calculate from the data of Table I that the mean rate of sterol synthesis from acetate was 2.25 f 0.3 pmol/l08 cells per hour corresponding to the generation of 13.5 f 0.6 pmol of mevalonate.
In addition, 3.6 f 0.6 pmol of mevalonate are accountable in squalene giving an hourly total of 17.2 pmol of mevalonate formed from acetate per 108 leukocytes.
This value is in very close agreement with the HMG-CoA reductase activities measured in "uninduced" cells, 10 to 20 pmol/lO* cells per hour. From the incorporation of mevalonate into sterols, one might conclude that the maximum sterol-synthesizing capacity of the leukocytes per lo8 cells was 12 to 16 pmol per hour (cf .  Table I). However, from the slopes of the curves on Fig. 2 beyond the 6th hour of incubation, we calculate that the rate of sterol synthesis in the normal cells increased to 17 to 20 pmol/ been shown in both fibroblasts and leukocytes to affect HMG-CoA reductase, the rate-controlling enzyme in sterol biosynthesis, there is no reason to suspect that the heterozygous cells have a greater complement of enzymes in the biosynthetic pathway to cholesterol after mevalonate; that is why we chose mevalonate rather than acetate for these experiments.
Proof that the sterol loss is associated with the primary defect in the leukocyte must await further study including the demonstration that the nonisotopic sterol content of the heterozygous cells is equal to or lower than the normal at a time when the heterozygous cells are producing more st,erol from acetate. Khachadurian and Kawahara have observed precisely such a situation in homozygous fibroblasts (26).
It is pertinent to discuss the observations of Goldstein and Brown and their associates (2-6) who have studied the repression of HMG-CoA reductase in human fibroblasts cultured from the skin of normal, heterozygous, and homozygous familial hypercholesterolemic individuals by serum lipoproteins.
They have found that the high levels of HMG-CoA reductase, induced in normal fibroblasts and in fibroblasts of heterozygous familial hypercholesterolemics by a serumless medium, or by a lipoprotein-deficient serum, could be repressed by low density lipoproteins. However, higher concentrations of LDL-cholesterol were needed for the repression of the enzyme in the heterozygous than in the normal fibroblasts. The fibroblasts of homozygous familial hypercholesterolemics grown in 10% fetal calf serum had HMG-CoA reductase levels that were 40 to 60 times higher than those of normal fibroblasts similarly grown and, in contrast to the behavior of the normal cells, the HMG-CoA reductase levels in the homozygous cells did not increase upon the change to a lipoprotein-deficient medium, neither could the enzyme be repressed in the homozygous cells by as much as 2 mg of LDLcholesterol per ml, whereas 25 to 40 pg per ml were sufficient to repress the enzyme in normal cells. Goldstein and Brown attribute the cellular defect in familial hypercholesterolemia to a defect of LDL binding to the cells and of LDL degradation, which they have demonstrated with the aid of '%labeled LDL (6). They attribute the abnormality to a genetic mutation that results in a failure of the synthesis of specific LDL receptors on the cell surface and a failure of degradation of LDL and hence transport of cholesterol into the cell.
It is not possible to decide at, present whether the phenomena we have observed in leukocytes, the abnormally high rate of induction of HMG-CoA reductase in heterozygous cells and the high rate of loss of sterols from such cells to a lipid-free serum, and those observed by Goldstein and Brown, the impaired binding and degradation of LDL by heterozygous and homozygous fibroblasts, are related or not, as we have not studied LDL binding to leukocytes.
Whatever might be the correlation of the recorded phenomena, the observations on cultured fibroblasts and on freshly isolated leukocytes indicate a genetic abnormality in the control of the synthesis of HMG-CoA reductase. Our interpretation of the observations differs from that given by Goldstein and Brown for the origin of the familial hypercholesterolemic abnormality in a fundamental way in that we propose that the abnormality results from a mutation with a defective binding for a sterol repressor synthesized within the cell and with a consequently greater loss of sterol from the cell. Once a sufficiently high level of extracellular cholesterol has been built up, rates of cholesterol synthesis and rates of loss of sterol from the cells, are depressed to normal values. Goldstein and Brown imply that the primary control of the phenotypic expression of HMG-CoA reductase lies not within the cell, but comes from without through the transfer of cholesterol from extracellular LDL to the cell. Only further work can decide between these contrary views.
In spite of the evidence for the existence of a genetic abnormality in the control of the phenotypic expression of HMG-CoA reductase in fibroblasts and leukocytes, a full understanding of the origin of familial hypercholesterolemias is still not at hand. The limited information available from in t&o studies does not suggest the existence of a derepressed or partially derepressed state for the synthesis of HMG-CoA reductase, at least not in heterozygotes.
All in wivo studies have failed to show an overproduction of cholesterol (27, 28). Moreover, Miettinen (29), using an in wivo test for HMG-CoA reductase activity finds that, if anything, the reductase in heterozygous familial hypercholesterolemics is lower than normal.
Indeed, the HMG-CoA reductase activity in the liver biopsy of one heterozygote taken to date was found to be lower than normal (30). The greatest puzzle has been provided by the case of the 12.year-old girl, J.P., a homozygous familial hypercholesterolemic and whose fibroblasts were the subject of the very first studies of Goldstein and Brown (2,3). She had suffered a myocardial infarction, had progressively debilitating angina and a serum cholesterol of 800 to 1000 mg/lOO ml which was refractory to all previous forms of therapy. Starzl et al. (31) tried prolonged intravenous alimentation of J.P. For the first time her serum cholesterol fell by 300 mg/lOO ml. Because of these encouraging results, an endto-side portacaval shunt was made on the child. Subsequently, her serum cholesterol fell to the 200 to 300 mg/lOO ml range, her xanthomas disappeared, she returned to school, and grew at a normal rate. It is difficult to reconcile these results with a structural gene mutation for LDL binding and degradation as has been suggested from the study of J.P.'s fibroblasts in culture.
The fibroblasts are studied under conditions which do not exist in tiuo. So are the leukocytes in lipid-free sera. However, the situation in wivo, i.e. a high concentration of LDL, is mirrored by the leukocytes before incubation or after incubation in full serum, and our results under these conditions are in perfect agreement with the in vivo studies in man (27,28).
We believe the transfer to lipid-free serum may also reflect an in viva phenomenon, derepression.
Because fibroblasts grown in 10% fetal calf serum have adapted to a cholesterol concentration one-tenth of that seen even in the interstitial fluid drained from the foot (32), and because they must be grown in culture for at least five generations before being studied, one cannot assess derepression as it must happen in tivo, namely, in the presence of a high LDL concentration.
That is probably why the fibroblasts were taken routinely to a completely derepressed state and then repression studied.
This cannot occur in &JO. What may occur in tivo is periodic derepression.
The diurnal variations of HMG-CoA reductase levels in the liver documented for the rat (e.g. 33-35) may also exist in man, and in familial hypercholesterolemia the HMG-CoA reductase levels may rise higher to the postprandial stimulus, whatever that stimulus might be, than in normal individuals. hilalamos et al. have shown (36) that when human leukocytes were separated from whole blood and then incubated in their own plasma with [I-14C1acetate, they released [%]sterols into the medium within 6 hours. We have shown that heterozygous cells lost [3H]cholesterol but not [3H]squalene in the lipid-free sera more readily than did normal cells. Under steady state conditions in tivo, the heterozygotes do not synthesize cholesterol more rapidly (27, 28), but they have a higher extracellular concentration of cholesterol.
We feel that this may be a homeostatic mechanism to compensate for a defect which leads to a higher rate of loss of cholesterol from cell membranes.
If this hypothesis is correct, one would expect that the cellular content of cholesterol in heterozygotes is no higher than in normals.
Indeed, except for specific sites of connective tissues and blood vessels, there is no evidence that cells of heterozygous hypercholesterolemits have a higher cholesterol content than normal.
The cholesterol concentration in a liver biopsy from a person with type II hyperlipoproteinemia was no higher than the cholesterol concentration in liver biopsies from persons operated on for cholelithiasis without biliary obstruction and with normal serum concentrations of cholesterol (30). Moreover, Maurizi ef al.
(37) did not find any correlation between liver cholesterol content and atherosclerosis in a large autopsy series. The cholesterol concentration in fibroblasts from two individuals with homozygous familial hypercholesterolemia was no higher than that in fibroblasts from six normals, despite the fact that the homozygotes incorporated 10 times more acetate into cholesterol than the normals (26). Hence, one must assume that the homozygotes were losing the synthesized cholesterol from their cells into their environment faster than the normals.
The defect in LDL binding that Brown and Goldstein (6) have observed in familial hypercholesterolemic cells may in fact be a consequence of the mass action effect of cholesterol loss from the 1% loss from the surface of those cells. The leukocytes, parallel with the cultured fibroblast, is, we 16. believe, a most useful and readily accessible human cell for prob-17.
ing the many unanswered questions as to the control of sterol biosynthesis in man.

18.
Acknowledgments-We thank most warmly all the patients 19.
and members of the UCLA Medical Center for their willing co-20. operation in our studies. We also thank Mr. V. P. Heuring of the Proctor & Gamble Co. for a gift of the KYRO EOB deter-21. gent. Dr. Alan Polito assisted with some of the experiments.