Low Density Lipoprotein Receptor Activity in Homozygous Familial ...

Low Density Lipoprotein Receptor Activityin Homozygous Familial Hypercholesterolemia Fibroblasts* (Received for publication, November5, 1981)

Clay F. Semenkovich, Richard E. Ostlund, Jr., Richard A. Levy, and Steven R. Osa From the MetabolismDit,ision, Department of Medicine, Washington Uniniuersity School of Medicine. St. Louis, Mwouri 6.3110

We have identified specific low affinity low density remains unknown. Goldstein and Brown ( 3 ) have postulated lipoprotein (LDL) receptors in skin fibroblasts from that the clinical phenotype can be accounted for by three two patients previously classified as having LDL recep- disorders that areclinically similar but biochemically distinct. tor-negative homozygous familialhypercholesteroleThus,culturedfibroblasts from patients with homozygous mia (FHC). K,,, and maximum capacity for cell-associ- FHC are classified as receptor-negativein which less than 21' ated and degraded '"I-LDL were determined by two of normal LDL receptor activityis detectable, receptor-defecindependent methods, a traditional technique in whichtive in which 2-25% of normal LDL receptor activity is deincreasing amountsof '"I-LDL were added until recep- tectable, and "internalization-defective" in which LDL receptor saturation was achieved and a new technique in tors are able to bind LDL particles normally but LDL is not which the displacement of a small amount of l"I-LDL taken up by the cell. This paper focuses on the receptortracer was observed during the addition of variable negative form of homozygous FHC. amounts of unlabeled LDL. The K , for specific cellLDL receptor-negative fibroblasts clearly have a gross reassociated 12'1-LDL in FHC cells was 3.5-7.3 times that is unresolved of normal cells and themaximum specific capacity was duction in "'I-LDL binding activity,butit is present. Severallines whether or not some residual activity reduced to 11%of normal. Thus, some FHC cells have of evidence suggest the existence of significant LDL receptor reduced affinity as well as reduced capacity for LDL. The FHC cell receptors share many but not all prop- activity in "receptor-negative'' cells. LDL receptor activity erties of the normal skin fibroblast LDL receptor. Spe- can be assessed either directly by the measurement of ""Icific degradation of bound lZ5I-LDLoccurred concomi- LDL bound to cultured fibroblasts or indirectly by a number effects of LDL on cellular tantly with LDL binding and was greatly reduced by of methodsthatmeasurethe the addition of chloroquine, an inhibitor of lysosomal metabolic functions. Indirect methods include determinations function. Preincubation of FHC cells with cholesterol of the effect of LDL on 3-hydroxy-3-methylglutaryl-Co A or LDL resulted in significant suppression of receptor reductase activity, cholesteryl ester formation, and the release function. Modification of lysine residues of LDL abol- of amino acids by LDLdegradation. Goldstein et al. (4) ished receptor activity in both normal and FHC cells. classify cells as "receptor-negative'' if LDL fails to stimulate Treatment of FHC cells with compactin, a cholesterol any incorporation of oleate into cholesteryl esters. However, synthesis inhibitor, resulted in significant increases in when receptor-specific "'I-LDL binding was studied in five specific '"1-LDL binding and degradation comparedto fibroblast strains that were later classified as LDL receptorFHC cells without compactin treatment. Normal cells negative on the basis of oleate incorporation studies(4), these also showed increases in '"I-LDL binding and degracells bound 11.2 +- 3.7% of the expected amount of LDL ( 5 ) . dation with compactin treatment, but the mean perBreslow et al. (6) found 12% of normal '"I-LDL specific centage increase in specific lZ5I-LDLdegradation was binding in fibroblasts from a patient apparently classified as significantly greater in FHC cells (strain GM 2000, 160 LDL receptor-negative. Using a dose-response curve, meask 18%)than in normalcells (29 2 8%). urable levels of specific degradation of LDL (15%' of normal) were demonstrated by Goldstein and Brown in fibroblasts from a typical homozygote withthe receptor-negativeform of Homozygousfamilial hypercholesterolemia is a well de- FHC (4, 7 ) .Haba et al. (31) found from 0-7.54>of the expected scribed clinical phenotype which is accompanied by a defispecific LDL internalization and degradation in receptor-negciency in the cell surface receptor for LDL' and accelerated ative fibroblasts. Fung et al. (10, 11) have repeatedly demonatherosclerosis with premature death (1, 2). Despite elegant strated LDL-mediatedsuppression of hydroxymethylglutaryl20 CoA reductase activity andof acetate incorporation into stercharacterization of the clinical syndrome over the past years by Khachadurian, Frederickson and Levy, and Brown ols in receptor-negative fibroblasts. This is evidence for speand Goldstein, the exact nature of the genetic defects involved cific receptor activity since LDL entering fibroblastsby nonspecific pinocytosis evidently does not regulate sterol synthe* This work was supported by a grant from the American Heart Association. The costs of publication of this article were defrayed in sis or esterification ( 7 , 12, 15). Using fluorescent and immuindividualfibroblasts, part by the payment of page charges. This article must therefore be nofluorescent microscopy tostudy hereby marked"adeertisement" in accordance with18 U.S.C. Section Kruth and Vaughn (35) reported that some homozygous re1734 solely to indicate this fact. ceptor-negative cells bind small amounts of LDL and accu' The abbreviations used are: LDL, low density lipoprotein; FHC, mulate intracellular cholesterol. familialhypercholesterolemia; LPDS, lipoprotein-deficienthuman The present studywas undertaken to characterize receptor serum;Hepes, 4-(2-hydroxyethyl)-l-piperazineethanesulfonicacid; activity present in fibroblasts from patients with the receptorNBCS, newborn calf serum; MEM, Eagle's minimum essential medium: T M U , tetramethvlurea; mEGF, mouse epidermal growth facnegative form of homozygous FHC. We have identified specific tor. low affinity LDL receptor activity in these fibroblasts which

12857

12858

LDL Receptors in Homozygous Hypercholesterolemia

squares. The slope of the line is -l/K,,, (K,,, = Michaelis-Menten constant) and the intercepton the abscissa is the maximum capacity. Determinations ofK,, and maximum capacity for ""I-LDL were also accomplished from analysis of independent data obtainedvia the EXPERIMENTALPROCEDURES displacement of asmall andfixed amount of labeled LDL by variable Materials-Thediphosphatesalt of chloroquine, Hepes buffer, quantities of unlabeled LDL. The rationalefor this technique is given and cholesterol were purchased from Sigma. Cholesterol was repuriin the text. In calculating the K,,,, the following assumptions were fied twice by the method of Fieser (32). Calcium chloride was purmade: 1) labeled and unlabeled LDL compete identically for binding chasedfromMallinckrodtChemicalWorks. Dextransulfate was sites; 2) binding follows Michaelis-Menten kinetics; 3 ) nonspecific purchased from Calbiochem-Behring. Compactin in the lactone form binding of ""I-LDL is not a function of total (labeled + unlabeled) was obtained from Dr. A. Endo, Sankyo Co., Ltd., Tokyo, Japan. The LDL concentration. The following definitions are given. Q = total lactone was converted to the sodium salt by heating in base; 0.385 concentration of LDL (labeled + unlabeled) added to the medium, mM compactin was subsequentlyadjustedto physiologic pH and pg/ml; T = concentration of "'I-LDL tracer in pg/ml; Bc2= specific stored a t -20 "C (14).Newborn calf serum was purchased fromGibco, cell-associated LDL (labeled + unlabeled), ng of LDL/mg of cell Inc., Grand Island, NY. ''?I was purchased from Amersham. Radiaprotein; BT = specific cell-associated '"I-LDL tracer, ng of '"I-LDL/ tion-sterilizedplastic 35-mm diameter six-well cluster dishes were = maximum specific capacity for cell-assomg of cell protein; Bmilr,, purchased from Costar Corp., Bedford, MA.LDL (density 1.019-1.063 ciated LDL(labeled + unlabeled LDL), ngof LDL/mg of cell protein. g/ml) from normal, healthy donors was prepared by ultracentrifugaFrom the Michaelis-Menten theory: tion and iodinated as previously reported (7, 15). Unlabeled LDL (Free LDL) (free receptors) prepared from an unrelated patient with homozygous FHC was used = K,,, for displacement of specifically bound "'I-LDL. This FHC LDL was (specifically cell-associated LDL) indistinguishable from LDL prepared from normal donors indisplacing "'I-LDL in cultured fibroblasts.Lipoprotein-deficient human Since the amountof LDL bound is negligible with respect to the total amount of LDL present, the free LDL present in the medium = 4. serum was prepared from the bottom fraction after ultracentrifugation And since the concentration of free receptors is B,,,.,,,, - Bu, this and washing a t a density of 1.215 g/ml (7). Fortheexperiment presented in Table VIII, LPDS was shaken for 4 h at room temper- equation becomes: ature with 125 mg/ml of silicic acid (SI-R,Sigma) followed by sedimentation of the silicic acid. This reduced the total cholesterol content of the LPDS toless than 1 pg/ml(30). Stock LPDS solutions were adjusted to 70 mg of protein/ml (100% LPDS). All media were Solving for BQ: sterilized by filtration through 0.22-pm filters. Receptor grade mouse epidermal growth factor (Collaborative Research, Waltham, MA) was iodinated as previously described (8).Human "'I-thrombin (9) was a gift of Dr. Douglas Tollefsen and topical bovine thrombin was purchased from Parke-Davis. Expressing specific cell-associated "'I-LDL ( B T )in terms of the total Cells-Skin fibroblast strains fromtwo patients with the "receptor- amount of total LDL (labeled + unlabeled, Bu). negative" form of FHC (lines GM 2000 and 1915) were purchased from the Genetic and Mutant Cell Repository, Camden, NJ. Consistent results were obtained using different shipments of the same cell types over 3 years. Skin fibroblasts from normal controls were obtained by skin biopsies of the deltoid area. Cell plating density was Substituting Equation 2 into Equation 3 yields: 10"cells/35-mm well. The cells weregrown for 4 days in growth media consisting of Eagle's minimum essential medium containing 15-20?; NBCS and antibiotics (15, 23). Cells were then washed twice with Puck's saline G (16) and 1 ml of MEM & 10%LPDS (7 mg of protein/ Inverting: ml) was added to eachwell to induce receptor activity.After 72 h, the medium was removed and cellswere then assayed for '""I-LDL binding and degradation as previously described (15, 23). Cells were incubated for 5 h a t 37 "C with "'I-LDL in MEM + 10% LPDS k unlabeled LDL. The medium was removed and assayed for trichloThis is the equationof a straight line ~ / B T as the dependentvariable roacetic acid-soluble non-iodidedegradation products. Degraded LDL and 4 the independent variable. TheX intercept = -K,,2. Therefore, was determinedaftersubtraction of ablank measured in dishes B,,,,, may be calculated as without cells (15, 23). Following aspiration of the medium, the dishes 1 were immediately placed on ice and washed. In some experiments, B,,,,,, = ( T )(slope as calculated from least squares analysisof line) the surface-bound LDL was eluted for 1 h at 4 "C with 4 mg/ml of dextran/S04 (Calbiochem) in Puck's saline G and internalized LDL was then determined by dissolving the cells in NaOH. Cell-associated RESULTS "'I-LDL was determined eitherby dissolving the washed cells directly in 0.625 N NaOH for 20 min at room temperature withoutfirst eluting Characterization of LDL Receptor Activity in FHC Cellsthe surface LDL or as the sumof dextran/S04-releasable + internalA dose-response curve for cell-associated '2511-LDLin LDL i :ed LDL in experiments where there was too little surface-bound receptor-negative GM 2000 cells is presented in Fig. 1. Al:.DL to analyze separately. Calculations-Statistical comparisons were performed by the Stu- though the absolute level is low, appreciable cell-associated "'I-LDL displaceable by unlabeledLDL is apparent and easily dent's t-test with mean e standard error of themeanpresented. Unless otherwise stated, specific "'I-LDL binding or degradationwas distinguishable from binding to culture dishes without cells computed as the difference of values determined in the absence (total) (blank dishes). In subsequent data, receptor-specific cell-asand thepresence (nonspecific)of 20-fold or greaterexcess of unlabeled sociated or degraded '"I-LDL was calculated by subtracting LDL. The estimated standard error of this difference was computed + SEM~l,,,,,;,,,,f,,where the number of nonspecific values(seenin the presence of cells but with as SEM (specific) = dSEM:,,L*~ excess unlabeled LDL) from total values (seen in the presence samples for total and nonspecific binding was identical. Experiments of cells without excess unlabeled LDL). The estimated stanwere routinely performed in triplicate. K,,, and maximum capacit.v for '"I-LDL were calculated from dard error of thisdifference is routinelypresented in the kinetic data determined in triplicate or sextuplicate by plotting acfigures and tables. Blank culture dishes reproducibly manicording toScatchard (17, 18, 34) theratio of cell-associated (or festeda minor degree of specific bindingwhich has been degraded) ""I-LDL (nanograms of LDL protein/mg of cell protein) to free '2'I-LDL (micrograms LDL protein/ml) on the ordinateuersus disregarded in subsequent calculations. No LDL specific degradation productswere observed in blankculture dishes. Data the cell associated (or degraded) "'I-LDL (nanogramdmg) on the of the experimentof Fig. 1were recalculated as nanograms of abscissa and fitting the points atostraight line by the method of least

is distinct from nonspecific uptake of LDL by cells and binding of LDL by blank culture dishes.

12859

LDL Receptors in Homozygous Hypercholesterolemia "'I-LDL specifically cell-associated and degraded per mg of cell protein and are displayed in Fig. 2. Specific LDL degradation occurred and increased in parallel with cell-associated LDL, suggesting that the cell-associated material was transferred to lysosomes. This was further supported by the data of Table I in which addition of 60 ~ L Mchloroquine (lysosomal enzyme inhibitor) to the assay solution during a 5-h incubation in the presence of 50 pg/ml of "'I-LDL resulted in the nearly complete inhibition of LDL degradation in FHC cells accompanied by a significant increase in the amount of internalized LDL. The time course of specific cell-associated and degraded LDL in receptor-negative GM 1915 cells is shown in Fig. 3. The FHC cells qualitatively resembled normal cells. Cellassociated '"I-LDL values rose progressively after additionof '"I-LDL and no degradation was observed until 45 min after which degradation roselinearlywithtime.Maximum cell associated '251-LDL was observed at 5 h (data not shown).

TABLEI Effect of chloroquine on LDL binding and degradation After 3 days of initial cell growth and 2 days of LDL receptor induction, the medium was removed and replaced with 0.7 ml of MEM + LPDScontaining 50 pg/ml of '"'I-LDL f 500 pg/ml of 5-h unlabeledLDLwithorwithout 60 p~ chloroquine.Aftera incubation at 37 "C, the cells and media were processed for LDL binding and degradation. ~

DextranAssay conditions

ble binding

50 pg/ml IL5II-LDL only 50 pg/ml '"I-LDL plus 60 PM chloroquine

LDL inter"lized

LDI, tlegraded

5.54 f 1.58 52.3 f 4.7 135 f 15.9 f 4.3 6.7 f 13.7 4.62 f 0.8298.2 ( p < 0.001) ( p < 0.001)

Little LDL binding and no significant degradation were observed in blank culture dishes. Specificity of FHC cell receptor for lipoproteins is demonstrated in Table 11. In both normal and FHC cells, unlabeled LDL reduced the specific degradation of ""I-LDL to a similar extent. However, methyl LDL and acetoacetyl LDL,lipoproteins modified a t critical lysine residues involved in receptor binding (27), did not compete significantly for either normal or FHC receptors. Since only a very small per cent of the radioactivity added to the culturedcells became cell-associated, the natureof this material was further examined. The '"I-LDL tracer added contained 3.6% of material that remainedin the organic phase after aFolchlipid extractionand wash (26) followed by evaporation of the organic solvent. Internalized "'I radioactivity dissolved in 0.5 ml of 0.625 N NaOH from an experiment identical with thatdescribed in Table IXwas neutralized with 0.5 ml of 1 N HCl and immediately vortexed with 1.0 ml of 2 5 7 5 15 30 100 chloroform. The chloroform phase was removed, dried, CONCENTRATION OF 1 2 5 1 - tI ~~~ ~ / ~ I I counted,andcomparedtothat remainingin theaqueous IN ASSAY MEDIUM phase. In GM 2000 cells, 7.5% of the radioactivity was chloFIG. 1. Dose-response curve of cell-associated '''I-LDL for roform-soluble, whereas in normal cells 3.6% was chloroformreceptor-negative GM 2000 fibroblasts. Four days after plating at soluble. Hence, this material did not appear to be primarily IO5 cells/well in MEM + 20'% NBCS, the medium was removed, the cells were washed twice, and 1 ml of MEM + 10%LPDS was added. lipid in either normal or FHC cells. Since apoprotein B, the After 3 days of LDL receptor induction, the medium was removed protein constituent of LDL, is insoluble in aqueous solutions and replaced with 0.7 ml of MEM + 10% LPDS containing 2.5-100 of TMU whereas other apoproteins are soluble in TMU (25), pg/ml of '"I-LDLwith or without a 20-fold or greater excess of the solubility of bound radioactivity in this solvent was tested unlabeled LDL and thecells were incubated for 5 ha t 37 "C. 0, GM 2000 cells without excess unlabeled LDL; 0, GM 2000 cells with excess (Table 111). In both normal and GM 2000 cells, over 87% of the bound counts were TMU-insoluble. Hence, the cell-assounlabeledLDL; 0, blankculturedisheswithoutexcessunlabeled LDL; B, blank dishes with excess unlabeled LDL. Brackets indicate ciated radioiodinated material appears tobe "'I-LDL apoproS.E. for triplicate culturewells. tein and not a contaminant. Saturability of FHC cell receptor activity is demonstrated by the displacement curveshown in Fig. 4. In the presence of a constant amount of '"I-LDL (50 pg/ml) in the medium, the amount of cell-associated '"'I-LDL tracer decreased from 104 k 2.8 ng/mg of cell proteinto 18.2 k 1.1 ng/mg asthe concentration of unlabeled LDL in the assay medium was increased from 0 to 2000 pg/ml. In the same experiment, the amount of degraded tracer decreased from 223 +- 15.4 ng/mg of cell protein to 51.9 k 12.9 ng/mg. The displacementof I2'ILDL by large amounts of unlabeled LDL does not appear to be an artifactsince similar amounts of unlabeled LDL did not displace specifically bound '"I-thrombin or I-mouse epidermal growth factor from normal cell receptors (Table IV). LDL binding in FHC cells was calcium-sensitive (Table V). The effect of calcium on specific LDL binding was assessed by a modification of the technique used by Goldstein and I I 5 75 15 10 Ion Brown to characterize high affinity receptor activity (7). In l 2 5 1 - t ~IN ~ ASSAY ( p g / m l ) the presence of 50 pg/ml of '"I-LDL, both specific dextranFIG. 2. Dose-response curve of specific cell-associated '"ILDL and degraded LDL in GM 2000 cells. Data from the experi- SO4-releasable surface-bound and internalized LDL were sigment of Fig. 1 werereplottedafternormalizationtocellprotein. nificantly increased in GM 2000, GM1915, and normal cells Brackets indicate S.E. after the additionof 2 mM calcium to theassay medium. This


12860

K,,, (reciprocal of affinity) of these cells was 127 pg/ml compared to 16.5 pg/ml in the normal cell type. The mean K,,, for receptor-negative cells (103 pg/ml) was 4.6 timesthat of normals, while the mean maximum binding capacity (364 ng/ TABLE I1 Specificity of the FHC LDL receptor Normal and GM 2000 cells were grown for 5 days in MEM + 159 NBCS and then incubated for 3 days in MEM + 10% LPDS. The mediumwasremovedandthecells (3-12 dishes/condition)were incubated for 5 h a t 37 "C in MEM + 10% LPDS containing I00 pg/ ml of"'1-LDL and the specified amount of unlabeled lipoprotein. Numbers in parentheses indicate percentage of no addition of unlabeled lipoprotein. -

""I-I,DI. degradation GM 2000 cells

Normal cells ng/mp/5 h

15

45

120

300

MINUTES

Noaddition of unlabeled lipoprotein +I20 pg/ml LDL +2000 pg/ml LDL +I20pg/mlmethylLDL +120 pg/ml acetoacetyl LDL

6250 rt 243 (100%)

251 f 12 (1009;)

2040 f 77 (33X) 102 f 5" 401" f 51 (6.4%') 19 f 4" 5690 -C 304 (919) 224 & 8 6060 f 202 (97%) 243 f 21

(414) (7.6%) (89%) ($477)

" p < 0.001 with respect to no addition.

TABLE I11 Solubility of bound "-'lcounts in TMU Triplicate dishes of fibroblasts were grown in MEM + 15%) NBCS for 5 days and then in MEM + 10% LPDS for 2 days. The medium was aspirated and the cells were incubated a t 37 "C for 5 h with 92 pg/ml of "'I-LDL f 2.0 mg/ml of unlabeled LDL. T h e cells were washed inthe usual fashion and then scraped with a Teflon policeman into 0.15 M NaCl and sedimented. The cell pellet was taken up in a 1:l mixture of 0.15 M NaCl and TMU, 150 p g of carrier LDL was added in 6.2 11, and the tubes were vortexed and incubated at 37 "C for 1 h. The TMU-soluble and insoluble fractions were separatedby centrifugation for 5 min in a Beckman microfuge and counted. Numbers in parentheses are percentages of total recovered counts/min found in the TMU oellet.

cpm/dish

Normal cells TMU supernatant T M U pellet GM 2000 cells TMU supernatant TMU pellet

1050 f 81 7270 f 520 (87.4%) 79 f 3 542 f 41 (87.3%)

86 f 14 576 f 18 (87.0$,) 17 f 3 182 f 16(91.5'F)

MINUTES FIG. 3. T i m e course of "'1-LDL interaction. After 4 days of initialgrowthand3days of receptorinduction,themediumwas removed and replaced with 0.7 ml of MEM + 10% LPDS containing 40 pg/ml of "'I-LDL with or without 750 pg/ml of unlabeled LDL. Cells were incubated at 37 "C and then harvested and processed for specific cell-associated ""I-LDL and degraded LDL a t specified time intervals.

effect of calcium on specific binding could not be accounted for by binding to blank culture dishesalone. Calcium sensitivity was theonepropertynot observed by Danaetal. in studving the binding of "'I-LDL to glass beads (33). Calculation of Binding Parameters-Two methods were i 1" I " ' IO00 2 000 used to calculate the apparent K,,, and maximum capacity for UNLABELED LDL ADDED ( p g / m l I specific cell-associated '"'I-LDL and degraded LDL in GM FIG. 4. Displacement of '"1-LDL b y unlabeled LDL in GM 2000 cells (Table VI). First,a Scatchard analysis was applied 2000 cells. Following growth and receptor induction asin Fig. 1, the to the equilibriumdose-response data of Fig. 2 in which medium was removed and replaced with MEM+ 10%LPDS containincreasing quantities of ""I-LDL were added with and without ing 50 pg/ml of '"1-LDL + 0-2000 pg/ml of unlabeled LDL, and the excess unlabeled LDL (Fig. 5). It is clear that only low affinity cells wereincubated at 37 "C for 5 h. Cell-associated '""I-LDL is binding (flat slope)is present in GM 2000 cells. The apparent presented without subtraction of nonspecific binding. -I

"I

1 "-"

LDL Homozygous Hypercholesterolemia Receptors in

12861

TABLE IV mg) was 11.5% of normal (Table VI). Similar results were Effects of unlabeled LDL on binding of thrombin and mouse found for '"I-LDL degradation. The receptor-negative cells epidermal growth factorto fibroblasts. thus have bothreduced affinity and capacity for LDL. A, normal human fibroblasts were plated at lO'/well in MEM + One limitation of the Scatchard method is that, it is not 15% NBCS. T h e medium was removed after 5 days, the cells were practical to use concentrations of added '"I-LDL significantly washedwithsalineG,and 25 mM Hepes-bufferedMEM, pH 7.3, above the K,, of the receptor-negative cells. Nonspecific II"1containing 0.1% bovine serum albumin was added for 5 min to deLDL binding rises in proportion to the added 12'II-LDLconcrease nonspecific binding. This medium was removed and 0.7 ml of thesamemediumcontaining0.83 nM '"I-a-thrombin f 200-fold centration, whereas the low level of specific binding in FHC excess of unlabeled bovine thrombin or unlabeled LDL was added. cells is constant. Therefore, the data become less reliable as T h e cells were incubated for 60 min at 37 "C in air and then washed large quantities of '"I-LDL are employed. To circumvent this, as for LDL binding experiments. The cells were dissolved in 0.625 N an analysis of the I2'I-LDL displacement curve of Fig. 4 was NaOH and counted for total cell-associated binding. B, normal human performed as described under"Experimental Procedures." fibroblasts were plated at 10r'/welL in MEM + 15% NBCS for 4 days. This technique has the advantage of a constant and low level The medium was removed, the cells were washed with saline G, and of nonspecifically bound '"I-LDL counts.The analysis is 1 ml of 10% LPDS in MEM was added. This media were removed after 3 days and 0.7 ml of the same medium containing 16 ng/nd of based upon the principle that ascold LDL is added in increas'"I-mEGF f 20-fold excess of unlabeled mEGF or unlabeled LDL ing amounts to I2"I-LDL in the assay medium, the unlabeled was added. Thecells were incubated 40 min at 37 "C in 5% CO?, then material will compete for receptor binding. However, the washed as for LDL binding experiments. T h e cells were dissolved in degree of competition will be imperfect depending on the per 0.625 N NaOH and counted for total cell-associated bindine. _

_

_

~

"'I-a-thrombin bound

~ " _ _ _

fm/mgprotezn

A. "'1-a-thrombin +200 pg/ml unlabeled bovine thrombin +100 pg/ml LDL +lo00 pg/ml LDL

338 f 3.5 8.1 1.8 325 f 12.i 337 f 12.6

*

1 2 ' ~ . m ~ ~ ~

"I-mEGF

bound pg/rngprofezn 2350 f 74

B.

85 -C 12.2 2160 237 2430 f 133

+320 ng/ml unlabeled mEGF + 100 pg/ml LDL +1 OOO wdml LDI,

TABLE

*

v

Calcium sensitiuity of LDL binding Cells were grown for 3 days in MEM + 15% NBCS and washed, and 1.0 ml of MEM + 10% LPDS was added to each well. After 2 days of receptor induction, the medium was removed. One group of cells was then washed with calcium and magnesium-free Puck's saline G and 1.5 ml of 50 mM Tris-CI (pH 7.5), 0.1 M NaCl buffer containing 50 pg/ml of "'I-LDL with or without 2 mg/mlof unlabeled LDL was added to each well. A second group of cells was treated identically except CaCL was added to the assay medium yielding a final calcium concentration of 2 mM. T h e cluster dishes were sealed with parafilm, incubated in airat 37 "C for 3 h. and then Drocessed for LDL bindine. 2 mM Ca'*

Dextran-SO,-releasable binding

-

6.84 -C 2.97 p < 0.01 26.3 k 1.93 4.58 f 1.56 30.2 -+ 8.87 0.05 27.7 f 2.16 54.0 f 4.97 < O.O1 6.70 f 3.51 6.04 2.16 n.s.'l 26.9

ng/ulell

GM 2000

GM 1915 Normal Blank dishes I'

w

,

~

-

~

~

~

ng/luell

+ -

+ + + -

*

20.5 f 2.67 p < 0.001 78.9 f 5.97 < o,O1 30.2 f 1.26 82.3 f 9.07 o.O1 202 6.36 353 k 27.4 16.0 -C 0.78 p < 0.05 2.92

*

Not significant.

CELL ASSOCIATED '251-LDL (ng / m g ) FIG. 5. Scatchard analysis of specificcell-associated '9LDL. The data of Fig. 2 for specific cell-associated ""I-LDL were plotted according to the methodof Scatchard (18). The slopeis the negative of the apparent affinity for LDL and the x-intercept is the maximum capacity. 0, GM 2000 cells; A,normal cells.

TABLEVI '"I-LDL processing characteristicsof n o r m a l a n dFHC fibroblasts r is the correlation coefficient. Method of determination of K,,,

Cells

K,,, for specific cell.as. sociated LDL

Mcell-associaled a i m u m specific LIII,

K,,, for specific LDL Maximum specific degradation LDL degradation

~~

.

~

~

W/ml

GM

analysis Scatchard Displacement analysis GM

.

.

_

_

w/mg

2000 310 127 ( r = 0.99) GM 1915 418 79.4 ( r = 0.97) Normals cells (7lines) 22.6 f 5.6 3150 f 378 2000 48.7 ( r = 0.99) 159 Normal 6.66 ( r = 0.99) 1500 -____

_

_

~

g/ml

116 (r = 0.87) 81.3 ( r = 0.97) 16.3 f 4.5 ~

w/mg

777 1460 6900 f 760 ___-__


12862

cent saturationof the LDL receptors, which is itself a function of binding affinity. If receptors are presentin great excess, no competition will occur; if receptors are saturated with LDL, perfect competition takes place. The data from Fig. 4 (GM 2000 cells) and a normal cell type are plottedin Fig. 6 in such a way as to linearize the data, analogous to the LineweaverBurk method of enzyme analysis. In Fig. 6, the x-intercept = -Km. It can be seen that the apparentaffinities of FHC cells (48.7 pg/ml) and a normal cell type (6.6 pg/ml) differ markedly. By this method, maximum capacity for cell-associated '"'I-LDL of receptor-negative cells was 10.6%of normal (Table VI). The above parameters were computed for cell-associated (surface-boundplus internalized) '"I-LDL and degraded LDL. Although it was possible to measure dextran-S04-releasable cell surface binding (Tables I and V), thelevel of binding was too low to determinereliable K , and maximum bindingvalues. More precise surface binding data could be obtained from the measurement of cell-associated '""I-LDL at 4 "C, a condition which prevents LDL internalization and degradation(1).Receptor-specific '"I-LDL binding at 4 "C in three normal cell strains performed as specified in the legend to Fig. 7 demonr

8

0

15

30

45

60

'251-LDL.

75

90

105

120

135

pg/rnl

FIG. 7. Binding of '"I-LDL at 4 "C. Normal fibroblasts (A), GM ZOO0 (a), and dishes without cells (W) were prepared for "'I-LDL binding as describedunder"ExperimentalProcedures." The cells were cooled to 4 "C for 40 min and all subsequent operations were performed in a 4 "C cold room. '"'I-LDL 2 unlabeled LDL in MEM + 10% LPDS buffered with 25 mM Hepes to pH 7.3 was added and the dishes were incubated 2 h,followed by the usual washing proce2 extra IO-min washeswithalbumin-containing dureexceptthat buffer were done before washing the cells with albumin-free buffer. The cells were dissolved in 0.625 N NaOH for 20 min and counted. Values reported are receptor specific binding of sextuplicate wells. Protein content was45 0.8 pg/22-mm diameterwell in normal cells and 51 f 1.0 pg/well in GM 2000 cells.

*

TABLEVI1 "'I-LDL-degradation by fibroblasts preincubated in physiologic amounts of LDL Cells were plated in growth medium which was replaced after 4 days with MEM+ 10% LPDS containing80 pg/ml of LDL. Thecells were preincubated for 3 days with medium changed 24 h before the MEM experiment. The medium was then aspirated and replaced with + 10%LPDS containing 80 yg/ml of '"I-LDL k 1 mg/ml of unlabeled LDL and the cells were incubated 21 h a t 37 "C. Mean f S.E. of triplicate dishes is presented. .__ '''I-LDL specific tlegrada-

Cell type

tion

ng/mg/21 h

Normal heterozygote FHC FHC homozygote (GM

5820 k 336 3230 261 1970 f 173

*

1915)

strated half-maximum receptor saturation at3.8 k 1.0 pg/ml and maximum capacity of 109 .+ 15 ng/mg of protein. When similar studies were done onGM 2000 cells it was not possible consistently to demonstrate specific le51-LDLbinding a t concentrations of added '"I-LDL less than 10 pg/ml. However, at higher amounts of added '"I-LDL (Fig. 7), appreciablespecific binding activity was observed and thebinding capacity/mg of protein a t 125 pg/ml of added '"I-LDL was 35.5 -+ 2.6% of a normal strainanalyzed simultaneously. Half-maximum receptor occupancy occurred at 75 pg/ml of added '"I-LDL, consistent with reduced affinity of the GM 2000 receptor for LDL. Effect o f PhysiologicManipulationson LDL Receptor FIG. 6. Displacement analysis of cell-associated '"I-LDL. Activity-Up tothispoint, only datafrom cells grown in FHC (GM 2000) and normal cells were grown and assayed, as deLPDS havebeen presented. To approximate moreclosely the scribed in the legend to Fig. 4, by adding"'I-LDLandvariable physiologic extracellular environment, normal fibroblasts and amounts of unlabeled LDL. Normal cells received 5pg/ml of '""ILDL GM 1915 FHC cells were preincubated for 3 days in MEM + and FHC cells,50 pg/ml of "'I-LDL. Specific cell-associated '"'I-LDL was determined by subtracting binding observed in the presence of 10% LPDS containing 80 pg/ml of unlabeled LDL. Then, the of LDL medium was replaced with MEM + 109 LPDS containing80 2000 pg/ml of unlabeled LDL. Values for the total amount (labeled plus unlabeled) in the medium are plotted on the abscissa. pg/ml of '"I-LDL with and without 1 mg/ml of unlabeled The reciprocal of nanograms of radiolabeled LDL cell associated per LDL and the cells were incubated for another 21 h (Table mg of cell protein is plotted on the ordinate. This treatment allows VII). Specific degradation of '"I-LDL in FHC cells was 34% onetolinearizethedisplacementdata of Fig. 4andgraphically of that observed in normal cells. This experiment suggests K,,, determine the K,,, (K,,, = -x-intercept). For GM 2000 cells (a), of FHC fibroblasts is relatively = 48.7 pg/ml; for normal cells(A), K,,, = 6.6 pg/ml (see "Experimental thatthereceptoractivity greater compared to normals when receptors are not artifiProcedures").

LDL Receptors in Homozygrous Hypercholesterolemia

12863

FHCandnormal cells were preparedas describedin the legend to Table IX by 2 daysculture ingrowthmedium followed by a 48-h incubation in 10%LPDS f 4 p~ compactin. Specific '"I-LDL binding was then determinedin triplicate by incubationwith 100 pg/ml of '"'I-LDL for 5 h at 37 "e. Dextran-S04-releasable surfacebinding increased from 9.1 1.3 to 34.2 f 3.2 ng/mg, and dextran-SO?-resistantinternalized 13.4 to 287 + 8.2 ng/mg. '"I-LDL increasedfrom 128 Protein content of the wells was unchanged by compactin TABLE VI11 treatment (control,48.5 -+ 1.1pg/well, + 4 pM Compactin, 48.7 Down-regulation of LDL receptor bypreincubation with f 1.5 pg/well). The specific degradation of '"I-LDL was also cholesterol or LDL increased after compactin treatment in GM 2000 cells (Table Normal and GM2000 cells were washed and preincubatedfor 72 h IX). In seven experiments, the mean percentage increasesin in MEM + 10% silicic acid-treated NBCS LPDS containing either specific '"I-LDL degradation in response to compactin was LDL, cholesterol, or no addition. LDL receptor activity was deter160 18%in GM 2000 cells, but only 29 f 8%>in three control mined by washing twice with saline G and incubating 5 h at 37 "C with MEM + 10% silicic acid-treated NBCS LPUS containing 100 cell lines. The greater effect of compactin preincubation on pg/ml of"'1-LDL s 1.0 mg/ml of unlabeled LDL. Results are the binding and degradation in GM 2000 cells is surprising since mean from 3-6 wells/condition f S.E. All statistical comparisons are in GM 2000 cells one would expect less suppression of LDL with untreated cells of the same type. receptor function by anytrace lipoproteins remaining in Dextran-So4Internalized Degraded LPDS and thereforeless receptor induction by compactin. releasable

cially induced by LPDS. The ability of LDL or cholesterol preincubation to down-regulate LDL receptors of normal and FHC cells grown in LPDS is presented in Table VIII. LDL receptor activity was reduced in both cell types, but the FHC cells demonstrated more resistance to down-regulation than normal cells. The effect of 48-h preincubation with compactin, an inhibitor of cholesterol biosynthesis, was also studied. GM 2000

*

*

*

ng "'I-LDL s p e r i f k d l y hound or degraded/mg protem

DISCUSSION

The data presentedin this paper concurwith the studiesof Brown and Goldstein (1, 5 , 7 ) in that no high affinity LDL I64 8.4 502 & 17.3 12.0 0.83 receptor activity was found in skin fibroblasts from patients 122 f 11.0" 322 -C 26.4" 6.72 f 0 . 7 6 with the receptor-negative formof homozygous FHC (Fig. 5). However, specific but low affinity LDL receptor activity was 1660 f 69.0 4440 f 251 137 f 0.73 identified which is distinct from the nonspecific bulk phase 259 f 1 2 . 8 589 f 38.5" 20.5 f 2 . 6 pinocytosis of '"I-LDL measurable in FHC cells in the presence of large amounts of unlabeled LDL (Fig. 4, Ref. 7) and 5.73 f 0.71 135 f 4.8 255 f 22.5 from nonspecific binding of '"'1-LDL by culture dishes without 3.34 f 0.41" 97.0 f 4.8" 90.4 f 23.2b cells (Figs. 1,3, and 7). The cell-associated "'I-labeled material appears to be apo-LDL since it is not extractable into chloroform and is insoluble in aqueous tetramethylurea (Table 392 f 27.2 2980 f 219 7540 f 537 110 f 5.3" 1310 f 50.1" 3530 +- 170" 111). Such specific low affinity receptor activity is consistent with previous findings of small amounts of specific binding of "'I-LDL (5, 6), small amounts of specific degradation of '251" p < 0.001. hp c 0.02. LDL ( 7 ) ,and regulation of sterol synthesisby LDL (10, 11) in receptor-negative FHC fibroblasts. TABLE IX TheLDLreceptors in FHC cells demonstrated by our Differential effect of compactin preincubation on LDL degradation experiments resemble the normal high affinity LDL receptor in GM 2000 and normal cells in several ways. 1) The binding of "'I-LDL is accompanied by Cells were plated in triplicateat 105/well inMEM + 15-20%,NBCS. appropriate degradation (Fig. 2). During a 5-h incubation at After 2-5 days of growth, the medium was removed, the cells were 37 "C with 100 pg/ml of '"I-LDL, the ratio of dextran-S04washed twice, and 1 ml of MEM + 10% LPDS f 2-4 ~ L Mcornpactin LDLinternalized LDLdegraded was added. After 2 days of preincubation, the medium was removed, releasable surface-bound the cells were washed twice, and to each well was added 0.7 ml of LDL was 10.6:100:152 in GM 2000 cells and 10.0:100:182 in MEM + 10% LPDS containing100 pg/ml of "'I-LDL with or without normal cells. The degradation of "'I-LDL was linear after a a 20-fold excess of unlabeled LDL. Cells were incubated for 5 h at 45-min delay in both normal and receptor-negativecells (Fig. 37 "C and then processed as described. 3B), consistent with a requirement that LDL be transferred PhenoCells -Compactin +Compactin Percentage to lysosomes before degradation. Chloroquine,alysosomal type increase enzyme inhibitor, reduced l2'1-LDL degradationto 5% of ng "'I-LDL degraded/mg/5 h normal in GM 2000 cells. Thus, the usual pathway of LDL FHC GM 2000 133 f 23.5 368 f 37.5 178 uptake is apparently presentin FHC cells. 2) The timecourse 225 f 25.8 675 f 25.3 200 446 f 195 122 201 f 48.3 for cell-associated "'I-LDL was similar in normal and FHC 634 f 181 179 228 f 75.1 cells (Fig. 3A). 3) Modification of lysine residues reduced the 830 f 78.1 60 490 f 23.9 capacity of LDLtointeractwithbothFHCandnormal 378 f 96.7 1120 & 47.7 196" receptors (Table11).4) Binding to both cell types was calcium61.9 f 27.5 170 f 12.9 175"." sensitive (Table V). 5) FHC cells demonstrated down-regulaMean increase = 160 f 1856 tion of LDL receptors in response to pretreatment with choNormal DWSr 4250 f 103 5850 f 152 35 lesterol and LDL (Table VIII) and increased receptor activity 9" /, 3830 f 225 4180 f 252 in response to compactin,a cholesterol biosynthesis inhibitor RO 111 18,900 f 1390 13" 21,300 f 874 (Table IX). 2540 f 138 3370 f 168 33"." 5y.t S 356 2270 f 107 3520 f 123 TheLDLreceptors of FHC cells are also easily distinMean increase = 29 k 8%: guished from those of normal cells. The maximum specific capacity for cell-associated'"I-LDL in FHC cells was 9.8-13% '' 5 days initial growth. of normal cells assayed in the same manner (Table VI). But 2 ~ L compactin M instead of 4 p ~ . ' p < 0.001 uersus GM 2000. receptor affinity of LDL was also reducedto 14-2870 of normal. Experiment 1 GM 2000 No addition +50 pg/ml LDL Normal No addition +50 pg/ml LDL Experiment 2 GM 2000 addition No +7.5 pg/mlcholesterol Normal No addition +7.5 pg/mlcholesterol ~

*

*

12864


The K , for LDL in FHC cells (49-127 pg/ml) was 3.5-7.3 times that of normal cells and much nearer the estimated normal extracellular LDL apoprotein B concentration of 70 pg/ml. Thus, FHCcells should metabolize significantamounts of LDL when grown chronically under conditions in which large amounts of LDL are present in the medium. This was confirmed in Table VI1 in which FHC cells metabolized 34% of the normal amount of LDL when 80 pg/ml of LDL was present. Reduced affinity ofl2'1-LDL in GM 2000 cells was also noted in assays conducted at 4 "C (Fig. 7), a condition preventing LDL internalization anddegradation. The data presented do not allow for a distinction between altered expression of the classic LDL receptor and a genetically separate class of low affinity LDL receptors that might be present in all cell types. It should be noted, however, that in another FHC mutation characterized by increased LDL receptor affinity but low capacity there is no evidence for a class of low affinity receptors even though they should have been moreeasily detected thanin normal cells (17). Of interest is the recent report by Beisiegel et al. (36) that some LDL receptor-negative fibroblasts have much morebinding of radiolabeled anti-LDL receptor antibody than of radiolabeled LDL, suggestingthat thehigh affinity LDL receptor mightbe present but modified. The regulation of receptors inGM 2000 cells wasalso altered. Preincubationof normal cells for 48 h with compactin in LPDS medium resulted in increased specific I2'II-LDLdegradation compared to cells grown in LPDS without compactin (Table IX). This confirms similar previous work with compactin in normal skin fibroblasts (29). Thus, LPDS medium apparently is not a maximum stimulus to LDL receptor accumulation. GM 2000 cells also demonstrated increased receptor activity after compactin preincubation. The mean percentage increase in specific LDL degradation in this cell type after compactin preincubation was significantly greater (160 f 18%)than in normal cells from three individuals (29 f 8%). Since the absolute increase in specific receptor activity after compactin addition was considerably greater in normal cells than in FHC cells, the greater percentage increasein specific LDL degradation in GM 2000 cells may be physiologically insignificant and merely represent an artifactof experimental conditions or data treatment. An alternative explanation is that the greater percentage increase in degradation in GM 2000 cells after compactin preincubation represents the unmasking of specific LDL receptors usually down-regulated by endogenouscellular production of cholesterol.Prolonged preincubation with compactin appears important in order to demonstrate increased receptors since Haba et al. (31) didnot observe any changesin LDL receptor activity in either normal or FHC fibroblasts incubated with both I2'II-LDLand compactin for only 6 h. Thus, our finding of specific low affinity LDL receptors andhyper-responsiveness to compactinin GM 2000 cells raises the possibility that the primary abnormality in FHC fibroblasts may not necessarily be a defect in the structural gene for the LDL receptor but rather a defect in the genes regulating LDL receptor expression or cholesterol metabolism.Amechanismfor FHC implicatingincreased endogenous production of cholesterol has been proposed previously by Fogelman et al. (22). Low affinity receptors may have been overlooked in previous experiments for several reasons. The curve describing "'I-LDL binding of fibroblasts is sigmoidal and very little binding is seen at '"I-LDL concentrations far below the K,. Thus, when '"I-LDL at concentrations of 5-20 pg/ml is used, normal cells have appreciable receptor occupancy whereas receptor-negative cells show very little occupancy. "'I-LDL concentrations above 50 pg/ml should be employed for FHC

cells and, for this reason, 1-2 mg/ml of unlabeled LDL in alternate wells is required to detect specific binding of the labeled material. We have often used large quantities of LDL prepared from the pheresisplasma of an FHChomozygote for this purpose. We also induced LDL receptorsby growing cells in 7mg/ml of LPDS for 72 h, a larger amount andlonger time than customarily used. The major classification scheme for FHC fibroblasts has utilized principally an indirect method for determination of LDL receptorfunction, eg., radiolabeled oleate incorporation into cholesteryl esters after exposure of cells to LDL (2, 4). Receptor-negative cells revealed no oleate incorporation in response to LDL whereas receptor-defective cells had 5-20% of normal incorporation. These are measurements of the acute effect of metabolized LDL on cells, not the actual LDLprocessed by surface binding, internalization,anddegradation. Since low levels of specific LDL receptor activity can be identified in at least some receptor-negative fibroblasts, we believe thatthephenotype of homozygous FHCmay be associated with a spectrum of qualitative and quantitative receptor abnormalitiesencompassing the functional categories receptor-negative and receptor-defective. Since it is agreed thatalmost allhomozygous FHCfibroblastsare severely deficient inLDL receptor function, future studies should more clearly define the exact natureof the LDL binding abnormalities present in these cells. Acknowledgments-We thank Stephen Block, Mark Mendelsohn, and Kay Zorn for excellent technical and secretarial assistance. REFERENCES 1. Goldstein, J. L., and Brown, M. S. (1977) Annu. Rev. Biochem. 46,897-930 2. Fredrickson, D. S., Goldstein, J. L., and Brown, M. S. (1978) in The Metabolic Basis of Inherited Disease (Stanbury, .J. B., Wyngaarden, J. B., and Fredrickson, D. S., eds) pp. 617-633, McGraw-Hill, New York 3. Goldstein, J. L., and Brown, M. S.(1979) Annu. Reu. Genet. 13, 259-289 4. Goldstein, J. L., Dana, S. E., Brunschede, G. Y., and Brown, M. S.(1975) Proc. Natl. Acad. Sci. U. S. A . 72, 1092-1096 5. Brown, M. S., and Goldstein, J. L. (1974) Proc. Natl. Acad. Sci. U. S. A . 71, 788-792 6. Breslow, J. L., Spaulding, D. R., Lux, S. E., Levy, H . I., and Lees, R. S. (1975) N. Engl. J. Med. 293, 900-903 7. Goldstein, J . L., andBrown,M. S.(1974) J. Biol. Chem. 249, 5153-5162 8. Carpenter, G., and Cohen, S. (1976) J . Cell Biol. 71, 159-171 9. Tollefsen, D. M., Feagler, J. R., and Majerus, P.W. (1974)J . Biol. Chem. 249,2646-2651 10. Fung, C. H., Khachadurian, A. K., Wang, C. H., and Durr, I. F. (1977) Biochim. Biophys. Acta 487,445-457 11. Fung, C. H.,Wang, C. H., andKhachadurian, A. K. (1978) Biochim. Biophys. Acta 528, 445-455 12. Goldstein, J. L., and Brown, M. S. (1976) Curr. Top. Cell Regul. 11. 147-181 13. Deleted in proof 14. Brown, M. S., Faust, J. It., Goldstein, J. L., Kaneko, I., and Endo, A. (1978) J. Biol. Chem. 253, 1121-1128 15. Ostlund, R. E., Jr., Pfleger, B., and Schonfeld, G. (1979) J. Glia. Invest. 63, 75-84 16. Puck, T . T., Cieciura, S. J., and Robinson,A. (1958) J. Exp. Med. 108,945-956 17. Ostlund, R. E., Jr., Levy, H . A,, Witztum, J. L., and Schonfeld, G. (1981) J. Clin. Invest., in press 18. Scatchard, G. (1949) Ann. N. Y. Acad. Sci. 51, 660-672 19. Deleted in proof 20. Deleted in proof 21. Deleted in proof 22. Fogelman, A. M., Edmond,J., Seager, J., and Popjak, G. (1975) J. Biol. Chem. 250, 2045-2055 23. Ostlund, R. E., Jr., Hajek, S. V., Levy, R. A., and Witztum, J . L.

LDL Receptors in Homozygous Hypercholesterolemia (1981) Metabolism 30, 285-289 24. Deleted in proof 25. Kane, J. P. (1973) Anal. Biochem. 53, 350-364 26. Folch, J., Lees,M., andSloane-Stanley, G. H. (1957) J . Biol. Chem. 226,497-509 27. Weisgraber, K. H., Innerarity, T. L., and Mahley, R. W. (1978) J. Biol. Chem. 253,9053-9062 28. Deleted in proof 29. Filipovic, I., and Menzel, B. (1981) Biochem. J. 196, 625-628 30. Oram, J. F., Albers, J. J., and Bierrnan,E. L. (1980) J . Biol. Chem. 255,475-485

12865

31. Haba, T., Mabuchi, H., Yoshimura, A,, Watanabe, A,, Wakasugi, T., Tatami, R., Ueda, K., Ueda, R., Karnetani, T., Koizumi, J., Miyamoto, S., Takeda, R., and Takeshita, H. (1981) J . Clin. Invest. 67, 1532-1540 32. Fieser, L. F. (1953) J. Am. Chem. Soc. 75, 5421-5422 33. Dana, S. E., Brown, M. S., and Goldstein, J . L. (1977) Biochem. Biophys. Res. Commun. 74, 1369-1376 34. Hofstee, B. H. J. (1959) Nature (Lond.)184, 1296-1298 35. Kruth, H. S., and Vaughn, M. (1980) J . Lipid Res. 21, 123-130 36. Beisiegel, U., Schneider, W. J., Goldstein, J . L., Anderson, R. G. W., and Brown, M. S. (1981) J. B i d . Chem. 256, 11923-11931