Cholesterol synthesis contributes significantly to

Transcription

1 247 Human Cholesterol Synthesis Measurement Using Deuterated Water Theoretical and Procedural Considerations P.J.H. Jones, C.A. Leitch, Z.-C. Li, and W.E. Connor Human cholesterogenesis is measurable as the rate of Incorporation of deuterium derived from deuterium oxide (D 2 O) within the body water pool into plasma or erythrocyte cholesterol pools. Oral D 2 O equilibrates across body water, thus enabling extracellular sampling of pools (such as urine) to serve as accurate indicators of intracellular deuterium enrichments at the point of synthesis. Required doses of D 2 O fall below the threshold associated with negative side effects. Deuterium/carbon incorporation ratios into cholesterol during biosynthesis have been established that are applicable in humans. Models using unconstrained and constrained curve fitting permit improved flexibility in interpretation of deuteriumuptake kinetics. However, sample-size restrictions presently limit the ability of the technique to examine the kinetics within individual lipoprotein species. Correction of enrichment data for proton exchange during the combustion and reduction phases of sample preparation is an additional important procedural concern. In summary, the deuterated-water procedure is a useful tool in studies of human cholesterol synthesis that offers the advantages of short measurement interval, relative noninvasiveness, and provision of a direct index of synthesis in comparison with other available techniques. (Arteriosclerosis and Thrombosis 1993;13: ) KEYWORDS cholesterol deuterium synthesis methodology erythrocytes cholesterol ester humans Cholesterol synthesis contributes significantly to total body input in humans 1 and thus exists as an important factor in determining body cholesterol balance. Cholesterol formation rates in humans are measurable using several techniques. These techniques include input-output analysis, 2 ' 3 kinetic approaches using labeled cholesterol, hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase activity measured directly in the liver or indirectly in skin fibroblasts 7 and mononuclear leukocytes, 8 and plasma sterol precursor-level quantification. 910 Steady-state requirements vary across methods. With input-output analysis or kinetic approaches, cholesterol pool steadystate conditions must exist for the measurements to be valid. With other methods, cholesterogenesis can be measured without steady-state cholesterol pools. Moreover, the time window of measurement varies from 7 days or more for input-output analysis to a few hours for sterol precursors. Although wide-ranging, such methods are limited in three ways: they either are accurate but require extended measurement periods, are immediate but overly invasive, or measure synthesis indirectly. From the Division of Human Nutrition (PJ.HJ., C.A.L., Z.-C.L), University of British Columbia, Vancouver, Canada, and the Division of Endocrinology, Metabolism and Clinical Nutrition (W.E.C.), The Oregon Health Sciences University, Portland. Supported by a grant from the British Columbia and Yukon Heart and Stroke Foundation. Address for reprints: Dr. Peter J.H. Jones, Division of Human Nutrition, 2205 East Mall, University of British Columbia, Vancouver, British Columbia, Canada VT 1Z4. Received March 11, 1992; revision accepted November 4, In contrast, the deuterium-uptake procedure uses an intermediate time window of measurement (-24 hours), is not overly invasive, and measures synthesis directly. Cholesterogenesis is measured as the increase in deuterium abundance in circulatory-pool cholesterol over time as a function of body water deuterium oxide (D 2 O) enrichment. The method offers various attractive features that lend themselves to use in in vivo human studies. The goal of the present review was to examine various theoretical issues related to the method and to present experimental results concerning related assumptions. History of Deuterium's Use in Cholesterogenesis Measurement The use of deuterium as a tracer to measure lipogenesis was pioneered by Rittenberg and Schoenheimer in n Mice maintained at constant body water D 2 O enrichment incorporated deuterium label into lipids to a plateau of approximately one half that of body water. Later, in humans, London and Schwartz 12 showed a parallel rise in plasma and erythrocyte cholesterol deuterium enrichment over several weeks in two subjects given regular doses of D 2 O. This study was the first reported use of a stable-isotope tracer in the study of human lipid metabolism; at that time, radiolabeled nuclides were tracers of choice in studying cholesterogenesis. 13 Taylor et aj 14 later refined the deuteriumuptake technique; as a result of improved precision of measurement using isotope ratio mass spectrometry, they were able to reduce the D 2 O dose required. Nevertheless, sensitivity limitations of this early instru-

2 248 Arteriosclerosis and Thrombosis Vol 13, No 2 February 1993 mentation necessitated doses of D 2 O large enough to result in vertigo in some subjects. More recently, improvements in isotope ratio mass spectrometry have enabled measurement of deuterium enrichment changes in cholesterol over comparatively short intervals using D 2 O dosages that, almost without exception, produce no side effects. 15 In addition, ongoing improvements in our understanding of human cholesterol turnover have enabled refinements in modeling and interpretation. Theoretical Considerations Water as a Tracer The deuterium-uptake method shares many of its theoretical underpinnings with a similar technique that uses tritiated water ( 3 H 2 O) developed for in vitro 1 and in vivo 17 application in animals. In both approaches, tracer water equilibrates between the precise intracellular site of synthesis and the extracellular physiological fluid, such as urine or plasma, that is used to determine the level of labeling. A primary advantage of either method is that the tracer that is taken up into de novo synthesized sterol originates from a pool of known enrichment. Potential errors arising from differential labeling of intracellular and extracellular precursor enrichments, such as occur with the use of labeled metabolites other than water, are thus obviated. For instance, incorporation rates of acetyl CoA units into identical rat tissue slices differ when the incubation contains 14 C- labeled glucose, acetate, or octanoate. Such differences reflect not synthetic rate changes, but variations in the amount of dilution of the labeled acetyl CoA pool derived from each metabolite. Certain of these problems have been overcome with the use of modern xenobiotic-probe approaches. 18 Thus, body water serves as an ideal precursor pool for whole-body studies of lipogenesis. Although water equilibrates across body pools, intercompartment transmission is not instantaneous. In animals, the interval required for plasma water to reach the plateau tracer concentration after intravenous injection of a bolus of labeled water is approximately 30 minutes. 17 ' 19 Previous data in human subjects have shown that equilibration of plasma water deuterium enrichment after a single oral dose of D 2 O occurs between 3- hours. 20 Data from a similar experiment using earlier time points are presented in Figure 1. Two fasted subjects drank a loading dose of D 2 O at time zero, after which saliva and urine samples were collected at hourly intervals for 4 hours. Values expressed as percent plateau enrichment show equilibration occurring by 1-2 hours after dosing. Changing enrichment during the first postlabeling hour would produce a measurement error in experiments of very short duration; however, for experiments of even a few hours, the relatively short equilibration interval would be expected to result in a minimal error. To avoid the problem of equilibration of the precursor, the baseline cholesterolenrichment sample could be obtained 1-2 hours after as opposed to before administration of the D 2 O dose. An additional advantage of the deuterium-uptake method is the absence of cholesterol proton exchange with aqueous proton pools. When commercial free cholesterol at natural abundance was incubated with Time (hr) Urine Saliva FIGURE 1. Line graph shows equilibration of urinary (m and solid line) and salivary (n and dashed line) water deuterium enrichment over time before and after a single dose ofdeuterated water (D 2 O) in two subjects. Plotted is the mean percent deuterium enrichment in each pool relative to the mean of 1-4-hour time points in subjects fasted overnight and then given 0.0 g/kg body wt D/D followed by a 50-ml rinse of tap water. Subjects consumed no food or water over the measurement period and provided hourly samples for 4 hours after dosing. 3 H 2 O 19 or with 99.8% D 2 O and shaken at 37 C overnight (P.J.H. Jones, unpublished data), incorporation into the isolated sterol was nil. Although it cannot be ruled out that deuterium exchange does not occur under any physiological situation, data strongly indicate that deuterium uptake must represent synthetic activity. Fractional Uptake Rate for Label To accurately measure net cholesterogenesis using D 2 O uptake, the ratio of incorporation of deuterium versus protium into cholesterol during biosynthesis must be established, as has been determined for fatty acids. 21 Not all of the various precursors along the cholesterol biosynthetic pathway incorporate the label at a level proportional to that of body water. Thus, knowledge of the fraction of the 4 protons incorporated per de novo synthesized cholesterol molecule that originates from labeled precursors is needed. This area has been thoroughly considered for the 3 H 2 O method. 17 In brief, hydrogens bonded to the growing cholesterol molecule are derived from three sources. Seven are obtained directly from body water, reflecting the exact tracer concentration therein. Fifteen originate from reduced nicotinamide adenine dinucleotide phosphate (NADPH) during the synthetic cascade, and the remaining 24 derive from cytosolic acetyl CoA. Previous studies comparing uptake rates from 3 H 2 O and 14 C-acetate have shown that protons derived directly from H 2 O and NADPH equilibrate with body water. 1 In contrast, those from acetyl CoA are unlabeled, unless the period of measurement is sufficiently prolonged to permit recycling of labeled acetate into acetyl CoA pools. Assuming negligible acetate recycling, the ratio of 0.81 labeled protons per carbon atom is accepted as most representative of synthesis. 17 This value corresponds to 22 deuterium atoms per 27 carbon atoms within the cholesterol molecule. Correcting for the ratio of hydrogen to carbon atoms in cholesterol, the fraction of deuterium incorporated at the

3 Jones et al D 2 O in Cholesterol Synthesis Measurement 249 level of whole-body enrichment per total hydrogen atoms (deuterium/total hydrogen) is Potential limitations exist in the estimation of deuterium/carbon when using either 3 H 2 O or D 2 O. While the theoretical assumptions concerning the enrichment of protons from water and acetate are reasonably well defined, there is less certainty regarding those derived from NADPH. Protons used in cholesterol biosynthesis from NADPH originate from two sources. The first is the pentose phosphate pathway, which does not exchange protons with cellular water and thus does not contribute to enrichment of sterol during synthesis. The second, the malic enzyme system, may result in exchange of NADPH with water protons at cellular concentrations. 17 A number of in vivo and in vitro studies using tritium suggest small but potentially significant variations in the deuterium/carbon ratio, suggesting significant contributions of pentose phosphate pathwayderived NADPH or some other change in NADPH equilibration A range from 21 to 25 3 H atoms per 27 carbon atoms within the cholesterol molecule has been reported, which translates into potential errors of about 10% in apparent measures of cholesterogenesis using labeled water. Although some studies have shown incorporation ratios to be consistent across changing metabolic conditions and acetyl CoA substrate supply, it has also been suggested that ethanol use as a substrate results in a reduction in the 3 H/carbon ratio. 2 Disparity between cholesterol synthesis rates determined by 3 H 2 O and sterol-balance methods previously reported in guinea pigs fed various dietary fats has been suggested as being caused by variability in the source of NADPH protons, secondary to differences in macronutrient oxidation rates. 27 Moreover, it cannot be ruled out that with D 2 O, deuterium/carbon ratios may differ across various tissues. Collectively, these findings suggest that the use of deuterium uptake to measure cholesterol synthesis may be unreliable under certain conditions that result in shifts in NADPH equilibration. Modeling Assumptions In addition to selecting the correct deuterium/carbon ratio for the system under study, choosing an appropriate model to accurately describe the pattern of change in cholesterol deuterium enrichment is central to successful data interpretation. The simplest, constrained model describing the deuterium-uptake rate assumes that monoexponential enrichment increases to an asymptotic value at the maximum theoretical plasma cholesterol enrichment, which is represented by the plasma water enrichment and can be corrected by using the appropriate deuterium/carbon enrichment ratio (Figure 2). For cholesterol, this plateau corresponds to the fractional uptake rate (0.478 x deuterium enrichment of body water). As a consequence of the low turnover rate relative to pool size, reaching this plateau requires considerable time. Indeed, in a single subject given continual deuterium doses, plasma total cholesterol deuterium enrichment continued to increase after 100 days. 12 The initial 1-2-day period after deuterium dosing is considered to yield the most interpretable synthesis data for several reasons. First, during the immediate postdose interval, movement of labeled cholesterol out of the rapid-turnover compartment is small; thus, uptake Time (days) FIGURE 2. Graph of model used to calculate cholesterol fractional synthesis rates by using constrained and unconstrained monoexponential approaches. Theoretical enrichment plateau is that of plasma water (line a). Enrichment curves with identical turnover rates plateauing at 75% (line b) and 50% (line c) of theoretical value are also shown. Note that when plateau values below the theoretical level are achieved, turnover rate represents that of the overall pool, not only that of cholesterol synthesis. rates are relatively linear. Considering the extended time required to reach the asymptote, this linear segment of the curve extends for many hours. In subjects consuming three meals per day, deuterium uptake measured every 4 hours was, with diurnal exception, relatively linear over 48 hours. 28 Second, during the shortterm postdose period, recycling of the label incorporated into cholesterol from other esters or free body pools was also small. Accounting for label reentry requires more complex kinetic analysis. Third, during earlier postdose periods there is minimal recycling of the label into acetate pools that can be expected to alter the deuterium/carbon incorporation ratio. In contrast, deuterium incorporation kinetics followed over longer periods follow an exponential curve, as has been demonstrated in pigs 29 and humans. 12 In pigs, exponential inflection in deuterium enrichment in erythrocyte cholesterol is detectable at 48 hours and pronounced at 9 hours. 29 In groups of pigs fed both cholesterol-free and 0.5% cholesterol-containing diets, monoexponential fits using the corrected water-enrichment level as the predicted plateau resulted in poor curve fits for deuterium enrichment. In each case the asymptote to the plateau value was well below that predicted by the simple model. Long-term enrichment data in humans from London and Schwarz 12 have also shown a tendency toward a plateau below the theoretical prediction when modeled with a monoexponential fit. There are several possible reasons for the departure of actual data from the predicted model over periods exceeding hours. First, the longer term deuterium-enrichment curve may not be definable by a single monoexponential function. Goodman et al 30 demonstrated the multiexponential nature of the kinetic behavior of elimination for 14 C-cholesterol injected into humans, reflecting interpool mixing. Second, incoming dietary and endogenously recycled cholesterol continuously dilutes the rapid-turnover pool. Dietary cholesterol represents a sizable fraction of overall body input.

4 250 Arteriosclerosis and Thrombosis Vol 13, No 2 February Time (hr) - Obxrved Enrichment - - Con*r*»d Modal Uncoratr^cwj Modd FIGURE 3. Line graph shows application of constrained and unconstrained models to deuterium enrichment data (authors' unpublished observations) obtained from six subjects over 48 hours. Subjects consumed prepared diets as three meals per day and were given 0.7 g-kg estimated body water' % Dfi at time zero (7:30 AM). Blood samples were collected before and at 4-hour intervals after the initial dose. Subjects consumed drinking water containing Dfi over the 48-hour blood-sampling period to maintain body-water enrichment at the plateau level. Plasma free-cholesterol enrichment was determined as described. 28 Monoexponential function fitting was performed with the SYSTAT curve-fitting program (SYSTAT, Evanston, III). Shown are mean observed enrichment values expressed as percent body-water enrichment and best-fit curves for each model Recycled cholesterol also contributes significantly to body input, as is suggested from the observation of negative deuterium-uptake rates in subjects during short-term fasting Negative uptake rates reflect entry of unlabeled cholesterol to a level that exceeds that of labeled, synthesized sterol. In a case in which dietary and recycled sterol enter the central pool at an identical rate as de novo synthesized cholesterol, the deuterium enrichment of plasma cholesterol would reach one half the predicted plateau value (Figure 2, line c). Last, as previously mentioned, the deuterium/ carbon incorporation ratio is altered as the label accrues within body metabolites over time. For these reasons interpretation of cholesterol deuterium-uptake data over longer periods is more complex than during the relatively linear immediate postdose period. For situations in which the theoretical plateau level is achieved, the deuterium incorporation rate is equal to the synthesis rate because there is no diluting source of unlabeled cholesterol. In cases in which true enrichment plateaus below theoretical predictions are achieved, the turnover rate for cholesterol, as determined from the rate of deuterium incorporation, reflects turnover of the total free-cholesterol pool, not only that of de novo labeled sterol. In such situations knowledge of the plateau value as a fraction of the theoretical maximum may be useful in determining the proportion of total pool turnover that can be attributed to de novo synthesis. The plateau value can be determined by extrapolating early time points or, preferably, by experiments of longer duration. The derived turnover rate can be corrected for the amount of unlabeled sterol entering the pool (Figure 3). Here, six healthy male subjects fed liquid diets as three meals per day consumed a loading dose of D 2 O followed by maintenance doses for 48 hours. Plasma free cholesterol deuterium enrichment was measured every 4 hours. Cholesterol deuterium-uptake curves were modeled using constrained curve fitting, in which plateau enrichment was assumed to be that of plasma water, and using unconstrained curve fitting, in which the curve was extrapolated to define its own plateau. With the constrained model, fractional synthetic rates (FSRs) were calculated as 0.075±0.005 pool day" 1. With the unconstrained model, the mean plateau value obtained through extrapolation was 7.8±3.% of the theoretical plateau. FSR values obtained directly from the curve fit averaged ±0.011 pool-day" 1. With the unconstrained model, the cholesterol synthesis rate, calculated as the product of the fraction of pool turnover attributable to synthesis and the turnover rate, was pool day" 1. Assuming a rapid-turnover pool size of approximately 25 g, one half of which is esterified and therefore turning over more slowly, net cholesterol synthesis rates were 0.94 and 1.00 g day" 1 for constrained and unconstrained approaches, respectively. The major disadvantage of the unconstrained model is its inaccuracy in predicting the plateau value from early time points. Optimally, derivation of the turnover rate from early time points and the plateau value from additional samples collected after deuterium dosing provides true synthetic FSR. Even over durations of <24 hours, significant movement of label occurs into the esterified cholesterol pool. 28 Such loss of label from the primary pool of measurement will result in some underestimation of synthetic rate with either the constrained or unconstrained approaches. With appropriate modeling,fluxof the label into the ester pool should permit calculation of the esterification rate. Procedural Considerations Sample extraction, purification, and combustion/reduction procedures associated with the deuterium-uptake methodology have been previously described Once extracted from tissue, cholesterol may be purified by thin-layer chromatography or high-pressure liquid chromatography. 29 Combustion and reduction steps are typically performed manually, off-line from the mass spectrometer. Regardless of the methodology used, the analysis is multistage, thus presenting numerous opportunities for sample proton contamination. Chief contamination sources are entry into the preparation system of unlabeled protons from solvent impurities during extraction and from water vapor during distillation. To minimize contamination artifacts, samples are analyzed in replicate. Given a typical intrareplicate coefficient of variation of 2-4%o, the minimal detectable difference in FSR is pool day" 1, assuming a plasma water deuterium enrichment of 3,500%o. Sampling Size and Site Considerations Both sample size and site limitations exist as constraints to studies of the deuterium-uptake method in humans, in contrast with animal experimentation. For humans, relatively large initial plasma volumes are required for cholesterol deuterium-enrichment analyses. In nonnolipidemic subjects, 2 ml plasma yields

5 Jones et al D 2 O in Cholesterol Synthesis Measurement Time (hr) - Ptoonw - Ejythrocyta FIGURE 4. Plasma free (m and solid line) and erythrocyte (a and dashed line) cholesterol deuterium enrichment (%o) over baseline (hour zero) and at, 12, and 24 hours after loading dose (0.7g/kg estimated body water) and maintenance doses of DjO. Isotopic enrichments over baseline are expressed as per mil (%c) versus standard mean ocean water (SMOW), where per mil is defined as [(R^vi t /R au^mi)-l]xl,000, where R is the ratio of heavy-to-light isotope and R^doi is SMOW. Data are mean±sd of six subjects. approximately 1 mg free cholesterol, the amount required per replicate to produce the 1 )i\ combustion water necessary for isotope ratio mass spectrometric analysis. The blood volume requirement can be reduced if free and esterified cholesterol are combined and considered as a unified pool. Alternatively, erythrocyte cholesterol, found solely in free form, enables use of smaller blood samples. Sample-size limitations have hampered studies in infants and of plasma lipoprotein subfractions. At present, plasma free, 14 ' esterified, 28 and erythrocyte 29 cholesterol deuterium uptakes have been investigated. Circulatory cholesterol is selected for sampling since it is both readily accessible and is a part of the rapidturnover pool that also comprises hepatic and intestinal sterol. 33 The free-cholesterol component of this pool equilibrates between plasma and other subpools with a turnover rate between the liver and plasma of 0.34 pool-hr" 1 in the rat. 34 Erythrocytes do not synthesize cholesterol' 2 ; thus, sterol in this subcompartment is obtained from the plasma. Erythrocyte cholesterol exists in the free form; however, the rate of equilibration between the plasma and erythrocytes is unknown. To examine this question six subjects consumed fixed diets for 4 weeks. During the final week subjects drank priming (0.7 g D 2 O/kg estimated body water) and maintenance doses of D 2 O over 24 hours. Plasma was obtained before the priming dose and at, 12, and 24 hours thereafter. Plasma free and erythrocyte cholesterol deuterium enrichments were determined at these time points. Results (Figure 4) indicate that deuterium incorporation was initially slower into erythrocytes compared with plasma due to interpool mixing, probably resulting from a delay in equilibration between the two compartments. These findings indicate that over the initial -12-hour postdose interval, use of erythrocyte deuterium uptake results in an underestimation of sterol synthesis into the rapidly exchangeable pool. Use of erythrocyte deuterium enrichments to assess sterol 24 synthesis should thus be made with caution over the initial postdose period. Within the plasma compartment, although deuterium uptake occurs into both free and ester pools, incorporation into the total free compartment exceeds that into ester by twofold. 28 The more rapid uptake into free sterol indicates its role as the primary compartment to receive newly synthesized sterol. Of the available options, free plasma cholesterol is thus likely to be most sensitive to perturbations in cholesterogenesis. For instance, diurnal rhythmicity observed in the free-cholesterol compartment 28 has not been reported in either the erythrocyte or ester compartment. Selection of Combustion/Reduction Tube Material Combustion and reduction procedures for converting sterol to hydrogen gas for mass spectrometric analysis have used quartz 28 ' 29 and Pyrex glass tubes Pyrex tubing is less costly and more readily available, although combustion temperatures for Pyrex are of necessity lower than those with quartz. Regardless of the tubing material selected, during both combustion and reduction stages cholesterol-derived protons exchange with the inner surface of the tube, resulting in isotopic dilution. The amount of exchange varies, depending on variations in the amount and isotopic composition of exchangeable hydrogen present on the inner wall of the tube. Representative effects of deuterium exchange during the combustion process are shown in Figure 5. Replicate 1-mg lipid samples were prepared with copper oxide and silver wire in preannealed 20-cm x 5-mm quartz tubes and combusted at 50 C for 4 hours. Tubes were of fixed length to equalize the interior surface area. The exchange associated with the use of quartz tubing resulted in a 12% offset in enrichment relative to theoretical predictions (Figure 5A). In a second experiment, replicate cholesterol samples of different deuterium enrichments were prepared in fixed lengths of both preannealed quartz and glass tubing and combusted at 50 C and 510 C, respectively, for 4 hours. An 18% offset in enrichment of samples combusted in glass relative to theoretical predictions (Figure 5B) was observed. These offsets appeared to be linear over the range of working standards in the range of the sample enrichments. Reduction procedures provide an additional opportunity for exchange of protons and attendant dilution of enrichment. To examine the extent of exchange, 2-fi\ samples of four representative water standards differing in deuterium isotopic enrichment were prepared in 12-cm x 5-mm glass or quartz preannealed tubes cut to standardized lengths. Tubes were reduced at 510 C for 30 minutes. Results (Table 1) showed a reduction tube effect on the water-standard enrichments. Regression analysis demonstrated linearity across the four standards with both glass and quartz tubes. However, data for each tube type fit different lines, necessitating the use of separate regression equations to produce calibration curves for the two tube types. Once calibrated, water samples prepared in either tube type yielded comparable enrichments. In sets of water samples compared by using each tube type with standards analyzed in the tubing of the appropriate material, the average difference in enrichment was <3%.

6 252 Arteriosclerosis and Thrombosis Vol 13, No 2 February O Actual Enrichment (o/oo) -based Enrichment (o/oo) FIGURE 5. Line graphs show effect of combustion-tube composition on observed net enrichment measurements. Panel A shows theoretical (u and solid line) versus observed (a and dashed line) deuterium enrichments for a [U-D]palmitate dilution series when combusted in quartz tubing. Panel B shows cholesterol samples of differing deuterium enrichments when combusted in quartz (m and solid line) and Pyrex (a and dashed line) tubes. Isotopic enrichments are expressed as net enrichments per mil (%o) versus standard mean ocean water. At both combustion and reduction stages of deuterium-enrichment analysis, proton exchange between the sample and the tubing surface may result in underestimation of cholesterol deuterium enrichment. To address this problem, such dilution must be corrected for at each stage by using standards of known enrichment. Conclusion A need exists for methods that are capable of direct, short-term measurement of human de novo cholesterol synthesis. D 2 O offers certain advantages over other available methods. Deuterium/carbon incorporation factors have been established; however, whether and to what degree such factors are alterable by changes in metabolic conditions within the organism remain to be fully defined. The theoretical cholesterol deuteriumenrichment plateau is considered to be about one half that of the body-water enrichment. Available data show that a time period of several months is required for cholesterol enrichment to reach its plateau value. Because of the increased complexity of modeling assump- TABLE 1. Comparison of and Reduction-Tube Material on Observed Versus Theoretical Deuterium Enrichments of FOOT Water Standards Sample GISP CTWS SMOW SDHS Tubing type Mean enrichment (%»)* ± ± ± ± ± ± ± ±10.8 n Theoretical enrichment (%») Data are mean±sd. GISP, Greenland Island standard precipitate; CTWS, Chicago tap water standard; SMOW, standard mean ocean water; SDHS, stock dilution heavy standard. 'Isotopic enrichments are expressed as per mil (%o) versus SMOW, where per mil is defined as [(R«? pt/r mn d^)-l]x 1,000, where R is the ratio of heavy-to-light isotope and R,,,,^ is SMOW. tions over longer periods, the initial -24-hour interval of deuterium uptake is considered to be a relatively linear portion of the overall incorporation curve, representing the optimal period for measuring synthesis and turnover. Using periods shorter than hours will result in artifacts, because of both the time lag required for equilibration of D 2 O across the body-water compartment and the mixing of deuterated cholesterol within erythrocytes and plasma cholesterol pools. Extending the period beyond 24 hours leads to more complex curve fitting as labeled sterol mixes among various pools, as well as allowing the precursor incorporation ratio to increase as acetate begins to accrue the label. However, over periods of longer duration, cross-compartmental mixing and entry into pools of exogenous and recycled cholesterol result in enrichment curves that tend toward a lower than predicted plateau value. Thus, cholesterol turnover rates determined by using deuterium uptake require correction for flux of unlabeled sterol through the cholesterol pool. Such correction can be obtained from the value of the final plateau relative to the predicted value. More advanced models will be needed to better delineate the kinetics of cholesterol synthesis using D 2 O in vivo. In summary, although in certain circumstances not all related assumptions have been proven, D 2 O uptake is a potentially useful tool in studies of human cholesterol synthesis, provided that theoretical and practical issues are considered. Acknowledgments The excellent technical assistance of Gayle Wickens and Qement Fung is acknowledged. The superior assistance of Lauren Hatcher and the staff of the Clinical Research Center, Oregon Health Sciences University, is also appreciated. References 1. Dietschy JM: Regulation of cholesterol metabolism in man and other species. Klin Wochenschr 1984;2: Samuel P, lieberman S, Ahrens EH: Comparison of cholesterol turnover by sterol balance and input-output analysis, and a shortened way to estimate total exchangeable mass of cholesterol by the combination of the two methods. J Upid Res 1978;19: Grundy SM, Ahrens EH, Davignon J: The interaction of cholesterol absorption and cholesterol synthesis in man. J Upid Ra 199;10:3O4-315

7 Jones et al D 2 O in Cholesterol Synthesis Measurement Goodman DS, Noble RP: Turnover of plasma cholesterol in man. J Clin Invest 198;47: Ferezou J, Rautureau J, Coste T, Gouffier E, Chevalier F: Cholesterol turnover in human plasma lipoproteins: Studies with stable and radioactive isotopes. Am J Clin Nutr 1983;3: Bjorkhem I, Miettinen T, Reihner E, Ewerth S, Angelin B, Einarsson K: Correlation between serum levels of some cholesterol precursors and activity of HMG CoA reductase in human liver. J Lipid Res 1987;28: Chait A, Bierman EL, Albers JJ: Low-density lipoprotein receptor activity in cultured human skin fibroblasts. J Clin Invest 1979;4: Nguyen L, Salen G, Shefer S, Tint GS, Shore V, Ness GC: Decreased cholesterol biosynthesis in sitosterolemia with xanthomatosis: Diminished mononuclear leukocyte 3-hydroxy-3- methylglutaryl coenzyme A reductase activity and enzyme protein associated with increased low density lipoprotein receptor function. Metabolism 1990^9: Popjack G, Boehm G, Parker TS, Edmond J, Ewards PA, Fogelman AM: Determination of mevalonate in blood plasma in man and rat: Mevalonate "tolerance" tests in man. J Lipid Res 1979;20: Parker TS, McNamara DJ, Brown CD, Kolb R, Ahren EH, Alberts AW, Tobert J, Chen J, Schepper PJD: Plasma mevalonate as a measure of cholesterol synthesis in man. / Clin Invest 1982;74: Rittenberg D, Schoenheimer R: Deuterium as an indicator in the study of intermediary metabolism: XI. Further studies on the biological uptake of deuterium into organic substances with special reference to fat and cholesterol formation. / Biol Chan 1937;121: London IM, Schwartz H: Erythrocyte cholesterol: The metabolic behavior of the cholesterol of human erythrocytes. / Clin Invest 1953^2: Gould RG, LeRoy GV, Okita GT, Kabara JJ, Keegan P, Bergenstal DM: The use of C l4 -labeled acetate to study cholesterol metabolism in man. J Lab Clin Med 1955;4: Taylor CB, Mikkelson B, Anderson JA, Forman DT: Human serum cholesterol synthesis measured with the deuterium label. Arch Pathol 19;81: Jones PJH, Leatherdale ST: Stable isotopes in clinical research: Safety reaffirmed. Clin Sa 1991;80: Lakshmanan MR, Veech RL: Measurement of the rate of rat liver sterol synthesis in vivo using tritiated water. J Biol Chem 1977;252: Dietschy JM, Spady DK: Measurement of rates of cholesterol synthesis using tritiated water. J Lipid Res 1984;25: Hellerstein MK, Christiansen M, Kaempfer S, Kletke C, Wu C, Reid JS, Mulligan K, Hellerstein NS, Shackleton CHL: Measurement of de novo lipogenesis in humans using stable isotopes. J Clin Invest 1991;87: Jeske DJ, Dietschy JM: Regulation of rates of cholesterol synthesis in vivo in the liver and carcass of the rat measured using ['HJwater. J Lipid Res 198O;21: Schoeller DA, Dietz W, van Santen E, Klein PD: Validation of saliva sampling for total body water determination by H 2 I8 O dilution. Am J Clin Nutr 1982;35: Jungas RL: Fatty acid synthesis in adipose tissue incubated in tritiated water. Biochemistry 198;7: Andersen JM, Dietschy JM: Absolute rates of cholesterol synthesis in extrahepatic tissues measured with 3 H-labeled water and 14 C- labeled Substrates. J Lipid Res 1979;20: Dietschy JM, McGarry JD: Limitations of acetate as a substrate for measuring cholesterol synthesis in liver. J Biol Chem 1974;249: Loud AV, Bucher NLR: The turnover of squalene in relation to the biosynthesis of cholesterol. J Biol Chem 1957;233: Pullinger CR, Gibbons GF: The relationship between the rate of hepatic sterol synthesis and the incorporation of [ 3 H]water. J Lipid Res 1983;24: Selmer J, Grunnet N: Ethanol metabolism and lipid synthesis by isolated liver cells from fed rats. Biochim Biophys Ada 197;428: Fernandez ML, Yount NY, McNamara DJ: Whole body and hepatic cholesterol synthesis rates in the guinea pig: Effect of dietary fat quality. Biochim Biophys Ada 1990; 1044: Jones PJH, Schoeller DA: Evidence for diurnal periodicity in human cholesterol synthesis. J Lipid Res 1990;31: Wong WW, Hachey DL, Feste A, Leggitt J, Clarke LL, Pond WG, Klein PD: Measurement of in vivo cholesterol synthesis from ^O: A rapid procedure for the isolation, combustion, and isotopic assay of erythrocyte cholesterol. / Lipid Res 1991^32: Goodman DS, Smith FR, Seplowitz AH, Ramakrishnan R, Dell RB: Prediction of the parameters of whole body cholesterol metabolism in humans. J Lipid Res 1980;21: Roe RP, Jones PJH, Frohlich JJ, Schoeller DA: Association between apolipoprotein E phenotype and endogenous cholesterol synthesis as measured by deuterium uptake. Cardiovasc Ra 1991; 25: Jones PJH, Dendy SM, Frohlich JJ, Leitch CA, Schoeller DA: Cholesterol and triglyceride fatty acid synthesis in apolipoprotein E2-associated hyperlipidemia. Arteriosclerosis 1992;12: Norum KR, Berg T, Helgerud P, Drevon CA: Transport of cholesterol. Physiol Rev 1983;3: Magot T, Frein Y, Champarnaud G, Cheruy A, Lutton C: Origin and fate of rat plasma cholesterol in vivo: Modelling of cholesterol movements between plasma and organs. Biochim Biophys Ada 1987;921: