Regulation of plasma lecithin:cholesterol acyltransferase in man. III. Role of high density lipoprotein cholesteryl esters in the activating effect of a high-fat test meal

Plasma 1ecithin:cholesterol acyltransferase (LCAT) activity is increased during the clearance phase of alimentary lipemia induced by a high-fat test meal in normal subjects. Ultracentrifugal fractionation of high density lipoproteins (HDL) into HDL2, HDL,, and very high density (VHD) subfractions followed by analyses of lipid and protein components has been accomplished at intervals during alimentary lipemia to seek associations with enzyme changes. HDLz lipids and protein increased substantially, characterized primarily by enrichment with lecithin. HDL,, which contain the main LCAT substrates, revealed increased triglycerides and generally reduced cholesteryl esters which were reciprocally correlated, demonstrating a phenomenon previously observed in vitro by others. Both changes correlated with LCAT activation, but partial correlation analysis indicated that ester content is primarily related to triglycerides rather than LCAT activity. The VHD cholesteryl esters and lysolecithin were also reduced. Plasma incubation experiments with inactivated LCAT showed that alimentary lipemic very low density lipoproteins (VLDL) could reduce levels of cholesteryl esters in HDL by a nonenzymatic mechanism. In vitro substitution of lipemic VLDL for postabsorptive VLDL resulted in enhanced reduction of cholesteryl esters in HDL, and VDH, but not in HDL2, during incubation. Nevertheless, augmentation of LCAT activity did not result, indicating that cholesteryl ester removal from substrate lipoproteinsis an unlikely explanation for activation. Since VHD and HDL3, which contain the most active LCAT substrates, were also most clearly involved in transfers of esters to VLDL and low density lipoproteins, the suggestion that LCAT product lipoproteins are preferentially involved in nonenzymatic transfer and exchange is made. The main determinant of ester transfer, however, appears to be the level of VLDL, both in vitro and in vivo. Rose, H. G., and J. Juliano. Regulation of plasma lecithin: cholesteryl acyltransferase in man. 111. Role of high density lipoprotein cholesteryl esters in the activating effect of a high-fat test mea1.J. Lipid Res. 1979. 20: 399-407. Supplementary key words trig1yceride:cholesteryl ester exchange . lysolecithin . primary hypertriglyceridemia * cholesterol ester equilibrium * HDL3 . HDL, . very high density lipoprotein (VHDL) Plasma 1ecithin:cholesterol acyltransferase (EC 2.3.1.43) activity is increased in normal subjects 510 hr after ingestion of a high-fat test meal (1-3), temporally coinciding with the clearance phase of alimentary lipemia. Activation of LCAT has been related to maximal increases in plasma triglycerides ( 1 , 2) and does not occur if fat is removed from the test meal (1). In addition to elevation of plasma chylomicrons and VLDL, fat feeding produces alterations in HDL. Have1 (4) reported elevations of HDL phospholipids which tended to persist even after 12 hr, with less consistent increases in cholesterol and protein (4). In more recent studies in which HDLz and HDL3 were isolated (5), increases in HDLz total cholesterol, phospholipids, and protein were noted. HDLB total cholesterol fell, while there was an increase in phospholipids which persisted even 10 hr after feeding. Similarly, Nichols and coworkers (6) reported elevation of S~l .z0 ,4-8 and Srcl,zo, 0-4 fractions 5 hr after oil ingestion. HDL glycerides and phospholipids were elevated and the fatty acid composition of the glycerides reflected the composition of the fed oil. HDL3 and very high density lipoproteins (VHDL, 1.2 1 < d < 1.25 g/ml) are LCAT substrates (7-9) while HDLz is nearly inactive (7) and perhaps inhibitory (3). According to Wallentin and Vikrot (2) the elevation of HDL phospholipids after fat ingestion correlates with LCAT response. Taken together, these observations point to the possibility that fat-induced alterations in levels or composition of Abbreviations: HDL,, high density lipoproteins3; HDL,, high density lipoproteinsl; VHD, very high density plasma fraction (d > 1.220 g/ml); LDL low density lipoproteins; LCAT, 1ecithin:cholesterol acyltransferase; VLDL, very low density lipoproteins; TG, triglycerides; UC, unesterified cholesterol; EC, ester cholesterol; PC, phosphatidylcholine; LyPC, lysophosphatidylcholine. Journal of Lipid Research Volume 20, 1979 399 This is an Open Access article under the CC BY license. LCAT substrates and/or products within the HDL are related to the activation phenomenon. We have accordingly isolated HDL,, HDL,, and a very high density fraction (d > 1.22 g/ml) at intervals after fat feeding in order to relate changes to activation of the enzyme. More specifically, we sought answers to the following questions. I ) Would activation prove to be correlated with new lipoprotein substrates appearing in HDL, and/or HDL, during lipolysis? 2) Would the lipid composition of HDLz and HDL, be changed during alimentary lipemia in a direction suggesting better substrate performance, i.e., enrichment in substrate lipids [unesterified cholesterol (UC) and phosphatidylcholine (PC)] and reduction in product lipids [ester cholesterol (EC) and lysophosphatidylcholine (LyPC)]? 3) Could those HDL lipid changes that correlate with LCAT augmentation in vivo be duplicated in vitro with resulting activation of the enzyme? MATERIALS AND METHODS

Plasma 1ecithin:cholesterol acyltransferase (EC 2.3.1.43) activity is increased in normal subjects 5-10 hr after ingestion of a high-fat test meal (1-3), temporally coinciding with the clearance phase of alimentary lipemia. Activation of LCAT has been related to maximal increases in plasma triglycerides ( 1 , 2) and does not occur if fat is removed from the test meal (1). In addition to elevation of plasma chylomicrons and VLDL, fat feeding produces alterations in HDL. Have1 (4) reported elevations of HDL phospholipids which tended to persist even after 12 hr, with less consistent increases in cholesterol and protein (4). In more recent studies in which HDLz and HDL3 were isolated (5), increases in HDLz total cholesterol, phospholipids, and protein were noted. HDLB total cholesterol fell, while there was an increase in phospholipids which persisted even 10 hr after feeding. Similarly, Nichols and coworkers (6) reported elevation of S~l.z0,4-8 and Srcl,zo, 0-4 fractions 5 hr after oil ingestion. HDL glycerides and phospholipids were elevated and the fatty acid composition of the glycerides reflected the composition of the fed oil. HDL3 and very high density lipoproteins (VHDL, 1.2 1 < d < 1.25 g/ml) are LCAT substrates (7)(8)(9) while HDLz is nearly inactive (7) and perhaps inhibitory (3). According to Wallentin and Vikrot (2) the elevation of HDL phospholipids after fat ingestion correlates with LCAT response. Taken together, these observations point to the possibility that fat-induced alterations in levels or composition of LCAT substrates and/or products within the HDL are related to the activation phenomenon. We have accordingly isolated HDL,, HDL,, and a very high density fraction (d > 1.22 g/ml) at intervals after fat feeding in order to relate changes to activation of the enzyme. More specifically, we sought answers to the following questions. I ) Would activation prove to be correlated with new lipoprotein substrates appearing in HDL, and/or HDL, during lipolysis? 2) Would the lipid composition of HDLz and HDL, be changed during alimentary lipemia in a direction suggesting better substrate performance, i.e., enrichment in substrate lipids [unesterified cholesterol (UC) and phosphatidylcholine (PC)] and reduction in product lipids [ester cholesterol (EC) and lysophosphatidylcholine (LyPC)]? 3) Could those HDL lipid changes that correlate with LCAT augmentation in vivo be duplicated in vitro with resulting activation of the enzyme?

In vivo studies
Eight normal young men (mean age, 24 yr), free of known systemic disease, served as subjects for fat-loading studies. All had normal plasma cholesterol and triglycerides, using age-adjusted values of Fredrickson, Levy, and Lees (10) and normal lipoprotein electrophoretic patterns by agarose electrophoresis (Pfizer Diagnostics, NY). Informed consent was obtained from each in a manner approved by the Human Studies Subcommittee of this hospital, according to regulations established by the Veterans Administration.
Duplicate venous blood samples (15-30 min apart) were obtained at 8-9 AM from subjects after a 12-14 hr fast and immediately chilled in ice water in tubes containing 1.0 mg of Na,EDTA/ml blood. The subjects then ingested a liquid high-fat test meal over approximately 10 min. The composition of the test meal was 36% cream, 4 ml; skimmed milk, 4 ml; sucrose, 0.25 g; and cocoa, 0.12 g. Sufficient volumes were mixed and blended to allow feeding at 1.5 g fat per kilogram body weight. A 70-kg subject would receive 105 g of fat, 47 g of carbohydrate, and 19 g of protein. A 2.0 g/kg dose was fed to obtain lipoprotein fractions for in vitro experiments. Blood samples were drawn at 2.5-hr intervals up to and including the 10th hour postfeeding. N o additional food was allowed during the test.

Chemical methods
Chilled blood samples were centrifuged at 4°C to remove cellular elements. Aliquots of plasma were taken for determination of a ) triglycerides by an enzymatic method (Boehringer-Mannheim, NY), b) cholesterol using the colorimetric method of Rude1 and Morris (1 l), and c) LCAT using a method previously described and validated (12). Precision of the LCAT method is *16% (2 SD). For analysis of lipoprotein fractions, diethyl p-nitrophenylphosphate (Paraoxon, K&K Labs, Plainview, NY) was added to blood (final concentration, 1.0 mM) to inhibit LCAT and lipase activities during handling. Aliquots of plasma (5 ml) were adjusted to 1.063 g/ml with NaBr-NaC1 (350: 150, w/w) and centrifuged for 30 hr at 40,000 rpm in a 40.3 rotor at 6°C with an L3-50 ultracentrifuge (Beckman Instruments, Spinco Division, Palo Alto, CA). The top fraction, VLDL plus low density lipoproteins (LDL), was removed by tube slicing and the bottom fraction was raised to a solvent density of 1.125 g/ml, using solid salts. This fraction was then centrifuged for 30 hr in the 40.3 rotor. The top fraction, HDL, (d 1.063-1.125 g/ml), was removed by slicing approximately 0.7-0.9 cm below the meniscus. The subnatant solution was raised to 1.220 g/ml with solid salts and 4.5-5.0 ml of the solution was layered under 1.0-1.5 ml of 1.220 g/ml salt solution using a syringe and long needle. After 30 hr of centrifugation the top fraction, HDL,, (d 1.125-1.220 g/ml) was removed by tube slicing. The subnatant fraction was designated very high density plasma fraction (VHD). The lipids in this fraction were those of very high density lipoprotein and albumin (LyPC and free fatty acids).
Purity of HDL,, HDL,, and VHD fractions was tested by lipoprotein and protein electrophoresis on agarose. HDLz contained only alpha-1 lipoprotein and a thin band in the pre-beta region (sinking pre-beta). HDL, yielded only a single alpha-1 lipoprotein band. Neither HDL subfraction contained detectable albumin by immunoelectrophoresis or by diffusion on Ouchterlony plates. However, HDL3 proteins did contain 0.7-1.1 % of albumin as judged by quantitative radial immunodiffusion (LC Partigen, Behring Diagnostics, Somerville, NJ). Both HDL fractions reacted with commercial anti-HDL antisera (Behring Diagnostics) on immunodifussion and immunoelectrophoresis, forming one arc when stained for lipid or protein. VHD contained only traces of lipid in the albumin region on electrophoresis and did not react with anti-HDL antisera.
The protein content of isolated lipoproteins was determined by a modification of the method of Lowry et al. (13). Lipid extracts, prepared by extraction with chloroform-methanol 2: 1 (v/v) as described by Folch, Lees, and Sloane Stanley (14), were separated on micro-columns of silicic acid essentially as described by Barron and Hanahan (15). Fractions containing cho-lesteryl esters, triglycerides plus unesterified cholesterol, and phospholipids were eluted sequentially. The purity of each fraction was assessed by thin-layer chromatography and showed negligible cross contamination. Mean recovery of ester and free cholesterol for 28 samples was 102 f 8% (SD). The TG content of the second fraction was determined; lipid phosphorus in the third fraction was estimated by the Bartlett method (1 6). The major individual choline-containing phospholipids were determined after separation on Adsorbosil-5 plates using chloroform-methanol-water 75:25:4 as the developing solvent mixture.
Inasmuch as the experimental design employed duplicate postabsorptive sampling and analysis, it was possible to appraise the variability of the lipoprotein fractionation by comparing duplicate results for total cholesterol content of lipoprotein subfractions. Biologic variation during the short period between bleedings is probably small. After lipoprotein separation, additional analytic error arises from lipid extraction, separation procedures, and chemical methods. For 14 sets of duplicates, mean coefficients of variation (kSD) were as follows: HDLz, total cholesterol, 3.8 k 2.4%; HDL,, total cholesterol, 5.0 +-3.1%; HDL2-EC, 3.4

In vitro net transfer experiments
The capacity of VLDL present in plasma 5 hr after fat ingestion to reduce HDL-EC levels by nonenzymatic transfer (17,18) was assessed in five studies with three normal and two primary hypertriglyceridemic male subjects. The latter patients were mildly hypertriglyceridemic (type IV patterns) without demonstrable etiology. In each study postabsorptive and alimentary lipemic plasma samples were incubated at 37°C for 3 hr to allow determination of LCAT. Paraoxon (2 mM) was then added to inhibit LCAT during an ensuing 12-hr transfer period. Samples at onset, 3 hr, and 15 hr time points were taken for determination of plasma LCAT, UC, and EC. Enzyme inhibition was nearly complete during the transfer period, since further consumption of UC was not observed. Duplicate 4.0-ml aliquots, obtained at 3 and 15 hr, were layered under 2.0 ml of phosphate-buffered saline (pH 7.4) for ultracentrifugal isolation of VLDL (d < 1.006 g/ml). After slicing the centrifuge tube to remove the VLDL, the subnatant solution was adjusted to I .063 g/ml with concentrated salt solution for isolation of LDL and the d > 1.063 g/ml fraction. The latter was designated HDL. Isolated lipoprotein solutions were extracted with chloroform-methanol 2: 1 (v/v) and the lipid extracts were fractionated by thin-2 2.7%; HDL,-EC, 4.6 +-2.7%; HDL2-TG, 11.2 layer chromatography using hexane-ether-acetic acid 90: 10: 1 for separation of EC and UC. After elution of the lipids from the plate with chloroform, portions of the eluate were used for cholesterol determination.

In vitro pre-incubation experiments
Studies were performed with plasmas from five normal subjects and one with primary type IV hyperlipoproteinemia in order to compare depletion of cholesteryl esters in HDL subfractions by pre-incubation with postabsorptive and alimentary lipemic VLDL and the effect, if any, on LCAT activity. In these experiments VLDL (d < 1.006 g/ml) was isolated by ultracentrifugation from fasting and lipemic plasmas obtained 5 hr post feeding from the same individual. Fasting and lipemic plasmas were then reconstituted at the concentration of original unfractionated plasma. In order to test the effect of alimentary lipemic VLDL, a third sample consisted of lipemic VLDL mixed with the d > 1.006 glml fraction from postabsorptive plasma. Prior to recombination, each d > 1.006 g/ml fraction was labeled with [3H]cholesterol as described previously (1 9, 20). 5,5'-Dithiobis-2-nitrobenzoic acid (DTNB, K&K Laboratories) was added to a concentration of 1.4 mM (21) in order to inhibit LCAT. Plasma samples were then pre-incubated for 5 hr at 37°C. At that time mercaptoethanol (11.8 mM) was added to activate the enzyme (21). Aliquots were taken for plasma lipids, radioactivity, and ultracentrifugal isolation of HDL2, HDL,, and VHD as previously described. Incubations were continued for 3 hr and the reaction was terminated by extraction of an aliquot with isopropanol and addition of Paraoxon (2 mM) to another aliquot prior to untracentrifugation. After lipoprotein fractionation, solutions were extracted with chloroform-methanol 2:l and ester and free cholesterol were separated on thin-layer plates for quantitation of mass and radioactivity.

Plasma LCAT and TG responses to fat ingestion
Mean postabsorptive plasma lipid levels for the group of eight normal subjects were as follows: total cholesterol, 192 k 13.0 (SEM) mg/dl; UC, 53.1 f 3.6 mg/dl; EC, 139 f 9.6 mg/dl; and TG, 75.8 k 13.5 mg/dl. Peak plasma T G response after fat ingestion was at 5 hr, averaging 108 f 30% (SEM) for the group. At 10 hr TG had fallen to below basal levels. Plasma LCAT was increased above the mean basal level of 61.3 ? 4.2 (SEM) pmol/l per hr at 2.5 hr, reached a broad plateau between 5 and 7.5 hr, and fell slightly   (2) have reported a similar temporal response pattern but appreciably smaller enzyme elevations.
Lipoprotein TG responses were more variable, since three of eight subjects displayed either negative responses or no change during the 2.5-to 7.5-hr primary response period. Negative responses occurred in those subjects exhibiting the smallest elevations of plasma TG. Cholesteryl esters increased by 5 hr, with a broad peak at 7.5-10 hr. Increases also occurred in proteins, UC, total phospholipids, PC, and the molar ratio of PC/UC. The mean percentage increase in PC was approximately double that of the other components exclusive of' TG. Fig. I , panel B , shows changes in HDL, during the period of alimentary lipemia. Basal levels of HDL3 components were as follows: UC, 7.0 k 0.7 (SEM) mg/dl; EC, 21.1 ? 2.2 mg/dl; PC phosphorus, 1.25 k 0.1 1 mg/dl; protein, 102 k 6.6 mg/dl; and TG, 3.7 ? 1.3 mg/dl. HDL,-TG followed a pattern similar to that of plasma and HDL2. As with HDL,-TG, both positive (% subjects) and negative (96 subjects) responses occurred. Individual HDL3-TG changes, however, did not correlate significantly with plasma TG changes. Cholesteryl esters showed a small mean reduction greatest at 7.5 hr (-5.4%). Fig. 1, panel B , also depicts the pattern of change in other HDL, components. Proteins and UC showed little change except for a small UC rise at 10 hr. Phosphatidyl choline showed a slow rise beginning at 5 hr with a 10-hr peak. The PC/UC molar ratio did not change significantly.
During alimentary lipemia, HDL2 maximal EC increments were highly reciprocally correlated with maximal HDL2-TG changes (r = -0.85, P = 0.01).  (17) from incubation experiments with normal plasma. It should be noted that subjects (three out of eight) who exhibited reductions in lipoprotein TG levels (Fig. 2, solid triangles and circles) were those showing largest increases in EC levels. Also of note is that the intercept of the HDL2 regression line on the x axis showed nearly 40% elevation of EC with no change in TG. This contrasted with the HDL, regression with an x-intercept near zero. Thus, TG changes appear to be prime correlates of EC levels in HDL,, while an additional source must be invoked to explain the large ester elevation in HDL2 unrelated to TG fluctuations. These esters might derive from lipolysis of VLDL and chylomicrons or by transfer from HDL, and VHD. Fig. 1, panel C, illustrates the effect of fat feeding on the VHD fraction. Basal levels of VHD lipids were as follows: UC, 2.4 k 0.2 (SEM) mg/dl; EC, 4.8 k 0.5 mg/dl; PC phosphorus, 0.34 2 0.02 mg/dl; and LyPC phosphorus, 0.33 & 0.04 mg/dl. Two and a half hours after feeding, only PC showed a small increase, while total phospholipids, UC, EC, and LyPC were unchanged. By 5 hr all components had fallen below baseline and remained at low levels throughout the period, except for elevation at UC at 10 hr. The fall in EC, which averaged 17% at 7.5 hr, was observed in seven out of eight subjects (Fig. 2). Lysolecithin fell by 23% at 5 hr.

Correlations between HDL changes and plasma LCAT responses
Although large changes occurred in HDLz during assimilation of dietary fat, including increases in the LCAT substrate lipids, PC and UC, there was no correlation between LCAT responses and any change in this subfraction. Similarly, there was no relationship between LCAT and any change in the VHD fraction. Maximal LCAT responses, however, were positively correlated with maximal changes in HDL3-TG (r = 0.74, P < 0.05) and negatively with maximal changes in EC ( r = -0.80, P < 0.05). As described in the preceding section, maximal EC and T G changes were inversely correlated (T = -0.88, P = 0.01). In order to separate the contribution of each LCAT correlate arising from each variable, partial correlation coefficients were calculated, alternately holding the third variable constant. The TG:EC correlation, with LCAT held constant, was reduced only modestly to r = -0.71. The LCAT correlation with TG, with EC changes held constant, was greatly reduced to r = 0.13. Thus, the possibility of an independent effect on LCAT through TG changes seems small. Similarly, the LCAT correlation with EC, with TG held constant, was reduced from r = -0.80 to -0.47. This residual correlation, although still substantial, was not statistically significant. It should be pointed out that three of eight subjects who exhibited negative TG responses and increased EC levels in their HDL, nevertheless had significant elevations of plasma LCAT, although less than the mean response. This might indicate that the activation mechanism is not removal of inhibition by cholesteryl esters, as suggested by in vitro studies of Fielding, Shore and Fielding (22). However, flotation at d 1.125 g/ml does not discretely separate two lipoprotein components. Consequently, LCAT substrates in HDL, could have experienced reduced EC levels which would be obscured by contamination with HDL, esters. In view of this uncertainty, in vitro experiments were next employed to test directly the possibility that activation of the enzyme was due to decreased inhibition as a result of EC depletion.

In vitro nonenzymatic net transfer experiments
Nonenzymatic net transfer of HDL-EC to VLDL in apparent exchange for triglycerides has been described by Nichols and Smith (17), and Glomset, Norum, and King (23) using both normal and endogenous hyperlipemia plasmas. To test for this phenomenon in alimentary lipemic plasmas, both fasting and alimentary lipemic (5 hr post feeding) plasma samples from the same individuals were incubated for 3 hr to allow measurement of LCAT activity. This was followed by a 12-hr transfer period during which LCAT was inhibited by Paraoxon. During this transfer period an absolute reduction in HDL-EC occurred in all samples, ranging from 2.9 to 12.6 mg/dl in postabsorptive plasmas and 4.3 to 13.8 mg/dl in lipemic samples. Paired t tests confirmed greater absolute EC decreases during the transfer period in HDL from alimentary lipemic plasmas compared with fasting (mean decrement, 8.5 k 3.9 mg/dl and 6.9 * 3.8 mg/dl respectively, P < 0.01). The difference in EC reduction between groups was 23.2%. This decrement did not correlate with pretransfer period LCAT activity (r = 0.53, P > 0.05), but did correlate inversely with VLDL cholesterol levels as determined after isolation from samples at 3 hrs of incubation ( r = -0.75, P < 0.05), immediately after the addition of Paraoxon.
Very low density lipoprotein cholesterol also correlated with LCAT activity (r = 0.86, P < 0.01). During incubation VLDL cholesteryl esters increased in all postabsorptive samples, but greater elevations in corresponding post-feeding samples were demonstrable only in those subjects who exhibited the largest post-feeding increases in VLDL.

In vitro pre-incubation experiments
The results of the transfer experiments indicated that increased VLDL, resulting from fat ingestion, would increase transfer of cholesteryl esters from the HDL to VLDL. Since depletion of HDL3 esters might activate LCAT, the possibility of in vitro activation by this mechanism was tested. The experimental design employed a 5-hr pre-incubation period in the presence of an LCAT inhibitor and radioisotopic dispersion, since this period is similar to the lag period for peak activation observed in vivo. Comparisons were made by replacement of postabsorptive VLDL with alimentary lipemic VLDL from the same subject. After the pre-incubation period, enzyme was activated for 3 hr by addition of mercaptoethanol and the activity was determined. Prior to these experiments, two experiments were performed replacing fasting with alimentary lipemic VLDL and one testing a lipemic VLDL plus LDL fraction (d < 1.063 g/ml), but without pre-incubations. No changes in LCAT resulted from these replacements.
In plasmas from five normal subjects and one primary hypertriglyceridemic subject, the effects of replacement by alimentary lipemic VLDL on postabsorptive HDL,, HDL3, and VHD-EC content were determined. The results (Table 1) revealed significantly greater reduction of HDL, esters (P < 0.02) after 5 hr of pre-incubation with lipemic VLDL. The mean of individual reductions was 10.3%, ranging from 23.2% to + 1.2%. The mean maximal HDL3-EC change observed post feeding in vivo was -3.4%, ranging from -16.0 to 14.0%. Of those displaying decreases (five out of eight, Fig. 2), the mean decrease was 11.9%. Thus, in vitro pre-incubation produced depletion of a similar magnitude to that noted in vivo. This change was not associated with in vitro activation of LCAT during the following period of enzyme activity (Table 1). However, some LCAT comparisons in lipemic plasmas exceeded the 16% limit previously established as a significant change (12), but the precision of the determination is probably reduced by processing, including use of an inhibitor and activator. Analysis by paired t test showed no significant activation. In contrast, addition of alimentary lipemic VLDL to the d > 1.006 g/ml fraction from lipemic plasma, which reconstitutes the lipemic plasma as obtained 5 hrs post feeding, did result in significantly increased activity (P < 0.025), even though the hypertriglyceridemic patient (exp. 4) did not respond to fat loading with increased activity. LCAT response to fat feeding in hyperlipemic subjects is variable and follows different temporal patterns compared to normals. Table 1 also shows that lipemic VLDL failed to deplete postabsorptive HDLz esters. A mean individual reduction of 5.0%, (range, -14.8% to +3.6%) was found (P > 0.05), but the hyperlipemic VLDL (exp. 4) did yield the largest individual reduction. Like HDL3, VHD esters were reduced (-0.6 mg/dl, -12%, P < 0.05). Rose < 1.006 giml) was removed from post-absorptive plasma from five normal subjects and one hypertriglyceridemic subject (exp. 4) and replaced with fasting VLDL or alimentary lipemic VLDL isolated 5 hr after fat ingestion. Concentrations are those in original unfractionated plasma. Mixtures were incubated for 5 hr in presence of DTNB and then activated for 3 hr by mercaptoethanol and LCAT was determined (pmol/l/hr). HDL, and HDL, ester cholesterol (EC) levels are in mg/dl as obtained at termination of the 5-hr pre-incubation period. T h e last column lists LCAT activity obtained by mixing lipemic VLDL with the d > 1.006 g/ml fraction from lipemic plasma and is, therefore, reconstituted lipemic plasma. P value is for difference with reconstituted fasting plasma.
In these experiments radioactive unesterified cholesterol had been added to samples and esterification was determined during the activation period. The distribution of radioactive esters was determined in HDL subfractions at the termination of the activation period. The order of specific activities relative to plasma in subfractions isolated from reconstituted postabsorptive plasmas was VHD > HDL, > HDL, (3.8/2.6/1.3). Similar ratios were obtained from fasting plasmas substituted with lipemic VLDL and reconstituted lipemic plasma. This order of labeling suggests that the VHD fraction contains the most reactive substrates (9). Fractional distribution of labeled esters in the fasting samples was HDL2, 12.3 2 2.3% (SD); HDL,, 40.7 5 8.3% (SD); and VHD, 13.3 k 2.4% (SD). Pre-incubation with lipemic VLDL resulted in a shift in distribution significantly decreasing the fraction in HDL, and VHD to 38.0% and 12.1%, respectively, with no change in HDL,. The fraction remaining in the total HDL was reduced from 66.3% to 62.7%, thereby indicating enhanced distribution of esters to lipoproteins of d < 1.063 glml during the period of enzyme activity.

DISCUSSION
Stimulation of LCAT activity during alimentary lipemia in man has been reported from three laboratories (1 -3). This phenomenon is reproducible and associated with large changes in HDL, and HDL,. These associations raise the possibility that new sub-strate lipoproteins of intestinal origin are responsible for the activation effect. HDL (24)(25)(26) and possibly LCAT (24) are synthesized by rat intestine. Furthermore, a substantial increase in HDL phospholipids, particularly lecithin, occurs during alimentary lipemia in man, which could reflect a characteristic of new substrates that would make for enhanced substrate reactivity. Thus, analysis of HDL subfraction changes that are simultaneous with LCAT measurements might reveal relationships with augmentation of enzymatic activity. We have performed such analyses during alimentary lipemia in eight normal male volunteers. Although the largest changes in protein and lipid content were observed in the HDL,, none was significantly correlated with LCAT response. This was true even though the HDL, were greatly enriched in lecithin, which is a substrate lipid, along with another substrate lipid, free cholesterol. The HDL3 also exhibited enrichment in lecithin, but once again this failed to correlate with LCAT response. However, the temporal pattern of lecithin increase in HDLz and HDL, generally paralleled the LCAT response pattern. In view of the report of Wallentin and Vikrot (2) that fat-induced LCAT responses correlated with HDL phospholipid changes and that phospholipid dispersions activate the enzyme in vitro, direct testing of substrate reactivities with purified enzyme will be needed to clarify the significance of this relationship.
Changes in triglyceride levels in the HDL, did, however, correlate with LCAT response while cholesteryl esters were inversely related. However, weak correlations between these individual lipid levels and enzyme response by partial correlation analysis suggest that neither lipid alone is instrumental in regulating activity. The ester response assumes mechanistic importance, since Fielding et al. (22) found reactivities of 1ecithin:free cholesterol dispersions to be strongly inhibited by the addition of cholesteryl esters. Thus, relief from product inhibition by increased removal of cholesteryl esters from LCAT substrate lipoproteins is a potential explanation for enzyme activation. Partial correlation analysis, however, did not support this possibility, since the correlation with LCAT was substantially diminished if HDL, triglycerides were held constant. The main correlate of cholesteryl ester content in HDL, is triglyceride content, as shown by a strong inverse relationship between these two lipid constituents (Fig. 2). The mechanism is speculative, but the phenomenon seems analogous to the nonenzymatic TG:EC interchange reported in vitro by Nichols and Smith (17) and Rehnborg and Nichols (18). In the experiments of these workers VLDL-TG transferred to the HDL and HDL-EC transferred to VLDL. This observation has been confirmed by Glomset et al. (23), who also noted that LCAT activity stimulated accumulation of TG in the HDL. Although the phenomenon is nonenzymatic, that is, it proceeds while LCAT is inhibited, it is nevertheless enhanced in the presence of active enzyme (1 7, 18, 23).
In vivo we have also shown that VHD cholesteryl esters fall during lipemia, but triglycerides in this fraction could not be measured. Interestingly, the lysolecithin content of the VHD also fell during lipemia. The explanation might be related to an observation of Minari and Zilversmit (27), who noted that dog chylomicrons lose phospholipids to the HDL during in vitro incubations, but chylomicron lysolecithin increases. In reverse binding experiments, lysolecithin has been reported to preferentially bind to lipoproteins rather than to apoproteins or albumin (28). This suggests a role for lipids in binding affinity. Since lysolecithin is a potent inhibitor of LCAT activity (22), accelerated removal by binding to chylomicrons and VLDL, leading to reduced levels in substrate lipoproteins, is a possible explanation for activation.
In vitro experiments were performed in order to distinguish between two possible explanations for in vivo associations between LCAT and HDL lipids. I ) LCAT activation is a result of depletion of HDL, and VHD cholesteryl esters by transfer to lipemic VLDL. 2) Transfer of esters to lipemic VLDL reduced HDL, and VHD ester levels, but without influence on enzymatic rate. A related question is whether the extent of ester transport to VLDL by a nonenzymatic mechanism, whatever its effect on LCAT activity, is influenced by ambient levels of enzyme activity. In the first experiment it was necessary to demonstrate that alimentary lipemic VLDL, like hyperlipemic VLDL (17,23), would reduce HDL cholesteryl esters and that newly synthesized radioactive cholesteryl esters formed in LCAT substrates within HDL would participate in transfer to the VLDL and LDL. The positive results of this experiment extend similar observations of Nichols and Smith (17), Glomset et al. (23), and Akanuma and Glomset (8) to alimentary lipemic VLDL. Both of these phenomena were observed and, in addition, partial correlation analysis suggests that LCAT levels do not condition the extent of subsequent EC transfers, which are primarily related to VLDL and plasma TG levels. A similar conclusion was also reached by partial correlation analysis of our in vivo HDL,-EC content data.
The second in vitro experiment utilized the demonstrated capacity of alimentary lipemic VLDL to reduce HDL cholesteryl esters. Cholesteryl ester levels were decreased during a 5-hr pre-incubation by the nonenzymatic mechanism in the HDL, and VHD fractions. Similar depletion of HDL2 esters could not be demonstrated (Table 1). Likewise, radioactive cholesteryl esters formed during the activation period were reduced in HDL, and VHD, but not in HDL2. 'These results favor the conclusion that endogenous LCAT substrate esters are among those removed. Although reduction of the HDL, esters was comparable to the largest reductions observed in vivo, in vitro activation of LCAT did not result. This corroborates negative reports by Rose and Juliano (12) and Wallentin and Vikrot (2) in experiments that did not employ preincubation and provides evidence against a modulating role for endogenous substrate cholesteryl ester content in the regulation of LCAT activity. The seeming paradox that increased LCAT activity is associated with reductions rather than increases of cholesteryl esters and lysolecithin in lipoprotein fractions containing substrate macromolecules may be explained by increased removal of product lipids and perhaps lipoproteins through transfer to VLDL, which are at elevated levels during the period of enzyme activation. It is possible that transfer of product cholesteryl esters to VLDL prevents accumulation of esters in substrate lipoproteins thereby permitting augmentation of enzymatic activity. Even though correlation analysis did not show a significant LCAT correlation with transfer in vitro or with changes in HDL, and VHD-EC levels in vivo, there exists the possibility that LCAT product lipoproteins constitute a pool of reactive lipoproteins that interact with VLDL.I Supported by funds from the Medical Research Service of the Veterans Administration. The authors are indebted to Ms. Polly Ellerby for excellent technical assistance and to