Reverse cholesterol transport is elevated in carboxyl ester lipase-knockout mice

Abstract

Mechanisms to increase reverse cholesterol transport (RCT) and biliary sterol disposal are currently sought to prevent atherosclerosis. Previous work with HepG2 cells and primary hepatocytes showed that carboxyl ester lipase (CEL), a broad-spectrum lipase secreted by pancreas and liver, plays an important role in hydrolysis of high-density lipoprotein (HDL) cholesteryl esters (CEs) after selective uptake by hepatocytes. The effect of CEL on RCT of HDL cholesterol was assessed by measuring biliary and fecal disposal of radiolabeled HDL-CE in control and Cel^−/− mice. Radiolabeled CE was increased 3-fold in hepatic bile of Cel^−/− mice, and the mass of CE in gall bladder bile was elevated. Total radiolabeled transport from plasma to hepatic bile was more rapid in Cel^−/− mice. Fecal disposal of radiolabel from HDL-CE, as well as total sterol mass, was markedly elevated for Cel^−/− mice, primarily due to more CE. RCT of macrophage CE was also increased in Cel^−/− mice, as measured by excretion of radiolabel from injected J774 cells. Increased sterol loss was compensated by increased cholesterol synthesis in Cel^−/− mice. Together, the data demonstrate significantly increased RCT in the absence of CEL and suggest a novel mechanism by which to manipulate plasma cholesterol flux.—Camarota, L. M., Woollett, L. A., Howles, P. N. Reverse cholesterol transport is elevated in carboxyl ester lipase-knockout mice.

The beneficial effects of high-density lipoprotein (HDL) derive, in part, from its central role in reverse cholesterol transport (RCT), the movement of cholesterol from peripheral tissues to the liver, where it is removed from circulation and eliminated from the body as biliary cholesterol and bile acids (1). In recent years, much attention has been focused on understanding the mechanisms that underlie this overall process, with the goal of developing means to enhance RCT as treatment to prevent or reduce atherosclerosis (2). One focus has been regulation of cholesterol efflux from macrophage foam cells to nascent or circulating HDL particles. The cholesterol transporters responsible for this step (ABCG1, ABCA1, SR-BI) have been identified (3–5), and regulation of the process involves nuclear receptors of the liver X receptor (LXR) and the peroxisome proliferator-activated receptor (PPAR) subfamilies, as well as cellular sterol and oxysterol levels (6–8). Increasing expression of the cytoplasmic cholesterol ester hydrolase (CEH) can also increase efflux by mobilizing the accumulated cholesteryl ester (CE) pool (9). Considerable attention has also been devoted to methods of raising HDL to increase the total capacity for RCT, and apoAI mimetics have shown some promise in this regard (10). Raising HDL levels has also been the focus of cholesteryl ester transfer protein (CETP) inhibitors, which prevent CE/triglyceride exchange between HDL and remnant lipoproteins. However, it has not been established that simply raising HDL cholesterol enhances RCT (11–13).

A key aspect of RCT is the selective uptake and subsequent metabolism of CE and cholesterol from HDL by hepatocytes. It is now well accepted that SR-BI plays a major role in the selective uptake of HDL-CE by hepatocytes and that HDL cholesterol is largely targeted for biliary sterols (14, 15). Approximately 75% of HDL cholesterol is esterified and must be hydrolyzed to free cholesterol by the liver before secretion into bile or conversion to bile salts for secretion. Thus, hydrolysis is an important first step in the hepatic processing of HDL-CE for the last critical stage of RCT. The identity of the lipases responsible for this hydrolysis is not clear. One candidate is the cytoplasmic CEH, extensively studied by Ghosh and others (16, 17). Most evidence indicates that CEH mobilizes cholesterol from an intracellular CE pool (17), although recent work showed that CEH can increase RCT by hydrolyzing HDL-CE, primarily for bile acid synthesis (18).

Another candidate for HDL-CE hydrolysis is carboxyl ester lipase (CEL), which is expressed by hepatocytes, as well as by pancreatic acinar cells (19). Our previous work with cell culture models and gene-knockout mice showed that CEL is associated with the SR-BI pathway in hepatocytes and that it plays a significant role in the hydrolysis of HDL-CE during or immediately after selective uptake via SR-BI (20). We have continued to investigate the role of this enzyme in the transhepatic movement of cholesterol from HDL into bile, with specific attention to its effect on RCT. The current report presents comparative data from Cel^−/− and control mice that show markedly increased fecal disposal of HDL-CE, as well as macrophage-derived CE in the absence of CEL. This increase in RCT appears to result from a combination of increased secretion of unhydrolyzed HDL-CE directly into bile and feces, and decreased hydrolysis and reabsorption of this CE by the intestine of Cel^−/− mice.

MATERIALS AND METHODS

Animals and diets

Control and Cel^−/− mice were derived by mating heterozygous males and females. Genotypes were determined by PCR, as described previously (20, 21). The CEL mutation is carried on the C57BL/6J genetic background and is backcrossed to the reference colony (Jackson Laboratories, Bar Harbor, ME, USA) yearly to minimize genetic drift. Mice were housed in the institution's vivarium with a 12-h light-dark cycle, controlled temperature and humidity, and ad libitum access to food and water. All procedures were approved by the University of Cincinnati College of Medicine Animal Care and Use Committee. All data are derived from experiments performed with male mice fed standard rodent chow without added cholesterol (Teklad LM485; HarlanTeklad, Madison, WI, USA).

Bile collection and biliary clearance of radiolabeled HDL

Flowing bile was collected from anesthetized mice via a polyethylene catheter inserted into the fundus of the gall bladder, as described previously (22). Collections were performed approximately midway through the light cycle (11 AM–2 PM) on non-food-deprived animals. Human HDL₃ (1.125<ρ<1.21) was isolated and radiolabeled with [³H]cholesteryl oleate, as described previously (20), and injected into the inferior vena cava just before insertion of the catheter (up to 750,000 dpm/mouse, maximum of 0.45 mg HDL protein). Bile was collected hourly for 2 h, and total radiolabel content was determined by scintillation spectrometry. Free and esterified cholesterol was resolved by TLC, the appropriate bands were scraped, and the relative amount of radiolabel in each fraction was determined by scintillation spectrometry (20). Gall bladder bile was also collected midway through the light cycle from non-food-deprived animals by excising the gall bladder immediately after euthanization.

Biliary lipid and bile salt analyses

CE mass in bile was demonstrated by TLC separation of gall bladder bile. Plates were stained with 10% phosphomolybdic acid in ethanol followed by heating at 80°C for 10 min to visualize bands. CE and total cholesterol were quantitated using the direct fluorimetric method described by Mizoguchi et al. (23), with some modifications. While CE was measured exactly as described, the method was modified to measure total cholesterol by leaving cholesterol oxidase out of the free cholesterol decomposition reagent, so that both free and esterified cholesterol are measured in the second step. The fluorescence intensities were measured using a multiwell plate reader equipped with a filter for excitation and emission at 544 and 590 nm, respectively. Phospholipid concentration was determined by colorimetric assay (Wako, Richmond, VA, USA). Total bile acid concentration was determined by colorimetric assay (Trinity Biotech plc, Bray, Ireland).

Fecal sterols

Animals were individually housed in cages fitted with a wire platform with full access to food and water. After a 1- or 2-d acclimation period, feces were collected for 72 h, and sterols were quantitated using a variation of the method described by Post et al. (24). After being dried and weighed, fecal material was ground to a fine powder. 5α-cholestane (40 μg) and [24-¹⁴C]-taurocholic acid (0.02 μCi) were added to 0.5 g of the ground fecal material as internal standards for extraction efficiency of neutral and acidic sterols, respectively. The fecal material plus standards was suspended in 10 vol of alkaline methanol (0.2 M NaOH in 80% methanol) and incubated at 80°C for 2 h, and neutral sterols were separated by 3 extractions with equal volumes of petroleum ether. The combined organic fractions were evaporated under nitrogen and resuspended in hexane, and a portion was used to measure neutral sterols by gas chromatography (Shimadzu Scientific Instruments, Kyoto, Japan). Data reported represent the sums of areas under the curves for cholesterol, coprostanol, and lathosterol. The aqueous residue from above was filtered, washed, dried, resuspended in water, and applied to prewashed C18 BondElut columns (Varian, Palo Alto, CA, USA). Bile acids were eluted from the columns with methanol, concentrated, and quantitated by colorimetric assay (Trinity Biotech). To measure fecal sterol esters, ground feces were resuspended in water, extracted with petroleum ether, concentrated, and resolved by TLC. Bands were either visualized by staining or (after HDL-[³H]CE injection) scraped and quantitated by scintillation spectrometry as described above.

RCT

Radiolabeled macrophages were prepared by a method similar to that described by Zhang et al. (25). Human LDL+IDL (1.006<ρ<1.063) was labeled with [³H]cholesteryl oleate (GE Life Sciences, New York, NY, USA) by the CETP exchange method (26), reisolated by isopycnic centrifugation, and then acetylated using acetic anhydride (27), followed by concentration and extensive dialysis. The radiolabeled, acetylated lipoproteins were incubated with J774 cells in serum-free medium for 18 to 24 h to achieve ≥0.5 dpm/cell. Cells were harvested, washed, resuspended in serum-free medium, and administered by i.p. injection such that each mouse received ∼1 × 10⁶ cells and ≥0.5 × 10⁶ dpm. Tail blood was drawn at 24 and 48 h to monitor plasma isotope levels, and total fecal output was collected every 24 h for 3 d. Radiolabeled neutral sterols were extracted from feces, as described above, and quantitated by scintillation spectrometry. To measure radiolabel in acidic sterols, a portion of the aqueous residue was dried onto filter paper and oxidized in an OX700 Biological Oxidizer (R. J. Harvey Instrument Corp., Tappan, NY, USA) with concomitant collection of the ³H in scintillation cocktail.

Bile acid pool size

Animals were deprived of food ∼12 h, and tissues were collected midway through the light cycle, immediately after euthanization. Bile acids were isolated from the liver, small intestine, and gall bladder by extraction into ethanol, as described by Schwarz et al. (28), and quantitated by the colorimetric assay described above.

Sterol synthesis

Hepatic sterol synthesis rates were determined essentially as described previously (29). Approximately 4 h into the dark phase, mice received 15 mCi of ³H₂O in saline by i.p. injection. After 60 min, animals were anesthetized, blood was collected from the vena cava, and the specific activity of [³H]₂O in plasma was determined. Livers were removed and saponified with alcoholic KOH, and ³H incorporation into digitonin-precipitable sterols was determined. From these data, synthesis rates were calculated in nanomoles per hour per gram liver based on 25 atoms of hydrogen from water being incorporated per molecule of newly synthesized cholesterol.

Plasma lipids

Serum or plasma was collected from mice deprived of food for 8 to 12 h, and total cholesterol was determined using a colorimetric assay (Wako, Richmond, VA, USA).

Statistics

All experiments except those presented in Figs. 3 and 5 were performed on multiple cohorts of mice with similar results. Data analyses were performed using Prism 5.0 (GraphPad Software, San Diego, CA, USA). All data are presented as means ± sd. Significance was assigned to differences between groups when t tests yielded P ≤ 0.05.

Figure 3.

Excretion of HDL-CE is increased in Cel^−/− mice, and the sterol remains esterified. Total fecal output was collected for 24 h after tail vein injection of radiolabeled HDL-[³H]CE. Neutral sterols were extracted, and radiolabel in free vs. esterified cholesterol was resolved by TLC and quantitated; n = 6 control and 4 Cel^−/− mice. *P < 0.006 vs. controls.

Figure 5.

RCT of macrophage cholesterol is greater in Cel^−/− mice. J774 mouse macrophage cells were loaded with radiolabeled acetylated LDL ([³H]cholesteryl oleate) and given to mice by i.p. injection. Total fecal output was collected for the subsequent 3 d, and radiolabel in neutral (A) and acidic (B) sterols was quantitated as described in Materials and Methods; n = 8 control and 10 Cel^−/− mice. *P < 0.001, ^#P ≤ 0.023.

RESULTS

Hydrolysis and biliary secretion of HDL-CE in Cel^−/− mice

In previous studies, Cel^−/− mice exhibited decreased hydrolysis of HDL-CE after selective uptake by hepatocytes (20). To determine the physiological consequences of this reduced hydrolysis, and thereby the importance of CEL, with respect to hepatic processing of HDL cholesterol into bile, we first measured the appearance of free and esterified cholesterol in flowing hepatic bile after injection of HDL that was radiolabeled with [³H]cholesteryl oleate (HDL-[³H]CE). Flowing bile was collected from control and Cel^−/− mice for 1 h after injection of radiolabeled HDL. Neutral sterols were resolved by TLC, and the relative amount of free cholesterol and CE was determined by scintillation spectrometry. As shown in Fig. 1A, 77.6 ± 12.6% of the biliary cholesterol from HDL-[³H]CE was hydrolyzed to free cholesterol, as has been reported previously (14). However, the opposite was true for Cel^−/− mice, in which only 24.0 ± 10.7% of the radiolabeled cholesterol secreted into bile was hydrolyzed (P<0.0001). This result suggested that CEL is the enzyme primarily responsible for hydrolysis of HDL-CE being delivered to the bile canaliculus. In addition to hydrolysis, the flow rate of radiolabel from HDL-[³H]CE into hepatic bile was monitored. Total bile flow rates (μl/min) were the same in both groups. However Fig. 1B shows that secretion of total radiolabel (cholesterol and bile salts) from the injected HDL-CE was ∼40% faster in Cel^−/− mice than in controls (0.80±0.14 vs. 0.56±0.05% of injected dose in 2 h, P=0.020), indicating increased biliary targeting of HDL-CE in the absence of CEL. Transport of radiolabeled CE from VLDL was similarly investigated but was negligible.

Figure 1.

Reduced hydrolysis of HDL-CE but increased transport to hepatic bile in Cel^−/− mice. A) Flowing hepatic bile was collected for 1 h after injection of HDL-[³H]CE into the vena cava of anesthetized mice. Lipids were resolved by TLC to separate cholesterol from CE, and the amount of radiolabel in each band was quantitated. Data are presented as the percentage of radiolabeled biliary cholesterol that was hydrolyzed; n = 4 control and 7 Cel^−/− mice. *P < 0.0001. B) Total counts transported from HDL-[³H]CE into hepatic bile were determined after 2 h of collection. Data are expressed as percentage of injected dose; n = 4 each. *P = 0.020.

Since biliary cholesterol is derived from HDL free cholesterol and hepatic or newly synthesized cholesterol in addition to HDL-CE (14, 30), the effect of CEL on biliary CE mass was determined. Gall bladder bile was collected and resolved by TLC to determine whether the increased radiolabeled CE seen in flowing bile from Cel^−/− mice reflected a change in CE mass. As shown in Fig. 2A, CE was clearly seen in bile from knockout mice but was barely detectable in bile from control animals. Cholesterol and CE in these biles were also determined by fluorimetric enzyme assays (see Materials and Methods). Figure 2B shows that the mass of CE in gall bladder bile was ∼2 times higher in Cel^−/− mice than in controls, both in absolute concentration (61.8±37.4 μM for Cel^−/− mice vs. 32.0±16.0 μM for controls, P=0.023) and as a fraction of total cholesterol (4.28±2.46% for Cel^−/− mice vs. 2.17±1.10% for controls, P=0.016). Total biliary cholesterol concentration was not significantly different. Bile acid and phospholipid concentrations were also not significantly different between groups. Together, these results indicate that absence of CEL significantly increases the flux of HDL-CE cholesterol from plasma into bile.

Figure 2.

CE content of gall bladder bile is greater in Cel^−/− mice. Gall bladder (GB) bile was collected midway through the light cycle. A) Equal volumes of bile from control and Cel^−/− mice were resolved by TLC and visualized with phosphomolybdic acid. CE is clearly seen in Cel^−/− bile but much less in control bile. B) Free and esterified cholesterol levels were determined in other samples by direct enzymatic assay, as described in Materials and Methods; n = 8 control and 13 Cel^−/− mice. *P ≤ 0.02 vs. controls.

Greater excretion of HDL-CE-derived sterol by Cel^−/− mice

As a first step to determine whether these changes in hepatic HDL-CE processing and bile composition affected RCT, fecal excretion of sterol originating as plasma HDL-CE was measured. Mice received HDL-[³H]CE by i.v. injection, and total fecal output was collected for the subsequent 24 h. Neutral sterols were extracted, and the amount of radiolabel was quantitated. Figure 3 shows that Cel^−/− mice disposed of significantly more cholesterol derived from the radiolabeled HDL-CE than did controls (3874±1560 vs. 1444±389 dpm, P=0.006). Since CEL is also made by the pancreas and is the only cholesterol esterase functional in the intestinal lumen (21, 31), the fecal extracts were analyzed by TLC to determine whether unhydrolyzed HDL-[³H]CE from bile was being excreted intact. After resolution, the bands of free and esterified cholesterol were scraped from the TLC plate, and the radiolabel in each was quantitated. As also shown in Fig. 3, 44.6 ± 7.9% of the radiolabeled fecal cholesterol from Cel^−/− mice remained esterified, while only 9.9 ± 2.4% was esterified in material from control mice (P<0.001). These data suggest that in the absence of CEL, more CE from HDL is transported directly through the liver to the bile and remains unhydrolyzed as it passes through the intestine. Because CE is poorly absorbed without hydrolysis (21, 31), excretion of HDL-CE cholesterol is greatly increased in Cel^−/− mice. Radiolabel in fecal bile acids was also greater in Cel^−/− mice than in controls (2124±461 vs. 1439±474 dpm), but the difference did not reach statistical significance (P=0.066).

Total fecal sterol mass is greater in Cel^−/− mice

To determine whether the increased flux of CE from radiolabeled HDL reflected a physiologically important increase of RCT in Cel^−/− mice, total mass of excreted sterols was quantitated for control and Cel^−/− mice. Fecal output was collected for 3 d from individually housed mice, and neutral sterols were extracted from the saponified material with petroleum ether and quantitated by gas chromatography. Data were normalized to output per day per gram of body weight (bw). As shown in Fig. 4A, fecal neutral sterol mass excreted by Cel^−/− mice was 59% greater than that excreted by control mice (131.2±25.6 vs. 82.6±14.1 nmol/d/g bw for controls, P<0.001). To determine whether the extra neutral sterol excreted by Cel^−/− mice was present as CE, neutral sterols were extracted from unsaponified portions of fecal material and analyzed by TLC. Figure 4B shows that the mass of esterified cholesterol in feces was markedly greater in Cel^−/− mice, in agreement with the analysis of radiolabeled fecal cholesterol from HDL-[³H]CE shown in Fig. 3. Figure 4A also shows that excretion of acidic sterols was increased by the absence of CEL (76.8±23.6 vs. 54.2±9.9 nmol/d/g bw for controls, P=0.034). Bile acid pool size did not differ between the groups (237±33 vs. 221±40 nmol/g bw for controls).

Figure 4.

Total mass of fecal sterols is greater for Cel^−/− mice, including increased mass of esterified cholesterol. A) Neutral and acidic sterols were extracted from feces and quantitated as described in Materials and Methods; n = 8 control and 8 Cel^−/− mice. *P < 0.001; ^#P = 0.034. B) Neutral sterols were extracted from feces without saponification and resolved by TLC to distinguish free and esterified cholesterol. FC, free cholesterol; CE, esterified cholesterol (loaded as markers in outside lanes).

Increased RCT of macrophage CE by Cel^−/− mice

Greater fecal disposal of cholesterol, particularly from HDL-CE, strongly suggested that RCT is increased in Cel^−/− mice. However, a more direct test of this process was performed to verify that cholesterol flux from peripheral (nonhepatic) tissues was increased, specifically the removal of cholesterol from macrophage foam cells. Mouse macrophages of the J774 cell line were radiolabeled by overnight incubation with acetylated LDL that had been prelabeled with [³H]cholesteryl oleate. Radiolabeled cells were harvested, washed, and administered by i.p. injection into control and Cel^−/− mice (8×10⁵ dpm, 1.25×10⁶ cells/mouse) as described previously (25). Total fecal output was collected for the following 3 d, and the amount of radiolabel in neutral and acidic sterols was determined. The data in Fig. 5A show that excretion of macrophage-derived neutral sterols was greatly increased by the absence of CEL, with a combined increase of 37% over the 3 d of collection (512±85 vs. 375±56 dpm/g bw for controls, P=0.001). Figure 5B shows that excretion of bile acids derived from macrophage cholesterol was also increased 30% (1326±170 vs. 1018±168 dpm/g bw for controls, P=0.003). Plasma levels of radiolabel were monitored during the experiment and did not differ between groups (data not shown). These results are consistent with those presented above and show that whole body RCT is elevated in Cel^−/− mice, including the clinically important removal of cholesterol from peripheral macrophage cells.

Increased hepatic cholesterol synthesis in Cel^−/− mice

Plasma cholesterol levels are normal in Cel^−/− mice (118±17 vs. 112±10 mg/dl for controls) as is their lipoprotein profile (FPLC data not shown). Liver cholesterol levels (free and esterified) are also not different (3.8±0.7 vs. 4.1±0.5 μmol free cholesterol/g; 1.9±0.6 vs. 2.3±0.6 μmol CE/g). Because of this, we reasoned that sterol synthesis would be elevated in Cel^−/− mice to compensate for their increased fecal sterol disposal. This was directly tested by measuring hepatic incorporation of [³H]₂O into digitonin precipitable sterols in control and Cel^−/− mice during the dark phase, when cholesterol synthesis is highest. Consistent with the above results, sterol synthesis by the liver was ∼2 times higher in Cel^−/− mice than in controls (1539±668 vs. 695±271 nmol/h/g, P=0.027), as shown in Fig. 6. This result was also corroborated by elevated levels of the cholesterol precursor, lathosterol, in fecal neutral sterols from Cel^−/− mice (included with neutral sterols in Fig. 4A).

Figure 6.

Hepatic cholesterol synthesis is higher in Cel^−/− mice. 15 mCi of [³H]₂O was injected into mice midway through the dark cycle, and digitonin-precipitable radiolabel in liver was quantitated after 1 h, as described in Materials and Methods; n = 5 control and 6 Cel^−/− mice. *P = 0.027.

DISCUSSION

Carboxyl ester lipase is a broad-spectrum lipase capable of hydrolyzing CEs, acylglycerols, lysophospholipids, and lipoamides. Synthesis and secretion of CEL are highest in the pancreas and lactating mammary gland, and the role of the enzyme in fat digestion and absorption has been extensively studied, especially with regard to neonatal nutrition (19). CEL is also made and secreted by hepatocytes and, in humans, by macrophage and endothelial cells (32), but its roles in plasma and hepatic cholesterol metabolism remain unclear. Two independent studies showed a direct correlation between plasma CEL and LDL cholesterol levels in humans (33, 34). On the other hand, studies with human HepG2 cells, as well as hepatocytes from control and Cel^−/− mice, demonstrated a role for CEL in intracellular hydrolysis of HDL-CE (20, 35). We have continued to use gene-knockout mice to further investigate the role of CEL in HDL metabolism, with particular focus on the potential physiological effects of reduced HDL-CE hydrolysis on the fate of these sterols. The key finding in this study is that lack of CEL results in a dramatic increase of RCT both from macrophage and from HDL.

The data show that lack of CEL markedly increased secretion of HDL-CE into bile (Fig. 1A), further indicating that this enzyme plays a key role in the hydrolysis of HDL-CE after selective uptake, as suggested by previous work (20). However, contrary to our original hypothesis, lack of CEL increases the flux of sterols from HDL-CE to bile and disposal in feces (Figs. 1B and 3), rather than reducing the process. Nonetheless, our finding that CE can be secreted into bile is not unique (14), and others showed essentially the same amount of CE from HDL in hepatic bile as in Fig. 1A (14). While it is well established that the twin half-transporters, ABCG5 and ABCG8, play the major role in biliary cholesterol secretion by hepatocytes (36), there is evidence that other mechanisms may also be involved (19, 37). For example, SR-BI has been shown to shuttle between the basolateral and canalicular membranes of hepatocytes (38, 39) raising the possibility of transcytosis of HDL lipids from plasma directly into bile. If so, this would also provide a vesicular carrier for the poorly micellized CE. Interestingly, apoA-I is also present in bile (40, data not shown). Thus, the data presented here are consistent with the notion that some HDL cholesterol is transported across the canalicular membrane by a mechanism other than ABCG5 and ABCG8 and that this alternative mechanism is a contributing factor to the phenotype of Cel^−/− mice. Another factor contributing to the difference between bile from control and Cel^−/− mice can be that the enzyme is present in bile (41) and can hydrolyze any CE secreted, which may be one reason why CE is not usually found in gall bladder bile. In addition to radiolabel, CE mass in gall bladder bile was also significantly increased in Cel^−/− mice (Fig. 2). Lower percentage of CE mass in gall bladder bile compared to percentage of CE radiolabel in hepatic bile reflects the contribution of other sources of biliary cholesterol (14, 30), as well as the different methods of collection. Nonetheless, increased CE in gall bladder bile supports the notion that the physiological effects of CEL on RCT are mediated, at least in part, by changes in hepatic processing of HDL-CE. Despite increased CE in bile, no increase in precipitates or turbidity has ever been detected in bile from Cel^−/− mice (n>50), indicating that the amount of CE present does not have pathological consequences.

In addition to increased secretion of HDL-CE by the liver, lack of CEL in the intestinal lumen prevents hydrolysis and reabsorption of the secreted CE by Cel^−/− mice (21, 31), thus increasing excretion and RCT of HDL cholesterol. Another possibility is that greater sterol loss by Cel^−/− mice results, in part, from an increase in nonhepatic, nonbiliary transport of cholesterol from circulation to the intestine. Recent reports from two independent laboratories have presented evidence that biliary cholesterol does not account for all neutral sterols excreted by wild-type mice under normal conditions (42–44). While the mechanism for the phenomenon remains unknown, if this transintestinal cholesterol efflux involves secretion of esterified cholesterol into the intestinal lumen, sterol excretion by this pathway would also be elevated in Cel^−/− mice.

Interestingly, the increased sterol excretion by Cel^−/− mice included bile acid excretion as well as neutral sterols, both in mass (Fig. 4) and in radiolabel from J774 cells (Fig. 5) and HDL-CE. These data indicate that the absence of CEL causes more HDL-CE to be hydrolyzed by a different lipase in the liver that mobilizes CEs from HDL directly for bile acid synthesis. This explanation is supported by previous investigations showing that ∼50% of HDL-CE hydrolysis is coupled to bile acid synthesis over 72 h in rats (15). While the time frames of most of the experiments described in this report were shorter (≤24 h), we have seen a comparable distribution of HDL-[³H]CE between neutral and acidic sterols in hepatic bile of mice. Consistent with the data in Figs. 4 and 5, the bile acid proportion was seen to increase in several experiments, but the difference did not reach statistical significance (data not shown). A likely candidate for this other activity is the neutral CEH that has been shown to play a role in mobilizing cytoplasmic CEs for bile acid synthesis and secretion by hepatocytes (18). Indeed, hepatic overexpression of this CEH has been reported to increase fecal disposal of cholesterol from injected J774 cells, largely through increased conversion of radiolabeled macrophage cholesterol into bile salts (18).

Plasma and liver cholesterol levels are normal in Cel^−/− mice despite increased fecal sterol disposal. This can be explained by increased hepatic cholesterol synthesis, as illustrated in Fig. 6. This result suggests that under normal conditions, some hydrolyzed HDL-CE enters the cellular cholesterol pool that equilibrates with the endoplasmic reticulum to modulate sterol synthesis. In the absence of CEL, this process is diminished, and synthesis is increased to support normal levels of lipoprotein synthesis and secretion, as well as to maintain cellular cholesterol pools.

Of potential clinical importance, the increased excretion of sterol mass by Cel^−/− mice also included material derived from injected J774 macrophage (Fig. 5). These data indicate that the increased flux of cholesterol in Cel^−/− mice is not limited to the plasma compartment or the liver, but derives from peripheral tissues as well, including macrophage foam cells. RCT, facilitated in ester form by HDL, is considered to be an important factor for the treatment and prevention of atherosclerosis. The data presented in this report suggest that modulating the activity of the lipases responsible for hydrolysis of this CE in hepatocytes may be a mechanism for manipulating hepatic cholesterol traffic and increasing sterol excretion. The results demonstrate that significant changes in sterol flux, specifically RCT, can be effected without raising or lowering static levels of HDL or LDL cholesterol. However, the data imply that lack of CEL activity combined with statin treatment might significantly lower plasma cholesterol, as well as increase RCT. Since mice lack CETP, these results also imply that decreased CEL activity combined with a CETP inhibitor might greatly enhance RCT. These intriguing possibilities await testing in CETP transgenic mice and other appropriate models.