点击显示 收起
1 From TNO Biomedical Research, Leiden, Netherlands (ID, PJV, PCNR, WvD, JJE, and LMH); TNO Innovative Ingredients and Products, Zeist, Netherlands (WFN); the Department of General Internal Medicine, University Medical Center, Leiden, Netherlands (ID, PCNR, and LMH); the Department of Endocrinology and Metabolic Diseases, University Medical Center, Leiden, Netherlands (PJV and JAR); and the Department of Cardiology, Leiden University Medical Center, Leiden, Netherlands (LMH)
2 Conducted in the framework of the "Leiden Center for Cardiovascular Research LUMC-TNO" and supported by the Netherlands Organization for Scientific Research (NWO-grant 903-39-179 to LMH, NWO VIDI grant 917.36.351 to PCNR, NWO VENI grant 916.36.071 to PJV, and program grant 903-39-291 to LMH and JAR) and the LUMC (Gisela Thier Fellowship to PCNR). 3 Reprints not available. Address correspondence to LM Havekes, Biomedical Research, Gaubius Laboratory, Zernikedreef 9, PO Box 2215, 2301 CE Leiden, Netherlands. E-mail: lm.havekes{at}pg.tno.nl.
ABSTRACT
Background: The prevalence of dyslipidemia and obesity resulting from excess energy intake and physical inactivity is increasing. The liver plays a pivotal role in systemic lipid homeostasis. Effective, natural dietary interventions that lower plasma lipids and promote liver health are needed.
Objective: Our goal was to determine the effect of dietary sphingolipids on plasma lipids and liver steatosis.
Design: APOE*3Leiden mice were fed a Western-type diet supplemented with different sphingolipids. Body cholesterol and triacylglycerol metabolism as well as hepatic lipid concentrations and lipid-related gene expression were determined.
Results: Dietary sphingolipids dose-dependently lowered both plasma cholesterol and triacylglycerol in APOE*3Leiden mice; 1% phytosphingosine (PS) reduced plasma cholesterol and triacylglycerol by 57% and 58%, respectively. PS decreased the absorption of dietary cholesterol and free fatty acids by 50% and 40%, respectively, whereas intestinal triacylglycerol lipolysis was not affected. PS increased hepatic VLDL-triacylglycerol production by 20%, whereas plasma lipolysis was not affected. PS increased the hepatic uptake of VLDL remnants by 60%. Hepatic messenger RNA concentrations indicated enhanced hepatic lipid synthesis and VLDL and LDL uptake. The net result of these changes was a strong decrease in plasma cholesterol and triacylglycerol. The livers of 1% PS–fed mice were less pale, 22% lighter, and contained 61% less cholesteryl ester and 56% less triacylglycerol than livers of control mice. Furthermore, markers of liver inflammation (serum amyloid A) and liver damage (alanine aminotransferase) decreased by 74% and 79%, respectively, in PS-fed mice.
Conclusion: Sphingolipids lower plasma cholesterol and triacylglycerol and protect the liver from fat- and cholesterol-induced steatosis.
Key Words: APOE*3Leiden mice • sphingolipids • steatosis • cholesterol • triacylglycerol • free fatty acids
INTRODUCTION
In our modern Western society, excess energy intake and physical inactivity are the leading causes of the epidemic prevalence of obesity. Obesity in turn is associated with an increased prevalence of cardiovascular disease risk factors, such as dyslipidemia, insulin resistance, and hypertension—collectively called the metabolic syndrome (1). Obesity-related dyslipidemia is characterized by mildly elevated concentrations of VLDL- triacylglycerol and LDL cholesterol and decreased concentrations of HDL cholesterol. The liver plays a central role in the maintenance of systemic lipid homeostasis because it synthesizes and secretes VLDL lipoproteins and, thus, is involved in the redistribution of lipids, primarily triacylglycerol, for storage and utilization by peripheral tissues. Lipid accumulation in the liver leads to the development of steatosis, a condition closely associated with insulin resistance (2).
The metabolic syndrome is usually treated by a combination of lifestyle and dietary changes. The drugs that are currently used target one aspect of the metabolic syndrome. For example, hypercholesterolemia is treated with HMG-CoA reductase inhibitors (statins) and cholesterol absorption inhibitors (ezetimibe, phytosterols, and stanols), hypertriglyceridemia is treated with peroxisome-proliferator activated receptor (PPAR) agonists (fibrates), hypertension is treated with ß-blockers (atenolol, metoprolol, and propranolol), and insulin resistance is treated with thiazolidinediones (pioglitazone and rosiglitazone) or metformin. Although recently developed compounds, such as the glitazars, target more than one aspect of the metabolic syndrome, treatment of multiple aspects of the metabolic syndrome with a single natural dietary compound could be an attractive alternative.
We recently observed in a pilot study that consumption of Western-type diet supplemented with sphingolipids lowered both plasma cholesterol and triacylglycerol in hyperlipidemic APOE*3Leiden mice. The APOE*3Leiden mouse has a lipoprotein profile that closely resembles the human profile. In these mice, plasma cholesterol can be titrated to various concentrations by varying the amount of cholesterol in the diet (3). Moreover, in contrast with wild-type mice, LDL receptor–deficient mice and apolipoprotein (apo) E–deficient mice, APOE*3Leiden mice, are highly sensitive to treatment with hypolipidemic drugs, such as statins (4, 5) and fibrates. We wondered, therefore, whether supplementation of the diet with sphingolipids could be used to treat the dyslipidemia characteristic of the metabolic syndrome.
MATERIALS AND METHODS
Sphingolipids
We used the 3 sphingolipids that represent the most abundant and simplest natural sphingolipid classes—the sphingoid bases (Figure 1; I, II, and III), which can be formed via the enzymatic breakdown of complex sphingolipids in the intestine—and the 3 complex natural sphingolipids (Figure 1; IV, V, and VI). Sphingomyelin (mainly N-palmitoyl-sphingosine-1-phosphocholine) from egg was obtained from Larodan Fine Chemicals (Stockholm, Sweden). Yeast-derived (semi)synthetic ceramide III (N-stearoyl-phytosphingosine), cerebroside (N-stearoyl-phytosphingosine-1-glucose), and phytosphingosine (PS) were from Cosmoferm BV (Delft, Netherlands). Sphinganine and sphingosine were from Avanti Polar Lipids (Albaster, AL).
FIGURE 1.. Structures of the sphingolipids used in this study.
Animals and diets
Female heterozygous 6-mo-old APOE*3Leiden transgenic mice (3) were fed a Western-type diet (Hope Farms, Woerden, Netherlands) for 5 wk, which provided 15% cocoa butter, 0.25% cholesterol, 1% corn oil, 40.5% sucrose, 20% acid casein, 10% cornstarch, and 5.95% cellulose (all by wt). The mice were housed during the experiment in clean conventional animal rooms (relative humidity: 50–60%; temperature: 21 °C; light cycle: 0600 to 1800) and were supplied with food and acidified tap water ad libitum. Mice were housed in macrolon cages (3 mice per cage). Body weight (20.9 ± 0.7 g) and food intake (2.5 ± 0.1 g/mouse per day) were monitored weekly. In the first experiment (Figure 2), the mice were randomized after the 5-wk diet to 1 of 7 groups (n = 6 per group) on the basis of plasma cholesterol and triacylglycerol and body weight. Subsequently, the mice were fed for 3 wk the same diet without or with 0.1% (by wt) PS, sphingosine, sphinganine, cerebroside, ceramide III, or sphingomyelin. Then, the sphingolipid dose was increased to 0.2% (by wt) for 3 wk and finally to 0.4% (by wt) for 3 wk. Tail vein blood samples were obtained after the mice were deprived of food for 4 h, at randomization, and at 3, 6, and 9 wk.
FIGURE 2.. Mean (±SD) plasma cholesterol and triacylglycerol concentrations in APOE*3Leiden mice after food deprivation for 4 h. After a run-in period of 4 wk, during which the mice were fed a Western-type diet, female APOE*3Leiden mice were fed a Western-type diet (control) or the same diet supplemented with increasing doses [0.1%, 0.2%, or 0.4% (by wt)] of various sphingolipids (see Figure 1) for 3 wk each (n = 6 per group). Baseline values for cholesterol and triacylglycerol were 15.0 ± 1.9 and 2.3 ± 0.5 mmol/L, respectively. *Significantly different from control, P < 0.05 (ANOVA followed by Dunnett's test).
In all subsequent experiments, 6-mo-old female APOE*3Leiden transgenic mice were fed the Western-type diet for 5 wk and then randomized as described previously. Subsequently, the mice were fed the Western-type diet for 5 wk or the same diet supplemented with 1.0% (by wt) PS before being subjected to experimentation. All experiments were approved by the TNO Animal Care and Use Committee.
Plasma variables
Tail blood samples were collected in EDTA-coated cups, or in paraoxon-coated capillaries to prevent lipolysis (6). Plasma lipid variables were measured by using commercial kits for total cholesterol (Roche Diagnostics, Mannheim, Germany), nonesterified free fatty acids (NEFA-C; Wako Chemicals, Neuss, Germany), triacylglycerol (Triacylglycerol GPO-Trinder; Sigma, St Louis, MO), ß-hydroxybutyrate (Sigma), and alanine aminotransferase (ALAT) (Reflotron GPT; Roche Diagnostics). Serum amyloid A (SAA) was measured by enzyme-linked immunosorbent assay (Biosource, Nivelles, Belgium) and fibrinogen by sandwich enzyme-linked immunosorbent assay, as described previously (7). For lipoprotein fractionation, groupwise-pooled plasma was size-fractionated by fast-protein liquid chromatography on a Superose 6 column (Äkta; Amersham Pharmacia Biotech, Uppsala, Sweden). Fractions were assayed for total cholesterol and triacylglycerol as described.
Intestinal cholesterol absorption
We assessed cholesterol absorption using the fecal dual-isotope method described by Borgstrom (8) and Wang and Carey (9) in mice fed the diet with or without 1% PS (6 per group) for 5 wk. The mice were housed individually. At 1700 the mice received a dose of 200 µL olive oil containing [14C]cholesterol (1 µCi/mouse; Amersham, Little Chalfont, United Kingdom) and [3H]sitostanol (1 µCi/mouse; ARC, St Louis, MO) by gavage. Body weight and food intake were monitored, and feces were collected for 4 d. Feces were freeze-dried, homogenized, pooled per mouse over the 4-d period, and dissolved in ethanolic potassium (3 mol/L, 60% ethanol). Radioactivity was determined in the fecal samples to assess the amount of radiolabeled cholesterol and sitostanol. Sitostanol was used as the reference compound because it is known to be poorly absorbed (<3%) in mice (9). The formula used to calculate cholesterol absorption was as follows:
RESULTS
Sphingolipids lower plasma cholesterol and triacylglycerol in APOE*3Leiden mice
For our initial experiment, to evaluate the effect of sphingolipids on plasma cholesterol and triacylglycerol concentrations in APOE*3Leiden mice, we used 3 simple and 3 complex sphingolipids (Figure 1). The data were analyzed by 2-factor ANOVA. After demonstrating that the interaction between dose and treatment was significant (P < 0.001), we analyzed the data for each dose separately by ANOVA, followed by Dunnett's test. With 0.1% of these sphingolipids in the diet, no significant effect on plasma cholesterol (P = 0.978) and triacylglycerol (P = 0.398) was seen (Figure 2). At a dose of 0.2% (by wt), the sphingolipids (except for ceramide and cerebroside) significantly decreased the plasma cholesterol concentration by 20–40% (P = 0.0096; Figure 2). At a dose of 0.4% sphingolipids, plasma cholesterol decreased even more, and ceramide also had a significant cholesterol-lowering effect (P = 0.0009; Figure 2). The decrease in triacylglycerol concentration was 40% for all sphingolipids at dietary sphingolipid concentrations of 0.2% and 0.4%—a significant decrease for all compounds at weeks 6 and 9 (Figure 2). No differences in food intake or body weight were observed throughout the experiment between the mice fed sphingolipids and the control animals (data not shown). Remarkably, the simplest sphingolipids, the sphingoid bases (Figure 1), had the same potent cholesterol- and triacylglycerol-lowering effect as their complex sphingolipid derivatives.
To study the mechanisms underlying the cholesterol- and triacylglycerol-lowering effects, we performed studies in mice fed the Western-type diet with or without 1% PS for 5 wk. This sphingolipid was chosen for all subsequent studies because it is one of the simplest in the sphingolipid class; it is the central structural element of ubiquitous sphingolipids of plants and yeasts that are part of our diet. PS can be formed in situ in the intestine by enzymatic degradation of complex sphingolipids. No effects on body weight or food intake were ever observed (data not shown). As can be seen in Table 1, mice fed a 1% PS–containing diet showed, as expected, a strong and significant decrease in plasma cholesterol and triacylglycerol. Plasma FFA concentrations also decreased significantly, whereas plasma ß-hydroxybutyrate (a liver-derived ketone body) did not change significantly (Table 1). Groupwise-pooled plasma was used to determine the lipoprotein profiles of these mice. The lipoprotein profiles showed that the decrease in cholesterol and triacylglycerol was confined to the VLDL and IDL-LDL fractions, whereas HDL cholesterol did not change (Figure 3).
View this table:
TABLE 1. . Plasma variables in APOE*3Leiden mice deprived of food for 4 h after being fed a control Western-type diet or the same diet supplemented with 1% phytosphingosine (PS) for 5 wk1
FIGURE 3.. Cholesterol and triacylglycerol concentrations in pooled plasma samples derived from 6 APOE*3Leiden mice deprived of food for 4 h after being fed a Western-type diet (control) or the same diet supplemented with 1% (by wt) phytosphingosine (PS) for 5 wk. Concentrations were measured in individual fractions after separation by fast-protein liquid chromatography. IDL, intermediate-density lipoprotein.
Phytosphingosine reduces intestinal cholesterol, triacylglycerol, and FFA absorption
We assessed the effects of dietary sphingolipids on intestinal cholesterol absorption by measuring absorption with the fecal dual-isotope method, using [14C]C and [3H]sitostanol, in mice fed a 1% PS–containing or control Western-type diet for 5 wk. No differences in body weight, food intake, and fecal output were observed between the 2 groups (data not shown). We observed that intestinal cholesterol absorption in PS-fed mice was only one-half that in the mice fed the control diet (Figure 4). This reduction was also reflected by fecal neutral sterol excretion, which was twice as high in the PS-fed mice as in the control mice (Figure 4). We next determined whether the observed decrease in plasma triacylglycerol and FFAs was also due to decreased intestinal absorption. Mice fed a control or 1% PS–containing Western-type diet for 5 wk were used. After being deprived of food overnight, the mice were intravenously injected with Triton WR1339 to inhibit lipoprotein lipolysis, followed by an intragastric gavage of [3H]triolein and [14C]oleic acid in olive oil. In the group fed 1% PS, the olive oil gavage contained, in addition to [3H]triolein and [14C]oleic acid, 1% PS. Serum triacylglycerol, 3H activity, and 14C activity were assessed over a 4 h-period. The time-dependent appearance of plasma 3H activity (a measure of intestinal triacylglycerol uptake) is depicted for the 2 groups in Figure 5. PS reduced the intestinal [3H]triacylglycerol uptake by 33% after 4 h. The plasma appearance of [14C]oleate (a measure of intestinal FFA uptake) was 43% lower in the PS-fed mice than in the control mice 4 h after administration.
FIGURE 4.. Mean (±SD) intestinal cholesterol absorption and neutral sterol excretion in mice (n = 6 per group) fed a Western-type diet (control) or the same diet supplemented with 1% (by wt) phytosphingosine (PS) for 5 wk. The mice received a dose of 200 µL olive oil containing [14C]cholesterol (1 µCi/mouse) and [3H]sitostanol (1 µCi/mouse) by gavage. Feces were collected for 4 d. Radioactivity was determined in the fecal samples to assess the amount of radiolabeled cholesterol and sitostanol. Sitostanol was used as the reference compound because it is known to be poorly absorbed (<3%) in mice. The formula used to calculate cholesterol absorption was as follows: percentage cholesterol absorption = ([14C]/[3H] dosing mixture – [14C]/[3H] feces)/[14C]/[3H] dosing mixture x 100. *Significantly different from control, P < 0.05 (Student's t test).
FIGURE 5.. Mean (±SD) intestinal triacylglycerol (TG) and free fatty acid (FFA) absorption in mice (n = 6 per group) deprived of food overnight and then fed a Western-type diet (control) or the same diet supplemented with 1% (by wt) phytosphingosine (PS) for 5 wk. The mice were injected with Triton WR1338 to block lipoprotein lipase–mediated triacylglycerol hydrolysis and then given an intragastric load of 200 µL olive oil containing [3H]triolein (12 µCi/mouse) and [14C]oleic acid (3.3 µCi/mouse). For the PS-fed animals, 1% PS was added to the olive oil load, because deprivation of food overnight might negate any direct effect that PS might have on intestinal absorption. Plasma samples were obtained to determine the appearance of triacylglycerol-derived [3H]fatty acids or [14C]FFAs in the blood. dpm, decays per minute. *Significantly different from control, P < 0.05 (2-factor ANOVA after a significant interaction between time and treatment was established).
Phytosphingosine increases hepatic VLDL-triacylglycerol production
Hepatic VLDL-triacylglycerol production was studied in mice deprived of food overnight with the use of the Triton WR1339 method and by measuring plasma triacylglycerol accumulation. In these experiments, plasma triacylglycerol concentrations increased faster in the mice fed a 1% PS–containing diet than in the mice fed the control Western-type diet. The VLDL-triacylglycerol production rate, as determined from the slope of the curves, was 20% higher in the PS-fed mice than in the control mice (Table 2). Analysis of the composition of the VLDL particles (isolated by ultracentrifugation) showed that the triacylglycerol content as well as the phospholipid content increased by 66% and 17%, respectively, in VLDL particles derived from PS-fed mice (Table 2). Total cholesterol, in contrast, decreased by 51% in the VLDL particles from the PS-fed mice. The de novo total apo B production rate in newly synthesized VLDL particles did not differ significantly between mice fed the PS-containing diet and the control mice (Table 2). The data indicate that the number of VLDL particles secreted by the liver was not affected; however, the VLDL particles from PS-fed mice contained less cholesterol but more triacylglycerol.
View this table:
TABLE 2. . Production rates of VLDL-triacylglycerol (TG) and VLDL-apolipoprotein (apo) B and VLDL particle composition in APOE*3Leiden mice fed a control Western-type diet or the same diet supplemented with 1% phytosphingosine (PS) for 5 wk1
Phytosphingosine increases the liver-mediated clearance of plasma cholesterol but not plasma triacylglycerol
Plasma cholesterol and triacylglycerol concentrations are not only determined by their production rate but also by their clearance rate, ie, by their uptake, lipolysis, or both. VLDL-like particles containing [3H]triolein and [14C]cholesteryl oleate, which were previously shown to mimic the metabolic behavior of triacylglycerol-rich lipoproteins (17, 28), were used to determine the effects of a 1% PS–containing Western-type diet on plasma clearance. PS accelerated the plasma clearance of [14C]cholesteryl oleate (half-life: 39.5 ± 5.3 and 74.5 ± 9.9 min in the PS-fed and control mice, respectively; P < 0.05; Figure 6). The enhanced removal of cholesterol from the blood was corroborated by the increased liver uptake of [14C]cholesteryl oleate (60% at 20 min; Figure 6). Although the LPL-dependent serum clearance of [3H]triacylglycerol was not affected (half-life: 4.7 ± 0.3 and 4.9 ± 0.3 min for he PS-fed and control mice, respectively; Figure 6), [3H]triacylglycerol uptake in the liver 10 and 20 min after injection increased significantly in the PS-fed mice (Figure 6). No effects were observed on the uptake of [14C]cholesteryl oleate or [3H]triacylglycerol-derived radioactivity by various peripheral muscle and adipose tissues (data not shown). Taken together, the results of this experiment indicate an increase in VLDL particle remnant uptake in the liver.
FIGURE 6.. Mean (±SD) in vivo clearance of VLDL-like emulsion particles in the serum and liver of mice (n = 6 per group) fed a Western-type diet (control) or the same diet supplemented with 1% (by wt) phytosphingosine (PS) for 5 wk. The mice were injected into the vena cava inferior with VLDL-like emulsion particles containing [3H]triolein (triacylglycerol; TG) and [14C]cholesteryl oleate (CO) at a dose of 300 µg triacylglycerol per mouse. *Significantly different from control, P < 0.05 (2-factor ANOVA after a significant interaction between time and treatment was established).
Hepatic mRNA concentrations indicate increased lipid synthesis
The hepatic expression of several genes was studied by using RT-PCR on liver samples of mice deprived of food for 4 h after being fed the 1% PS–containing or the control Western-type diet for 5 wk (Table 3). The mRNA concentrations of genes involved in FA and triacylglycerol synthesis and secretion (srebp1c, fas, mttp, and dgat2) increased. Acc1, dgat1, aco, and apob transcription concentrations did not change. Furthermore, mRNA concentrations of cholesterol homeostasis genes, such as srebp2, ldlr, and hmgcoAred, also increased. In strong contrast, the 2 genes involved in bile salt formation studied decreased (cyp7a1 and lxrß), whereas the expression of lxr and fxr increased. Ppar, ppar, and abca1 expression did not change (Table 3). Overall, these changes suggest increased hepatic lipid and cholesterol synthesis and decreased bile formation, which indicates a shift in hepatic triacylglycerol and cholesterol homeostasis compared with the control situation.
View this table:
TABLE 3. . Hepatic messenger RNA (mRNA) expression of genes involved in fatty acid (FA) and triacylglycerol (TG) and cholesterol homeostasis in APOE*3Leiden mice fed a control Western-type diet or the same diet supplemented with 1% phytosphingosine (PS) for 5 wk, as determined by real-time polymerase chain reaction1
Phytosphingosine protects livers from steatosis
Increased remnant uptake and increased mRNA expression of genes involved in FA and triacylglycerol synthesis suggested that intrahepatic triacylglycerol and cholesterol concentrations increased in our setting. However, at autopsy of these mice, it was noted that the livers of the mice fed the 1% PS–containing Western-type diet were of normal size and had a dark-red appearance, whereas the livers of the control mice fed the Western-type diet without PS were enlarged and yellowish (Figure 7, A and B). This indicates that the hepatic lipid content was lower in the PS-fed mice than in the control mice. Microscopical examination of HPS-stained sections showed that, compared with the controls, the PS-fed mice had less lipid-filled vacuoles in the liver cells (Figure 7, C and D). The livers of PS-fed mice weighed significantly less (22%) than those of the control mice (Table 4). Lipid analysis showed that the liver triacylglycerol content of PS-fed mice was less (56%) than that of the control mice. Furthermore, liver cholesteryl ester concentrations decreased by 61% and free cholesterol concentrations decreased by 11% in PS-fed mice. All these differences were significant (Table 4).
FIGURE 7.. Macroscopic appearance of a liver from a mouse fed the control Western-type diet for 5 wk (A) and from a mouse fed the same diet supplemented with 1% phytosphingosine (PS) (B). Hematoxylin-phloxine-saffron–stained histologic micrographs of paraffin-embedded livers from a mouse fed the Western-type diet for 5 wk (C) and a mouse fed the same diet containing 1% PS (D) are also shown.
View this table:
TABLE 4. . Liver lipid variables and plasma inflammation markers in APOE*3Leiden mice deprived of food for 4 h after being fed a control Western-type diet or the same diet supplemented with 1% phytosphingosine (PS) for 5 wk1
Phytosphingosine lowers plasma inflammatory markers
Hepatic steatosis is often associated with liver inflammation. We measured plasma ALAT concentrations as a measure of liver damage. The ALAT concentration decreased by 79% (P < 0.05) after PS feeding. A 74% (P < 0.05) decrease in the concentration of the acute phase marker SAA was found in the PS-fed mice. The fibrinogen concentration was 42% higher (P < 0.05) in the PS-treated mice than in the control mice (Table 4).
DISCUSSION
From this study we conclude that sphingolipids dose-dependently decrease plasma cholesterol and triacylglycerol concentration in APOE*3Leiden mice fed a Western-type diet. This cholesterol- and triacylglycerol-lowering effect is mediated through inhibition of intestinal absorption of both cholesterol and triacylglycerol and, eventually, leads to protection of the liver from fat and cholesterol-induced steatosis.
Because the intestinal absorption of FAs and triacylglycerols were inhibited to the same extent, we conclude that intestinal triacylglycerol lipolysis per se is not inhibited by sphingolipids but that the absorption of FAs is impaired. We do not know the exact mechanism underlying the PS-mediated inhibition of FA absorption. However, the FA-PS complex formation via an ionic interaction between the negatively charged carboxylic acid group of FAs and the positively charged primary amine of PS might be an explanation. Although complex sphingolipids do not have such a primary amine group, in the intestine they can be lipolyzed to some extent into sphingoid bases and, thus, the same mechanism will, eventually, also hold for these complex molecules. However, it has been observed that some dietary sphingolipids are not fully digested and are partly excreted via the feces (29).
The intestinal FA absorption experiments (Figure 5) were performed with the use of an intragastric olive oil load. As compared with conditions of normal ad libitum food intake, intragastric olive oil loads represent extreme conditions of intestinal FA absorption. Thus, quantitative extrapolation of the observed inhibitory effect of PS on the FA absorption during a intragastric fat load to FA absorption under normal feeding conditions is hazardous. Under normal feeding conditions, the effect of PS on FA absorption is assumed to be more modest, and the mice may compensate with increased glucose utilization. The observation that a PS-containing diet does not significantly affect food intake or body weight in the long term is in line with this reasoning.
Intestinal absorption of cholesterol depends on bile salts and is favored by the presence in the intestine of triacylglycerol-derived fatty acids that form mixed micelles with bile salts in which cholesterol is solubilized (30). Because we observed no effect of PS on intestinal triacylglycerol hydrolysis (Figure 5), disturbance of the composition of these mixed micelles by sphingolipids, which leads to hampered solubilization of cholesterol, seems unlikely. The formation of stable cholesterol and sphingomyelin (or sphingosine) complexes has been described (29, 30) and could, as such, be the cause of reduced intestinal absorption of cholesterol. However, whether the formation of cholesterol-sphingolipid complexes also occurs in the intestine has been questioned, because the high affinity of cholesterol for sphingomyelin is lost in the presence of bile salts (31, 32). We found that both simple and complex sphingolipids (Figures 1 and 2) decrease plasma cholesterol and triacylglycerol concentrations in APOE*3Leiden mice. Because of the diversity in chemical structure between the various sphingolipid species, a wide range of physical and chemical properties may be expected and, thus, the inhibition of intestinal absorption of cholesterol and triacylglycerol by PS is not likely to be explained by specific complex formation with bile salts or disturbance of bile salt micelles in the intestinal lumen. However, other yet unknown biochemical processes may be influenced by dietary sphingolipids or their metabolites. For example, the effect of the sphingolipid diet on the bile salt profile was not investigated, but changes in that profile (and production) potentially may affect the intestinal lipid absorption and may thus influence plasma lipid concentrations. This hypothesis is sustained by the observation that genes involved in the bile salt synthesis are indeed affected by the sphingolipid diet (Table 3).
The lowering effect of PS feeding of APOE*3Leiden mice on both cholesterol and triacylglycerol clearly differs from the effects of feeding stanol esters, which results in the lowering of cholesterol only. In APOE*3Leiden mice, stanol ester feeding leads to a reduction in plasma cholesterol, whereas plasma triacylglycerol concentrations remained unaffected (5). Importantly, in contrast with the present dietary sphingolipid study, the expression of hepatic key genes involved in lipid metabolism (ie, ldlr and cyp7a1) were not affected in APOE*3Leiden mice fed plant stanol esters (33). In addition, PS proved to be about twice as effective at decreasing plasma cholesterol in APOE*3Leiden mice as stanol esters. At a dose of 1% plant stanol esters for 9 wk, a 33% decrease in total plasma cholesterol was observed (5), whereas PS feeding at the same concentration reduced plasma cholesterol by 57%. Reciprocally, as did PS, dietary stanol esters decreased hepatic cholesteryl ester, free cholesterol, and triacylglycerol concentrations in APOE*3Leiden mice (33).
Because blood samples were taken from animals that were deprived of food for 4 h and were thus not expected to have intestine-derived chylomicrons in their plasma, the decreased concentrations of plasma cholesterol and triacylglycerol could not be directly ascribed to the PS-mediated inhibition of intestinal cholesterol and triacylglycerol absorption. We reasoned that a reduction in the absorption of dietary plus biliary cholesterol leads to a reduction in the liver cholesterol pool, as presented in Table 4. A reduction in the cholesterol pool in the liver leads to a reduction in bile acid synthesis as reflected by a reduced expression in the liver of bile acid synthesis genes such as lxrß and cyp7a1, concomitant with an increased expression of genes involved in hepatic cholesterol synthesis (hmgcoA reductase) and hepatic cholesterol uptake from plasma (ldlr) (Table 3). It can also be concluded from the results presented in Table 3 that the expression of genes involved in lipid synthesis (srebp1c, fas, and dgat2) and VLDL production (mttp) increased. Indeed, as presented in Table 2, the measurement of VLDL-triacylglycerol production in vivo showed a 20% enhanced secretion of VLDL-triacylglycerol by the liver, whereas the number of VLDL particles secreted into the circulation (apo B synthesis) remained constant.
The observation that these newly synthesized VLDL particles exhibit a 50% reduction in cholesterol content (Table 2) and a strongly increased hepatic uptake of their remnants (Figure 6), the latter in line with the increased expression of hepatic ldlr (Table 3), offers an explanation for the cholesterol-lowering effect of dietary PS. On the contrary, the increased hepatic secretion of VLDL-triacylglycerol does not sustain the triacylglycerol-lowering effect of PS. From Figures 6 it can be concluded that, after LPL-mediated lipolysis, >80% of VLDL-derived triacylglycerol is taken up by peripheral tissues within 20 min; the remainder is taken up by the liver (Figure 6). A difference in the plasma clearance of VLDL-triacylglycerol between the PS-fed and control mice could not be observed under the experimental conditions used (Figure 6). However, as depicted in Figure 6, the hepatic uptake of VLDL-triacylglycerol increased significantly in the PS-fed group, thereby sustaining the triacylglycerol-lowering effect of PS feeding. In PS-fed mice, the newly synthesized VLDL particles were larger and relatively enriched in triacylglycerol compared with control fed mice (Table 2). Because the affinity of triacylglycerol-rich emulsion particles (17) and lipoproteins (34) for LPL is higher with increasing particle size, this change in lipid composition could also (partly) explain the triacylglycerol-lowering effect of PS feeding.
Although sphingolipids have been suggested to be PPAR agonists (35), we found no changes in expression of PPAR or in genes that are under PPAR control, such as aco (Table 3). Furthermore, we did not observe enlarged livers in PS-treated mice, as would be expect with PPAR agonists in rodents (36, 37). In contrast, compared with the control mice, the livers of PS-fed mice weighed significantly less, contained less lipid (cholesteryl esters and triacylglycerol, Table 4), and contained smaller lipid-filled vacuoles in the parenchymal cells (Figure 7D). To maintain its lipid homeostasis, the liver compensates for the lower PS-mediated dietary and biliary cholesterol and triacylglycerol supply from the intestine by increasing its endogenous cholesterol and fatty acid synthesis, as is reflected in the increased hepatic mRNA concentrations of fas and hmgcoAred. However, hepatic lipid synthesis is well regulated by a feedback mechanism. Consequently, on PS-feeding, hepatic cholesterol or triacylglycerol synthesis increases but lipids do not accumulate in the liver (Table 4).
In steatotic livers, lipid accumulation is accompanied by an increase in plasma ALAT and SAA, which are markers of liver damage and liver inflammation, respectively. In PS-fed mice, both markers decreased dramatically, pointing to a true hepatoprotective effect of dietary PS under conditions of Western-type diet feeding. Liver steatosis and elevated ALAT and SAA concentrations are associated with insulin resistance and the metabolic syndrome in mice and humans (38, 39). Furthermore, inflammatory reactions are involved in the atherosclerotic process (40). Therefore, dietary sphingolipids should be considered as compounds for treating or ameliorating not only the lipid component of cardiovascular disease but also the inflammatory processes involved in atherosclerosis and insulin resistance. We suggest that dietary sphingolipids hold great potential to treat multiple aspects of the metabolic syndrome, such as dyslipidemia, insulin resistance, and cardiovascular diseases.
ACKNOWLEDGMENTS
We acknowledge the technical assistance of Lars Verschuren, Karin Toet, Kirsty Brachel, and Erik Offerman in the SAA, fibrinogen, neutral sterol, and histologic analyses and the practical assistance of Annemarie Maas in the experiments.
ID, PJV, PCNR, and WvD were responsible for data collection and analysis. ID was responsible for drafting the manuscript. JAR contributed intellectual input. LMH and PCNR were responsible for project supervision. LMH, PJV, PCNR, JJE, and WFN were involved in the project concept and design and provided intellectual input. All authors reviewed the final manuscript. None of the authors had a conflict of interest.
REFERENCES