| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Medicine, University of Lund, University Hospital, S-22185 Lund, Sweden
Published, JLR Papers in Press, October 26, 2005.
1 To whom correspondence should be addressed. e-mail: ake.nilsson{at}med.lu.se
 |
ABSTRACT |
|---|
20% of the phospholipids in human plasma lipoproteins, of which two-thirds are in LDL and VLDL. It is secreted in chylomicrons and VLDL and transferred into HDL via the ABCA1 transporter. Plasma SM increases after periods of large lipid loads, during suckling, and in type II hypercholesterolemia, cholesterol-fed animals, and apolipoprotein E-deficient mice. SM is thus an important amphiphilic component when plasma lipoprotein pools expand in response to large lipid loads or metabolic abnormalities. It inhibits lipoprotein lipase and LCAT as well as the interaction of lipoproteins with receptors and counteracts LDL oxidation. The turnover of plasma SM is greater than can be accounted for by the turnover of LDL and HDL particles. Some SM must be degraded via receptor-mediated catabolism of chylomicron and VLDL remnants and by scavenger receptor class B type I receptor-mediated transfer into cells.
Supplementary key words chylomicron • very low density lipoprotein • low density lipoprotein • sphingomyelinase • ceramidase • phospholipid transfer protein
Abbreviations: apoB, apolipoprotein B; BSSL, bile salt-stimulated lipase; CE, cholesteryl ester; CETP, cholesteryl ester transfer protein; ER, endoplasmic reticulum; IDL, intermediate density lipoprotein; LRP, low density lipoprotein receptor-related protein; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PLTP, phospholipid transfer protein; SM, sphingomyelin; SMase, sphingomyelinase; SR-BI, scavenger receptor class B type I; TG, triglyceride
|
INTRODUCTION |
|---|
This review focuses on the absorption and transport of SM and on the biological effects of SM and its metabolites that may be exerted during these processes.
|
ABSORPTION OF SM |
|---|
150 mg of SM per day. SM accounts for 2% of the phospholipids in human bile, which means that 100–200 mg is delivered to the gut every day (24), mainly the palmitoyl and stearoyl species (25). The brush border of the mucosal cells is rich in sphingolipids and is an additional source of endogenous SM. In dog and human chyle chylomicrons, SM accounts for 5% and 7%, respectively, of the total polar lipids (26, 27). Therefore, the question was early raised whether dietary SM could be absorbed intact and contribute to the chylomicron and plasma SM pools. Studies on lymphatic duct-cannulated rats that were fed [3H]sphingosine- or [14C]stearic acid-labeled SM, however, showed that little or no labeled SM was absorbed intact into the chyle (28). SM was sequentially hydrolyzed to ceramide and then to sphingosine and free fatty acids (29). Similarly, after feeding [3H]palmitoyl-sphingosine, no evidence was obtained for the incorporation of intact dietary ceramide into chyle lipoproteins either as ceramide or as SM, but the [3H]palmitic acid appears primarily in chyle TG. A small amount of radioactive ceramide was found in intestinal tissue, but it was not confirmed whether it was located intracellularly as a result of a slow absorption of intact long-chain ceramide or just associated to the mucosal surface. In contrast to SM and ceramide, free sphingosine is well absorbed and rapidly metabolized in the mucosal cells. Most of the absorbed sphingosine is converted to palmitic acid and incorporated into chylomicrons. The key enzymes in this reaction [i.e., sphingosine kinase (30), sphingosine-1-phosphate lyase, and the palmitaldehyde oxidase] are all found at high levels in the intestinal mucosa (31). A smaller portion of the sphingoid bases are reincorporated into mucosal ceramide and more complex sphingolipids (28, 32), and some of this newly formed ceramide appears as chyle ceramide rather than chyle SM (28). The pathways for the metabolism of dietary SM in the gut are summarized in Fig. 1.
|
Factors affecting SM digestion
The course of SM digestion and thus the exposure of distal small intestine and colon to SM and its metabolites can be influenced by multiple factors, such as the amount of SM given, the presence of bile salts and other lipids, and the levels of the enzymes involved in SM digestion. Early studies indicated an extended course and a limited capacity of SM digestion. The recovery of SM [14C]stearic acid in chyle was lower than is generally obtained for glycerolipid fatty acids, and 25% of the sphingosine part appeared in feces, mainly as ceramide (28). This proportion, however, did not increase when the given dose was increased from 1 to 25 mg. Nyberg et al. (46) fed 0.2 to 32 µmol of milk SM together with [3H]sphinganine-labeled dihydrosphingomyelin and analyzed the tissue and the content of four different levels of small intestine after 1–8 h as well as feces collected during 24 h. Increasing the dose increased the proportion and amount of undigested SM in the lower half of the gut, decreased the reincorporation of fatty acids in gut and liver tissue, and increased the output of SM and ceramide in feces.
In a human study, Hertervig (47) found that feeding a dose of 250 mg of milk SM in a standardized meal significantly increased the output of both ceramide and intact SM in ileostomy content in humans, indicating an incomplete digestion and absorption of SM. Although the factors that are rate-limiting for the digestion of SM and ceramide are not fully understood, some relevant observations have been made. When radiolabeled SM was sonicated together with cholesterol or sitosterol and given orally to rats, the digestion of SM was delayed and colonic exposure was increased (48). Studies in vitro using purified alkaline SMase showed that the presence of either polar or nonpolar glycerolipids, or sterols, but not of free fatty acids, inhibited SM hydrolysis by purified alkaline SMase (49). The higher alkaline SMase in jejunum and ileum than in duodenum as well as the presence of higher concentrations of dietary glycerolipids in the upper part may thus favor the location of SM digestion primarily to the middle and lower small intestine.
The regulation of alkaline SMase and ceramidase expression has not been studied extensively. Data from some studies in our laboratory indicate that expression of alkaline SMase and ceramidase can be affected by dietary factors and drugs. It was found that alkaline SMase was increased by psyllium and ursodeoxycholic acid (50, 51) and decreased by a fat-rich diet. Psyllium also decreased ceramidase activity and may thus increase ceramide levels in the gut. Variations in the key apoptotic enzyme caspase 3 correlated positively to alkaline SMase but if anything negatively to acid SMase (52). Insoluble fiber increased acid but not alkaline SMase.
SM digestion in the neonate
Because SM accounts for almost 40% of the polar lipids in human and cow milk, it is important to know whether the neonate can utilize the different components of the SM molecule and whether metabolites formed during the digestion of milk SM may influence normal gut function, maturation, and differentiation. This question has not been thoroughly investigated. Both alkaline SMase and neutral ceramidase, however, were early found to be present in human meconium (29), and studies on fetal and neonatal rats showed that alkaline SMase was promptly expressed soon before birth and the levels kept increasing to a plateau at 4 weeks after birth (53). Three week old suckling pigs had high levels of alkaline SMase in jejunum and ileum (54). The enzymes that digest SM are thus expressed when suckling starts. Acid SMase was shown to be secreted in milk (43). Gastric and duodenal intubation studies of 11 suckling newborns, however, indicated that digestion of milk SM in stomach and upper duodenum is negligible, whereas three jejunal samples and two ileal samples obtained from babies undergoing surgery indicated ceramide formation in these regions of the gut (54). The transfer of nervonic acid, which occurs specifically in sphingolipids, to tissues of newborn rats from mother's milk supports the notion that milk SM is indeed digested and the fatty acids absorbed (55).
During the period from birth to weanling, there is a continuous change in the lipid composition of both the brush border and the basolateral membrane of the gut epithelium. Whereas the brush border increases its cholesterol and SM content during suckling, there is a progressive decrease in the SM/PC ratio in the basolateral membrane (56, 57). The physiological relevance of these changes and whether the supply of SM in milk contributes to these changes are unknown.
|
EFFECTS OF DIETARY SPHINGOLIPIDS IN THE INTESTINE |
|---|
SM has been shown to inhibit cholesterol absorption both in vivo and in CaCo2 cell cultures. When equimolar amounts of cholesterol and SM were fed orally to rats as sonicated dispersions, only 10% of the radioactive cholesterol was absorbed, as estimated by the dual isotope method (48). The inhibitory effect was mutual in the sense that cholesterol and plant sterols also decreased the digestion and absorption of SM (48). Incorporation of cholesterol into SM-PC vesicles was shown to have profound effects on detergent-induced phase transitions, indicating a physicochemical cause of the absorption inhibition (60). In support of this possibility, Eckhardt et al. (61) correlated the effect of SM on the partitioning of cholesterol from bile salt solutions to polymerized silicone to an inhibitory effect of SM on cholesterol uptake in vivo and in differentiated CaCo2 cells. When cholesterol absorption was studied in lymphatic duct-cannulated rats, both egg SM and milk SM decreased cholesterol absorption, milk SM being most effective (62, 63). Interestingly, the lymphatic output of SM decreased rather than increased during SM feeding. Whether this reflects a decreased incorporation of SM into chylomicrons secondary to the decreased cholesterol absorption is unknown. Studies by Kirby et al. (45) and by Field and colleagues (64, 65) showed an increased uptake of cholesterol by CaCo2 cells after hydrolysis of SM or ceramide. Recently, sphingosine was shown to decrease cholesterol absorption in CaCo2 cells and to downregulate the Niemann-Pick-Like Protein 1, which has been implicated as a key protein in cholesterol absorption and a target for the cholesterol absorption inhibitor ezetimibe (66). Not only the physical interaction of SM in gut lumen but also regulatory influences of the metabolites may thus influence cholesterol absorption and lipoprotein composition.
The conclusion is that dietary SM may influence cholesterol absorption and lipoprotein metabolism in the gut by several mechanisms. The important questions of whether SM may influence the absorption of endogenous cholesterol and postprandial lipoprotein metabolism in humans have not been answered.
Effects on cell growth and differentiation
Because ceramide and sphingosine are well-known regulators of cell growth, differentiation, and apoptosis, the questions have been asked whether the metabolites formed from dietary SM may influence the cell cycle of the gut epithelium under normal and tumorigenic conditions, and whether sphingolipid metabolites regulate normal proliferation and differentiation in the crypt cell progenitor compartment and cell fate along the crypt villous axis. Choi et al. (42) found increased apoptosis in absorptive villous cells after feeding a short-chain ceramide, if ceramidase was blocked, and suggested that the brush border ceramidase is needed to protect absorptive cells from the apoptotic effect of absorbed ceramide. Whether this can be applied to long-chain ceramides that are expected to be less effectively absorbed than the short-chain ceramide is not known.
SM in the milk may have important roles in the mucosal development of intestine in the newborn. Milk replacement formulas regularly contain far less SM than does mother's milk. Motouri et al. (67) fed SM-enriched milk formulas or an equivalent amount of PC to suckling rat pups and found lower lactase levels, a larger Auerbach nerve plexus, and restriction of vacuolated epithelial cells to the villous tip in the SM group. More extensive studies in this area, including examination of the intestinal immune system, are needed.
As reviewed by Schmelz (68) and by Duan (69), several pieces of evidence indicate that sphingolipid metabolism may mediate anticarcinogenic mechanisms in the colon. Feeding sphingolipids to animals treated with the chemical carcinogen dimethylhydrazine or to MIN mice, an animal model for human familial adenomatous polyposis, showed an inhibitory effect on both the initiation and propagation of colon tumors. Studies on colon cancer HT9 cells showed that sphingosine induced apoptosis and the inhibition of proliferation by interacting with the ß-catenin system (70). Antitumor effects of SM in the gut may thus be exerted by increasing the exposure of the gut to its degradation products ceramide and/or sphingosine. Although alkaline SMase is expressed at the highest level in the middle and lower small intestine (71), it is found also in the colon mucosa. The enzyme activity was low in many colon tumors, in mucosa from familial adenomatous polyposis patients (72, 73), and in stools of colorectal cancer patients (74). Purified alkaline SMase was shown to inhibit the proliferation of colon cancer HT29 cells (75). Recently, we found that this cell line has an inactivating mutation in the alkaline SMase gene (76). Under normal conditions, notably substantial amounts of alkaline SMase and intestinal ceramidase reach the colon in the intestinal content. These enzymes may protect the colonic mucosa from tumorigenesis by generating ceramide and sphingosine in the colon.
SM is present in bile, where it accounts for a small minority of the total phospholipids in most species examined (24). There is convincing evidence that phospholipids in bile prevent cytotoxicity induced by bile salts (77). Although SM accounts for a small part of the bile phospholipids, it may contribute to this effect (78), particularly in the lower small intestine, where some undigested SM persists but where most other lipids but not the bile salts have been absorbed. SM was also found to prevent deoxycholate-induced apoptosis and hyperproliferation in CaCo2 cells. This provides an alternative mechanism by which SM may have anticarcinogenic effects in colon (79).
Conclusion
SM digestion is an extended process with limited capacity and is catalyzed primarily by alkaline SMase and neutral ceramidase, which are in part released into and act in the gut lumen. The generated sphingosine is well absorbed and effectively metabolized in the mucosal cells. SM and its metabolites in the gut may inhibit TG hydrolysis and cholesterol absorption and affect cell proliferation and maturation. The slow and extended digestion of SM makes it possible to increase the exposure of the whole gut to SM and its metabolites by dietary means. However, many questions regarding the biological effects or the mechanism of the effects remain to be solved, such as the factors limiting ceramide assimilation and the effects on cell growth and differentiation in both the neonate and the adult. It is still not known whether SM has a significant effect on postprandial lipoproteins in humans.
|
SM IN PLASMA LIPOPROTEINS |
|---|
416 µg/ml), accounting for 20% of the total plasma phospholipids (81). Animals with high HDL and low LDL levels, such as dog and rat, have rather high SM levels but a lower SM/PC ratio than humans. Low levels of plasma phospholipids and SM are found in herbivores such as guinea pig and rabbit and in chicken (e.g., plasma SM in the guinea pig amounts to only 5% of the values found in humans). In the plasma, SM is present mainly in different lipoproteins. In human VLDL, SM accounts for 12% of the phospholipids, in LDL plus VLDL it accounts for 28%, and in HDL it accounts for 18%. The SM/PC ratio is lowest in HDL2 and is 2-fold higher in LDL than in VLDL (82). A total of 63–75% of plasma SM is in LDL and VLDL and 25–37% is in HDL (83).
|
SM thus occurs in all lipoprotein classes and has important roles as an amphiphilic surface component with certain specific properties. Being resistant to lipases and LCAT and participating in spontaneous and phospholipid transfer protein (PLTP)-mediated transfer between lipoproteins, its key function may be to preserve particle stability during the wide range of lipoprotein particle modifications that are linked to TG and cholesterol transport. Although SMases are known to be present in blood, their quantitative capacity to metabolize lipoprotein SM has not been established.
Secretion of SM in TG-rich lipoproteins
SM is secreted into blood mainly in chylomicrons and VLDL. On an ordinary Western diet, 3–6 g of phospholipids enters blood via intestinal lipoproteins every day (93), of which 7% is SM (26). The daily input of chyle SM into blood is then on the order of 0.3–0.4 g/day. In nascent VLDL, SM accounts for 5% of the polar lipids (94). If nascent VLDL contains 10% polar lipids, of which 5% is SM, 0.1 g of SM per 20 g of VLDL is secreted. Recent stable isotope kinetic studies of VLDL TG production indicate a fasting production rate of 5–46 g/day (95), or 8–30 g of VLDL1 and 0.7–4.6 g of VLDL2 per day (96). The average production rate of VLDL apoB-100 was 14 mg/kg/day and 23 mg/day in the nonfasting state (97). Although the relation between apoB, TG, and SM secretion is not known, SM secretion is expected to exhibit a postprandial increase. The amount of SM secreted into blood as chylomicrons and VLDL should then be on the order of 0.5 g/day [i.e., approximately one-half plasma SM pool per day (24 h)]. It is expected to be higher in obese individuals having high VLDL secretion. Some SM is also secreted in discoidal HDL by liver and intestine (98). This secretion may involve the ABCA1 transporter, which was shown to be important by lipidating apoA-I in the liver (99). The amount of SM secreted into blood by these pathways is unknown, but it is likely to be smaller than the secretion in TG-rich lipoproteins.
Although SM is a component of nascent VLDL and chylomicrons, there is no evidence that it is essential for the secretory process. Hepatocytes secrete newly synthesized SM and ceramide in VLDL, but inhibition of SM and ceramide synthesis by the fungal toxin fumonisin did not inhibit VLDL secretion (100). The assembly of VLDL is initiated in the endoplasmic reticulum (ER), where apoB-100 has to be translocated to the lumen of the ER. ApoB-100 will interact with the lipids during this process, and the mature lipoprotein enters the secretory pathway when apoB-100 synthesis is completed and leaves the ribosome. The assembly process is dependent on the active formation of PC and is highly dependent on the rate of TG synthesis and on the action of the microsomal TG transfer protein (101). Of the enzymes necessary for SM biosynthesis, palmitoyl-CoA-serine transferase and ceramide synthetase are located primarily in the ER. In contrast, the SM synthetase catalyzing the transfer of phosphocholine from PC to ceramide occurs in the Golgi vesicles and in the plasma membrane, in different isoforms (102). An intracellular protein that may transport ceramide to the site of SM synthesis has been characterized (103, 104). However, the regulation of VLDL SM secretion is poorly characterized and may occur at different synthetic steps as well as during the intracellular transfer of SM and ceramide, and SM might even be formed from lipoprotein ceramide after secretion.
The chylomicron surface coat contains apoA-I as a predominant apolipoprotein. The output of phospholipids increases after fat feeding, the phospholipid/apoA-I ratio being significantly higher than in circulating HDL. It is not known, however, how the ingestion of fat and cholesterol specifically influences SM secretion in chyle lipoproteins. An optimal chylomicron secretion depends on the formation of PC by reacylation of absorbed lysophosphatidylcholine and de novo synthesis, which increases with fat feeding. The role of SM, unesterified cholesterol, and apoA-I in the secretion of chylomicrons is not well known. In chyle, the proportion of apoA-I found in chylomicrons increases with lipid feeding, but the secretion in denser lipoproteins, including HDL, continues during fasting (93). Even bile diversion, which depletes rat intestinal epithelial cells and mesenteric lymph of TG-rich lipoproteins, does not decrease mesenteric apoA-I and HDL output (105). The fact that the polar surface material of chylomicrons has a composition like that of the material released from many cell types by active ABCA1 transporter-mediated transfer raises the question of whether this transporter may contribute PC, SM, and unesterified cholesterol during chylomicron formation. No analysis is available on chyle lipoproteins from patients with Tangier disease, which have a nonfunctioning ABCA1 transporter, or from ABCA1–/– mice, but both have decreased plasma TG values, primarily because of the rapid clearance of apoB-containing lipoproteins (106, 107). The ABCA1 transporter is present in the Golgi membrane of intestinal epithelial cells and transfers lipids from Golgi structures to the plasma membrane (108). The hypothesis that this localization is linked to a role in determining chylomicron composition and possibly in hepatic VLDL formation (109) is intriguing but unproven. Redgrave (110) found that Golgi "prechylomicrons" contain fewer phospholipids than the secreted particles, indicating that some are indeed added at a late stage.
SM in HDL is derived from chylomicron and VLDL secretion but also from peripheral tissues. The SM of VLDL and chylomicrons contains 16:0 and 18:0 as predominant fatty acids and fewer 20–24 carbon fatty acids than tissue SM (26, 111). If SM in LDL is derived primarily from VLDL and/or chylomicrons, one would expect a similar fatty acid composition, which was found to be the case (112). In HDL, a difference was noticed between HDL2 and the denser HDL3, which contains less phospholipid and cholesteryl ester (CE) and more protein than HDL2. Whereas the fatty acid composition of SM in HDL2 is similar to that of LDL, HDL3 contains a larger proportion of 20–24 carbon fatty acid species. This suggests that HDL2 contains relatively more polar surface phospholipids from VLDL and chylomicrons and HDL3 relatively more tissue-derived SM, which contains more 20–24 carbon fatty acids. An alternative explanation is that there is a selective metabolism of different molecular species of HDL SM.
The fatty acid composition of plasma SM is rather stable. Myher et al. (113) fed diets rich in saturated and polyunsaturated fat to healthy young volunteers for 5 weeks. These conditions, which were sufficient to change the PC fatty acid pattern considerably, caused few changes in the fatty acid composition of SM.
In conclusion, chylomicron and VLDL secretion are major sources of plasma SM, although the mechanisms by which SM is inserted into these lipoproteins have not been established.
|
TRANSFER OF SM FROM CELLS TO LIPOPROTEINS |
|---|
|
|
In vitro, PC and SM but not phosphatidylethanolamine (PE) are exchanged between plasma lipoproteins and erythrocytes (118), and there is evidence also from other cell types that SM may be released to HDL by non-energy-dependent exchange/transfer, analogous to cholesterol (119) (Fig. 2). Partitioning of SM between lipoproteins and the erythrocytes by this pathway may be influenced by plasma lipoprotein concentration and composition. For example, in type II hypercholesterolemia, the SM/PC ratio of erythrocytes increases in parallel with an increase in the SM/PC ratio of the plasma lipoproteins. The rapid normalization after apheresis treatment of severe type II hyperlipidemia suggests a rather rapid and concentration-dependent partitioning equilibrium of SM between plasma lipoproteins and the erythrocyte membrane (120). Whether similar equilibria influence the SM level of other cell types is not known. For instance, it is not known whether the increased SM/PC ratio in erythrocytes and in the adipocyte plasma membrane, which is found in obese patients and which correlates to insulin resistance (121, 122), is linked to an increased transfer of SM from lipoproteins to cells. The exchange/transfer of SM between cells and HDL may occur mainly in rafts and/or the caveolar region and is not mediated by the ABCA1 transporter. The quantitative importance of this transfer in vivo is not known but may be large.
It is well known that apoA-I releases both cholesterol and SM from cells (123, 124), and there is compelling evidence that this transfer involves the ABCA1 transporter. Preß-migrating HDL, which is high in apoA-I and low in lipids, is a key acceptor during the active efflux of cholesterol and polar lipids, including SM (99, 125) (Fig. 2). An alternative cholesterol acceptor is the apoE-containing small SM-rich pre
-migrating HDL (92). The ABCA1 transporter, which is known to be important in cholesterol efflux from cells to lipoproteins, also transfers polar lipids, including SM, to apoA-I, occurring in monomolecular form or as poorly lipidated oliogomeric aggregates (125–127) (Fig. 2). The transfer of SM is not selective, however, because the proportion of SM in lipoprotein complexes formed by ABCA1 activity (128) is generally lower than that of PC. The proportion of SM in the released material is also lower than in the acceptor lipoproteins [i.e., preß-migrating HDL (129) and -lipoprotein E, which is rich in SM and contains apoE as the only apolipoprotein (130, 131)]. In some cell types, such as J774 macrophages, material released via ABCA1 activation contains a rather high proportion of SM among polar lipids (132).
Because loading of cells with cholesterol is known to increase ABCA1-associated cholesterol release to HDL, the question has been raised whether it also increases the proportion of SM in the released material. Bielicki, McCall, and Forte (133) showed that the differentiation of monocytes to macrophages increased SM content in the nascent HDL produced, but this did not increase further in foam cells produced by loading macrophages with acetylated LDL. Gillotte-Taylor et al. (134) enriched fibroblasts with cholesterol by incubating the cells with LDL or lipid/cholesterol dispersions. A doubling of the plasma membrane cholesterol was associated with a 6-fold increase in ABCA1-mediated efflux to apoA-I of both cholesterol and phospholipids, but the proportion of SM in the released material was decreased in cholesterol-loaded cells, and cholesterol loading increased PC rather than SM synthesis. Cholesterol excess in fibroblasts thus increased ABCA1-mediated SM efflux, but less than the cholesterol and PC efflux. Cholesterol loading of J744 macrophages increased the ABCA1-mediated cholesterol efflux and the cholesterol-phospholipid ratio of the released material severalfold but did not increase the proportion of SM in polar lipids (132).
The ABCA1 transporter is located in nonraft regions of the cells, and its function involves endocytosis and the resecretion of apoA-I (135). When lipidating an extracellular apolipoprotein acceptor, the transporter was shown to migrate between late endocytic vesicles and the cell surface, indicating that these vesicles play a role in cellular lipid efflux. In Tangier disease cells, both cholesterol and SM are accumulated, demonstrating a role of the ABCA1 transporter in the transport of SM. This model may explain why a deficient acid SMase decreases HDL (136), probably as a result of a defect of lipid efflux secondary to the inhibition of SM hydrolysis by acid lysosomal SMase (137, 138). It also supports the conclusion, drawn earlier by Mendez et al. (139), that cholesterol efflux to lipidated HDL particles but not the ABCA1-dependent efflux to lipid-free apoA-I involves the rafts. This was based on fibroblast studies showing that cholesterol efflux to HDL or plasma but not the ABCA1-mediated cholesterol efflux to apoA-I was associated with a reduction in raft cholesterol. Liu et al. (132), using J774 macrophages, demonstrated the apoA-I-induced release of apoA-I-containing particles, which were 9 and 12 nm compared with 6 nm for the lipid-free apoA-I and with a PC/SM ratio of 2:1. Furthermore, larger particles likely to be membrane vesicles, with a PC/SM ratio of 1:1, were formed. The conclusion was that the lipids of the apoA-I-containing particles were derived from nonraft regions, whereas the vesicles may be formed from cholesterol- and SM-rich lipid rafts. If so, ABCA1 may mediate the efflux of SM-rich lipid material in macrophages by two different mechanisms.
When another transporter, ABCA7, which is highly homologous to ABCA1, was overexpressed in HEK 293 cells, the apoA-I-mediated efflux of PC and SM, but not that of cholesterol, was increased (140).
In conclusion, SM may be released to HDL via the ABCA1 and possibly the ABCA7 transporter, but in much smaller amounts than PC. The quantity of SM supplied to HDL via these pathways is probably much less than the amount secreted in TG-rich lipoproteins and is weakly related to the reverse cholesterol transport. Little is known about whether a net transfer of SM from tissues to LDL, IDL, and/or VLDL occurs and whether HDL acceptors forward SM to LDL (115).
|
CAUSES OF HYPERSPHINGOMYELINEMIA |
|---|
SM levels are also increased in hyperlipidemia animal models induced by diet or genetic defect. In the cholesterol-fed rabbit and in the hyperlipidemic LDL receptor-deficient WHHL rabbit, severalfold increases in plasma SM in lipoproteins are seen (145–147). In the apoE–/– knockout mouse, SM-rich lipoproteins, primarily cholesterol- and TG-rich remnants, accumulate (7). Whereas the WHHL rabbit and the apoE–/– mouse are characterized by a slow lipoprotein clearance attributable to a defective lipoprotein receptor-lipoprotein interaction, the SMHL rabbit is characterized by increased VLDL production. This rabbit also develops a moderate increase in plasma SM (148).
In chicken treated with ethinylestradiol, which drastically increases the secretion of VLDL TG, the plasma SM increases in parallel with TG (149), emphasizing that a high rate of VLDL secretion is an alternative cause of a high plasma SM. Because in this animal model VLDL that is relatively resistant to LPL and specifically targeted to the egg is formed, a decreased clearance rate of VLDL may also contribute to the hypersphingomyelinemia (150).
Suckling neonates of all mammalian species ingest large amounts of milk lipids at frequent intervals. Suckling rats have 10-fold higher levels of LDL than normal adult rats and also high levels of lipoproteins in the LDL-VLDL area containing relatively large amounts of SM (111). During the neonatal phase, the rat expresses LPL in liver but has little adipose tissue at birth. Fatty acids are thus canalized to the liver for ketogenesis and VLDL production during postnatal growth, when adipogenesis is prominent in the rat (151). The findings in suckling rats must reflect both a high rate of chylomicron formation and an extensive recirculation of fatty acids as VLDLs.
Schlitt et al. (152) recently reported a postprandial increase in SM after ingestion of a lipid-rich meal (102 g, of which 52 g was saturated fat and 300 mg of cholesterol) that correlated to the postprandial increase in TG. Part of the increase in SM was in chylomicron remnant particles. Thus, also in adult humans, a large load of milk fat can induce postprandial hypersphingomyelinemia as a consequence of a high rate of chylomicron and VLDL production.
The consequences of a high-fat load for SM metabolism are summarized in Fig. 3. In this situation, increased amounts of chylomicrons and VLDLs and their remnants will compete for LPL and for apoB/E receptor- or low density lipoprotein receptor-related protein (LRP)-mediated elimination from plasma. As a consequence, the residence time of both types of intermediary TG-rich lipoproteins will increase. If one postulates that the fractional disappearance rate of SM in the TG-rich particles is lower than for PC, the SM/PC ratio will increase with the time needed for the particles to undergo lipolysis by LPL, clearance by receptors as remnants, or conversion to LDL. SM in plasma may be linked to the rate of VLDL production, which may explain why in the study by Schlitt et al. (152) the clearance of postprandial remnant-like particles was related to serum SM level. Using this model, one expects plasma SM to increase in situations with an exaggerated postprandial lipemic response as combined hyperlipidemia (153) and obesity (154). The number of VLDL particles clearly exceeds the number of chylomicrons, even during the postprandial phase, but still chylomicrons strongly inhibit VDL clearance, possibly as a result of their high affinity for LPL (155). Therefore, the effect on SM accumulation may be strongest in VLDL remnants. This model for the influence of large dietary fat loads on plasma SM does not need to postulate that the inhibitory action of SM on LPL and on apoE-mediated remnant clearance, described below and in Table 2, occurs in vivo. If it does, it will influence the clearance of those particles that have the lowest affinity for LPL and for apoE-mediated clearance most markedly.
|
|
METABOLISM OF PLASMA SM |
|---|
-migrating lipoprotein E increased, suggesting that these particles may be formed from surface material during lipolysis (167). Most chylomicron remnants are rapidly catabolized, primarily by the apoE-mediated action of the apoB/E receptor and the LRP receptor in the liver (168). In the case of VLDL, which contains both apoE and apoB when secreted in nascent form, some remnants formed during lipase and PLTP action are cleared and some are transformed to IDL and finally LDL with a slow turnover (97, 169). What regulates the SM content of these SM-rich atherogenic lipoproteins is important but not known. In most species, detailed information on the SM/apoB ratio during the transformation of VLDL to IDL and LDL is lacking. Thus, it is unknown whether in humans SM is enriched by the continuous lipolysis of other lipids only or by the net transfer of SM during the exchange with other lipoproteins as well. Interestingly, recent data on the phospholipid/apoB ratio and the percentage of SM in polar lipids in VLDL, IDL, and LDL (148) in the WHHL, the SMHL, and the normal New Zeeland White rabbit indicate that the SM/apoB ratios in VLDL, IDL, and LDL continuously decrease during the transformation of VLDL to IDL and LDL. The ratio in VLDL and LDL decreased from 0.57 to 0.06 in the SMHL rabbit, from 0.13 to 0.08 in normal rabbits, and from 1.07 to 0.43 in the WHHL rabbit. Thus, there was no evidence of a net acquisition of SM from LDL but a continuous loss of SM. As a consequence, any receptor-mediated catabolism of intermediary particles during the VLDL-LDL interconversion will eliminate relatively more SM per lipoprotein particle than the LDL catabolism.
SM is transferred between different lipoprotein classes
Early studies demonstrated an exchange of SM between VLDL, LDL, and HDL (170). This bidirectional transfer may be both spontaneous and mediated by PLTP (92). Rao et al. (166) studied the molecular and macromolecular specificity of human PLTP and found that the most rapid transfer of phospholipids occurred between different HDL subfractions. The spontaneous transfer of SM was faster than that of PC, and the PLTP-catalyzed transfer was slower. In studies of the action of PLTP and HL on HDL2, denser SM-rich particles were formed as a minority of the particles (171). Therefore, one can postulate that during the continuous transformation of TG-rich lipoproteins to remnants and LDL, there is a partitioning of SM that depends on available SM mass as well as LDL, HDL, and remnant composition. The factors that regulate this partitioning in vivo need to be characterized.
|
METABOLISM OF SM IN LDL AND HDL |
|---|
Bentejac et al. (174) incorporated [14C]stearate- plus [3H]choline-labeled SM into HDL in vitro and studied the fate of the doubly labeled lipoprotein SM in vivo in rats. The decay of plasma radioactivity followed a triexponential curve with one fast component, one intermediary with a half-life of 100 min, and a slow component. Tissue radioactivity was examined after 24 h and was greatest per gram of tissue in liver and adrenals. Studies from the same group also compared the fate of 18:0 and 24:0 SM and found only small differences (175, 176). Stein et al. (177) injected unilamellar liposomes of SM with defined fatty acids and labeled the 18:0, 18:1, or 18:2 fatty acid intravenously and found evidence for some reutilization in the liver of palmitoyl-sphingosine but not of linoleoyl-sphingosine for SM synthesis. This conclusion was supported in experiments with [3H]sphingosine-, [14C]palmitic acid-, and [14C]linoleic acid-labeled SM in vivo and in HepG2 cells (178).
The conclusion from these studies is that SM is cleared from plasma faster than apoA-I of HDL. These results are compatible with some uptake by the SR-BI receptor but at a slower rate than HDL CE. The ability of SR-BI to mediate the cellular uptake of SM from HDL has actually been demonstrated in COS-7 cells (179) and monocytes (180) using LDL, HDL, and lipid vesicles as carriers of pyrene-labeled phospholipids. Overexpression of SR-BI in BHK cells was found to increase the uptake of LDL SM most. Interestingly, overexpression of SR-BI in apoE–/– mice, which accumulate SM-rich lipoproteins, did not improve the defective lipoprotein clearance of apoB-containing lipoproteins in these animals (181). The conclusion is that SR-BI participates in the clearance of SM from HDL and possibly also from LDL, but the uptake is less efficient than the SR-BI-mediated uptake of CE.
Studies in cell cultures indicate that LDL SM can be degraded after uptake of the lipoprotein particle via the LDL receptor (182). SMase treatment of LDL caused an enhanced uptake by cultured cells, at least in part mediated by the LDL receptor (9). Evidence was provided that in lymphoblastoid cells, LDL SM could be processed by both an endosomal pathway involving lysosomal SMase and an alternative pathway by which LDL SM was degraded by neutral SMase (183). Interestingly, the uptake of LDL by fibroblast cultures was also found to be linked to the feedback suppression of sphinganine biosynthesis (184).
Studies of the clearance of HDL and LDL SM in humans are lacking. An approximate calculation from known data on apoA-I turnover in humans (0.20–0.25 pools/d) and a total plasma HDL SM pool of 0.3–0.4 g indicates that one-third to one-fourth of the HDL SM can be eliminated by particulate metabolism of HDL. A similar calculation for LDL SM, postulating an average apoB turnover of 0.46 pools/day (97) and a total plasma LDL SM pool of 0.7–1 g, indicates that less than half of the LDL SM could be cleared from plasma via the elimination of LDL particles, including apoB.
Thus, particulate degradation of LDL and HDL may account for a part but not all of the plasma SM turnover. Both the SR-BI receptor and lipoprotein receptor-mediated uptake of chylomicron remnant particles and VLDL remnants are likely to contribute, but the quantitative importance of these pathways needs to be defined. Clearly, the fractional turnover of HDL SM is lower than for HDL-PC, the SM/PC ratio being higher in circulating HDL than in polar lipids received from chylomicrons, VLDL, and the ABCA1 transporter.
|
FUNCTIONS OF SM IN LIPOPROTEINS |
|---|
SM decreases the association of apoE with remnants
In apoE–/– mice, SM-rich remnant particles accumulate (7). Because apoE has a key role in the rapid removal of remnant particles, the question has been asked how SM and apoE interact under normal conditions. Studies by Arimoto et al. (191) using artificial lipid emulsions with different SM content suggested that SM may delay remnant clearance by decreasing the binding of apoE, thereby decreasing interaction with the LRP and the apoB/E receptor. In artificial lipid emulsions of TG-PC-SM, apoE was found to prevent particle aggregation induced by SMase. Furthermore, particles consisting of triolein and a polar lipid mixture of PC/SM (2:1) were found to bind apoE less efficiently than an emulsion containing PC as the only polar lipid (192). The presence of ceramide in the emulsion with PC/SM (2:1) increased the binding. The incorporation of SM in the emulsion decreased the capacity for binding of apoE without changing affinity. ApoC-II and apoC-III inhibited the apoE-mediated cellular (HepG2) uptake of the particles more efficiently if SM was present in the particles (193).
Thus, there is evidence that SM inhibits both the action of LPL and the interaction of TG-rich particles with apoE. Although the in vivo relevance of these findings is not fully established, they may be highly relevant during the SM increase after a period of large lipid load (152).
SM inhibits LCAT but not cholesteryl ester transfer protein action
During reverse cholesterol transport, cholesterol is transferred from cells to HDL, where CE is formed by LCAT and metabolized primarily by the action of the SR-BI receptor in liver and steroidogenic tissues (194, 195) or transferred by the action of cholesteryl ester transfer protein (CETP) to LDL and TG-rich lipoproteins in exchange for TG. A number of studies indicate that this procedure is affected by SM in HDL. SM may inhibit LCAT by decreasing enzyme binding (196). A negative correlation between SM content and LCAT activity was observed in studies with proteoliposomes or reconstituted HDL (197). Rye, Hime, and Barter (92) found that SM influences the structure of discoidal and spherical HDL and confirmed that SM inhibits the LCAT reaction but not CETP. The metabolism of lipoprotein SM may in turn be influenced by LCAT, because inhibition of LCAT minimized SM exchange among lipoproteins (170). ApoA-ISeattle, which is linked to a decrease in LCAT activity, recruited lipids with a higher proportion of SM and PE (198). In children with kwashiorkor, a decrease in LCAT action could be related to changes in HDL composition, including an increased proportion of SM (199).
Thus, there is no evidence that SM affects CETP function (92), and CETP and PLTP have no overlapping function in vivo. In PLTP knockout mice that were made transgenic for CETP, the phospholipid transfer was still zero (200). Thus, if more SM is partitioned to CE-rich particles in hypercholesterolemia, this has to occur by other mechanisms than a CETP-mediated cotransport with CE. Nor is there any evidence that SM influences the SR-BI-mediated uptake of CE. Generally, the uptake of CE in the liver is more effective for the largest HDL particles, which should indicate that it is less efficient for small SM-rich particles (201). A key function of plasma SM may be in retaining HDL structure during the transformation of HDL, which regenerates SM-rich subfractions of particles that are optimized as acceptors for cholesterol (171).
In summary (Table 2), the inhibitory effects of SM on LPL, LCAT, and apoE binding to lipoproteins are well established, but there is still uncertainty about their physiological importance in vivo in humans. After a large fat load, when TG-rich lipoproteins compete for the LPL- and apoE-mediated catabolism, an SM-mediated inhibitory effect on these catabolic pathways may be important both for the clearance of the most TG-rich lipoproteins and for the partitioning of liver-derived VLDL remnants between rapid apoE-mediated remnant catabolism and the formation of LDL with a slow turnover.
|
EFFECTS OF DIET AND AGE ON SM TRANSPORT |
|---|
TOP
ABSTRACT
INTRODUCTION
