Potential conflict of interest: Nothing to report.
Scavenger receptor class B type I (SR-BI) mediates selective uptake of cholesterol from high-density lipoprotein (HDL) particles by the liver and influences biliary cholesterol secretion. However, it is not clear, if this effect is direct or indirect. The aim of this study was to determine the impact of SR-BI on biliary cholesterol secretion, especially in a functional context with ATP-binding cassette transporter g5 (Abcg5)/Abcg8 and Abcb4. SR-BI was overexpressed by means of adenovirus (AdSR-BI) in livers of wild-type, liver X receptor–null (Lxr−/−), Abcg5−/−, and Abcb4−/− mice. Consistent with previous reports, AdSR-BI decreased plasma HDL cholesterol levels in all models (P < 0.001). Hepatic cholesterol content increased (at least P < 0.05), whereas expression of sterol regulatory element binding protein 2 target genes was decreased (at least P < 0.05,) and established Lxr target genes were unaltered. Biliary cholesterol secretion was increased by AdSR-BI in wild-type as well as in Lxr−/− and Abcg5−/− mice, and considerably less in Abcb4−/− mice (each P < 0.001), independent of bile acid and phospholipid secretion. Immunogold electron microscopy and western blot showed a substantial increase of SR-BI protein localized to basolateral and canalicular membranes in response to SR-BI overexpression. Subcellular fractionation revealed a significantly higher cholesterol content of canalicular membranes (P < 0.001) upon SR-BI overexpression. Inhibition of microtubule function did not affect SR-BI–mediated biliary cholesterol secretion, indicating that transcytosis pathways are not involved. Conclusion: Our data indicate that SR-BI mediates biliary cholesterol secretion independent of Abcg5, yet largely depends on Abcb4-mediated phospholipid secretion and mixed micelles as acceptors in bile. SR-BI–mediated biliary cholesterol secretion has a high capacity, can compensate for the absence of Abcg5, and does not require transcytosis pathways. (HEPATOLOGY 2009.)
The scavenger receptor class B type I (SR-BI) has been characterized as a receptor that mediates cholesterol transport across membranes.1, 2 In nonpolarized cells, namely macrophages and the hepatoma cell line Fu5AH, SR-BI expression either results in selective uptake of cholesterol, mainly from high-density lipoprotein (HDL), or in cholesterol efflux toward suitable acceptors.3–6
In hepatocytes, which is a highly polarized cell type, SR-BI is the main receptor responsible for selective uptake of cholesterol from plasma HDL.7 Consequently, hepatic overexpression of SR-BI results in decreased plasma HDL cholesterol levels,8–10 whereas SR-BI knockout mice have increased plasma HDL cholesterol.11, 12 Interestingly, hepatocyte SR-BI appears to accelerate reverse cholesterol transport in vivo in the face of decreased plasma HDL cholesterol levels,13 which is in line with studies demonstrating that hepatic SR-BI expression protects against atherosclerosis development in mouse models.14, 15 Hepatic SR-BI expression is also linked to biliary cholesterol secretion in a process that is less well understood than selective uptake. Hepatic overexpression of SR-BI has been associated with increased cholesterol secretion into bile,8, 9 whereas biliary cholesterol secretion in SR-BI knockout mice is decreased.16 However, it is not clear if SR-BI modifies intracellular cholesterol fluxes via increased selective uptake to promote biliary cholesterol secretion or if the SR-BI protein under certain conditions directly contributes to biliary cholesterol secretion in its role as an efflux transporter.
Several transport systems have been shown to contribute to biliary cholesterol secretion. It is generally accepted that phospholipid secretion mediated by ATP-binding cassette transporter b4 (Abcb4) is required for cholesterol to be secreted into bile as evidenced by the phenotype of the Abcb4 knockout mouse model, in which hepatic cholesterol secretion is virtually absent.17 Regarding direct cholesterol secretion under steady-state conditions, the quantitatively highest contribution comes from the heterodimer Abcg5/Abcg8: Abcg5/Abcg8 knockout mice have a significant reduction in biliary cholesterol secretion,18 whereas Abcg5/Abcg8 overexpression results in elevated biliary cholesterol levels.19 However, in the absence of Abcb4, even high-level overexpression of Abcg5/Abcg8 does not increase biliary cholesterol secretion,20 delineating the crucial role for biliary phospholipids in mixed micelles with bile salts as acceptors for cholesterol solubilization. Importantly, an Abcg5/Abcg8-independent cholesterol secretion pathway has been described;21–23 however, the mediator of this process remains unknown.
The aim of this study was to investigate the potential contribution of SR-BI to biliary cholesterol secretion in vivo, especially in the functional context with Abcg5/Abcg8 and Abcb4. Our results demonstrate that hepatic overexpression of SR-BI increases biliary cholesterol secretion independent of Abcg5/Abcg8. The data generated are consistent with a model in which SR-BI mediates cholesterol transport into bile in a process that is functionally able to largely compensate for the loss of canalicular cholesterol translocation mediated by Abcg5/Abcg8, but requires mixed micelles as acceptors.
Abcb4, ATP-binding cassette transporter b4; Abcg5/Abcg8, ATP-binding cassette transporter g5/g8; Apo, apolipoprotein; FPLC, fast protein liquid chromatography; HDL, high-density lipoprotein; HMGCoA, 3-hydroxy-3-methyl-glutaryl-coenzyme A; Lxr, liver X receptor; Ldlr, low-density lipoprotein receptor; mRNA, messenger RNA; sPLA2, type IIA secretory phospholipase A2; Srebp2, sterol regulatory element binding protein 2; SR-BI, scavenger receptor class B type I.
Materials and Methods
For a detailed description of materials and methods used, please see Supporting Information. Please note that all four genotypes of mice used in this study are on different genetic backgrounds. Therefore, although internal comparisons between mice administered control adenovirus (AdNull) and mice administered adenovirus expressing SR-BI (AdSR-BI) within a given strain are rigorously controlled for, intergenotype comparisons have certain limitations and caution has to be exerted when comparing these.
Hepatic SR-BI Expression Results in Decreased Plasma HDL Cholesterol Levels and Increased Biliary Cholesterol Secretion in Wild-Type Mice.
In wild-type C57BL/6J mice, hepatic overexpression of SR-BI resulted in significantly decreased plasma levels of total cholesterol (73 ± 9 mg/dL versus 19 ± 10 mg/dL, P < 0.001; Fig. 1A), mainly due to a decrease in HDL cholesterol as revealed by fast-protein liquid chromatography (FPLC) analysis (Fig. 1B). Plasma phospholipids were also significantly decreased (P < 0.001), whereas plasma triglycerides remained unchanged.
Because hepatic SR-BI mediates selective uptake of cholesterol into the liver, we next assessed whether SR-BI overexpression would translate into changes in hepatic lipid composition. Body weight was identical in the groups injected with AdSR-BI and AdNull (data not shown); however, hepatic SR-BI overexpression resulted in a significant increase in liver weight (P < 0.001; Fig. 1C, Supporting Table 2). Expressed per gram of liver, triglycerides, phospholipids (data not shown), and free cholesterol (Fig. 1D) remained unchanged, whereas liver total (P < 0.05) and esterified cholesterol (P < 0.05; Fig. 1D) increased in response to SR-BI overexpression. Expressed per whole organ, hepatic cholesterol content was significantly increased in the AdSR-BI–injected mice due to the increase in liver weight (P < 0.001; Fig. 1E). Hepatic SR-BI overexpression did not affect messenger RNA (mRNA) expression levels of Abca1, Abcg5/Abcg8, Abcb4, and Abcb11 (Table 1). However, mRNA expression of the sterol regulatory element binding protein 2 (Srebp2) target genes low-density lipoprotein receptor (Ldlr) and 3-hydroxy-3-methyl-glutaryl-coenzyme A (HMGCoA) reductase was significantly decreased (each P < 0.001; Table 1), and consistent with these data, Ldlr protein expression was also decreased (Supporting Fig. 4).
Table 1. Hepatic Gene Expression Levels in Response to SR-BI Overexpression
AdNull (n = 6)
AdSR-BI (n = 6)
AdNull (n = 6)
AdSR-BI (n = 6)
AdNull (n = 6)
AdSR-BI (n = 6)
AdNull (n = 6)
AdSR-BI (n = 6)
Values are means ± SEM. n.d. = not detectable.
Significantly different from the respective AdNull-injected control groups: at least
Liver SR-BI overexpression has been shown to result in increased cholesterol concentrations in gallbladder bile.8 Therefore, we performed continuous bile cannulation to assess biliary sterol secretion rates (Table 2; for values normalized to liver weight, see Supporting Table 3). Bile flow was significantly increased in response to SR-BI overexpression (P < 0.001). Biliary secretion of bile acids remained unchanged in mice injected with AdSR-BI, whereas biliary phospholipid (P < 0.001) as well as cholesterol secretion (P < 0.001) was significantly increased. The ratio of biliary cholesterol to phospholipid secretion was also significantly higher in mice overexpressing SR-BI (P < 0.001), indicating a phospholipid-independent effect of SR-BI on biliary cholesterol secretion. SR-BI overexpression did not affect the biliary content of apolipoprotein A-I (apoA-I) and apoE (Supporting Fig. 3A,B).
Table 2. Biliary Secretion Rates in Response to SR-BI Overexpression
AdNull (n = 12)
AdSR-BI (n = 12)
AdNull (n = 6)
AdSR-BI (n = 6)
AdNull (n = 5)
AdSR-B I(n = 6)
AdNull (n = 6)
AdSR-BI (n = 6)
Bile flow is given in microliters per minute per 100 g body weight. Biliary secretion rates are given in nanomoles per minute per 100 g body weight. Values are means ± SEM. BA, bile acids; n.d., not detectable; PL, phospholipids
Significantly different from the respective AdNull-injected control groups: at least
Hepatic SR-BI Increases Biliary Cholesterol Secretion Independent of Lxrα.
Hepatic SR-BI expression has been linked to Lxr activation and Abcg5/Abcg8 expression,24 which could conceivably represent an underlying mechanism for the observed effect of SR-BI on biliary cholesterol secretion in wild-type mice. Although, however, in our study hepatic Abcg5/Abcg8 expression was not induced in wild-type mice by hepatic SR-BI overexpression, we tested the concept that SR-BI indirectly mediates biliary cholesterol secretion in an Lxr-dependent fashion by overexpressing SR-BI in the liver of Lxrα knockout mice on a mixed genetic background. Injection of AdSR-BI into Lxrα knockout mice resulted in decreased plasma total cholesterol levels (75 ± 11 mg/dL versus 17 ± 10 mg/dL, P < 0.001; Fig. 2A), mainly due to decreased plasma HDL cholesterol as revealed by FPLC analysis (Fig. 2B).
Although body weight was not different between the experimental groups, liver weight increased significantly in response to SR-BI overexpression (P < 0.001; Fig. 2C, Supporting Table 2). Expressed per gram of liver, triglycerides, phospholipids (data not shown), cholesteryl ester, and free cholesterol remained unchanged, whereas total hepatic cholesterol content was significantly increased (P < 0.05; Fig. 2D). Expressed per whole organ, hepatic cholesterol content was significantly increased (P < 0.01; Fig. 2E). SR-BI overexpression in Lxrα knockout mice did not result in changes in the hepatic expression level of Abca1, Abcg5/Abcg8, Abcb4, or Abcb11 (Table 1). However, comparable to the results in wild-type mice, expression of the Ldlr and HMGCoA reductase was decreased by about 50% (each P < 0.001; Table 1).
Bile flow was not influenced by SR-BI expression in Lxrα knockout mice, and the biliary secretion rates of bile acids as well as, in contrast to the results in C57BL/6J mice, phospholipids did not change (Table 2, Supporting Table 3). However, biliary secretion of cholesterol was significantly increased by hepatic SR-BI overexpression (P < 0.001) and also the secretion ratio of biliary cholesterol to phospholipid was higher (P < 0.001), consistent with the results in wild-type mice.
These data demonstrate that the SR-BI–mediated increase in biliary cholesterol secretion is not mediated by an Lxrα-dependent induction of canalicular cholesterol transporters.
Hepatic SR-BI Expression Normalizes the Biliary Cholesterol Secretion Deficit of Abcg5 Knockout Mice.
Thus far, the Abcg5/Abcg8 obligate heterodimer is considered the rate-controlling transporter system mediating biliary cholesterol secretion.18, 19 To investigate whether SR-BI would work in concert with or independent from Abcg5/Abcg8, Abcg5 knockout mice on a mixed genetic background were injected either with the SR-BI–expressing adenovirus or the control adenovirus AdNull. SR-BI overexpression resulted in decreased plasma cholesterol levels (84 ± 2 mg/dL versus 10 ± 8 mg/dL, P < 0.001; Fig. 3A), mainly because of decreased HDL cholesterol (Fig. 3B).
Liver weight increased in response to SR-BI overexpression (P < 0.001; Fig. 3C, Supporting Table 2). Expressed per gram of liver, hepatic phospholipid, triglyceride, and cholesterol content was not different between the experimental groups (Fig. 3D), whereas expressed per total liver, hepatic cholesterol content was increased (P < 0.01; Fig. 3E). SR-BI overexpression in Abcg5 knockout mice left mRNA levels of Abca1, Abcb4, and Abcb11 unchanged (Table 1), whereas the expression of Abcg8 (P < 0.05; Table 1) as well as Ldlr (P < 0.001; Table 1) and HMGCoA reductase (P < 0.01; Table 1) was significantly decreased.
SR-BI overexpression affected neither bile flow in Abcg5 knockout mice nor biliary secretion of bile acids and phospholipids (Table 2, Supporting Table 3). However, hepatic SR-BI expression resulted in a significant increase in biliary cholesterol secretion (P < 0.001), thereby restoring the reduced cholesterol secretion of the Abcg5 knockout mouse to levels observed in wild-type controls injected with the control adenovirus AdNull. The secretion ratio of biliary cholesterol to phospholipid was again significantly higher in mice with hepatic SR-BI overexpression (P < 0.001).
These data indicate that SR-BI expression has the potential to increase biliary cholesterol secretion in a quantitatively relevant fashion largely independent of the Abcg5/Abcg8 heterodimer.
Hepatic SR-BI Expression Increases Biliary Cholesterol Secretion Considerably Less in the Absence of Biliary Phospholipid Secretion in Abcb4 Knockout Mice.
Abcb4 knockout mice virtually lack biliary cholesterol secretion secondary to their inability to secrete phospholipids into bile. The Abcg5/Abcg8-mediated biliary cholesterol secretion has been demonstrated to be dependent on functional Abcb4, even under conditions with high-level hepatocyte Abcg5/Abcg8 overexpression.20 To investigate whether the SR-BI effect on biliary cholesterol secretion would also depend on Abcb4 expression, SR-BI was overexpressed in livers of Abcb4 knockout mice on a Friend virus B-type genetic (FVB) background. Hepatic SR-BI overexpression resulted in a significant decrease in plasma total cholesterol levels (43 ± 2 mg/dL versus 13 ± 1 mg/dL, P < 0.001; Fig. 4A), due to a reduction in HDL cholesterol (Fig. 4B).
Consistent with the effect in the other mouse models, SR-BI overexpression in Abcb4 knockout mice resulted in increased liver weight (P < 0.01; Fig. 4C, Supporting Table 2). Expressed per gram of liver, hepatic phospholipid, triglyceride, and free cholesterol content did not change, whereas hepatic total (P < 0.001; Fig. 4D) and esterified (P < 0.01) cholesterol content increased in response to SR-BI overexpression. Expressed per total liver, cholesterol content was significantly higher in the AdSR-BI injected group (P < 0.001; Fig. 4E). In response to hepatic SR-BI overexpression, mRNA levels of Abca1, Abcg5/Abcg8, Abcb4, and Abcb11 remained unchanged (Table 1), and although comparable with the data obtained in the other mouse models investigated, expression of Ldlr and HMGCoA reductase were decreased by about 50% (each P < 0.001; Table 1).
Bile flow was not different between control and SR-BI–overexpressing Abcb4 knockout mice (Table 2, Supporting Table 3). Biliary bile acid secretion also did not differ between controls and AdSR-BI–injected mice, whereas biliary phospholipid secretion was absent in the Abcb4 knockout model. The low biliary cholesterol secretion in Abcb4 knockouts was increased by almost five-fold in response to hepatic SR-BI expression (P < 0.001). However, the maximum quantitative effect was still considerably less than in the other mouse models investigated.
These data demonstrate that hepatic SR-BI increases biliary cholesterol secretion to a certain extent in the absence of phospholipid secretion in Abcb4 knockout mice, but is largely dependent on the presence of mixed micelles in bile as cholesterol acceptors.
Hepatic SR-BI Overexpression Results in Increased SR-BI Localization to the Canalicular Membrane.
To gain further mechanistic insights into SR-BI–mediated cholesterol transport into bile, we first sought to localize SR-BI within the hepatocyte by electron microscopy using immunogold labeling. In AdNull-injected mice, SR-BI was detectable at the basolateral (Fig. 5A) as well as the canalicular membrane (Fig. 5B) of hepatocytes. Expression at both localizations increased in response to SR-BI overexpression (Fig. 5C,D). Related to the length of the membrane, SR-BI expression at the basolateral membrane was not different from the canalicular membrane in AdNull-injected mice (1.4 ± 0.3 dots/μm versus 1.2 ± 0.2 dots/μm, n.s.; Fig. 5E). After administration of AdSR-BI, expression of SR-BI increased significantly by 3.9-fold at the basolateral (5.5 ± 0.6 dots/μm, P < 0.001; Fig. 5E) and by 6.8-fold at the canalicular membrane (8.1±0.8 dots/μm, P < 0.001; Fig. 5E), and was significantly higher at the canalicular compared with the basolateral localization (P < 0.01; Fig. 5E). To further substantiate these data, western blots were performed. SR-BI protein localized to plasma membranes increased substantially following overexpression, and SR-BI expression within canalicular membranes was especially elevated (Fig. 5F). Similar results were obtained for the other mouse models investigated (data not shown). These data demonstrate that coinciding with an increase in biliary cholesterol secretion in response to hepatic SR-BI overexpression the localization of SR-BI protein to the canalicular membrane was increased.
The Cholesterol Content of Canalicular Membranes Is Increased in Response to Hepatic SR-BI Overexpression.
Because SR-BI is able to mediate bidirectional flux of cholesterol in nonpolarized cells,5 a prerequisite for increased biliary cholesterol secretion would be the availability of sufficient substrate. Therefore, we next examined the change in cholesterol content of the basolateral as well as the canalicular membranes in response to SR-BI overexpression. In relation to phospholipid content, the cholesterol content of the basolateral membrane fraction did not increase following injection of AdSR-BI into wild-type mice (0.31 ± 0.06 versus 0.38 ± 0.08; Fig. 6A). On the other hand, cholesterol content of the canalicular membrane fraction was higher compared with the basolateral membrane fraction and increased almost two-fold in response to SR-BI overexpression (0.57 ± 0.06 versus 0.99 ± 0.13, P < 0.001; Fig. 6A). Also when different methods of normalization were employed, cholesterol enrichment of canalicular membranes remained substantial in response to SR-BI overexpression, i.e., when related to protein content (0.44 ± 0.05 μmol/mg protein versus 0.66 ± 0.07 μmol/mg protein, P < 0.01; Fig. 6B) or to alkaline phosphatase (AP) activity (0.12 ± 0.01 μmol/nmol/hour AP activity versus 0.32 ± 0.04 μmol/nmol/hour AP activity, P < 0.001; Fig. 6C). These data demonstrate that SR-BI overexpression results in cholesterol enrichment of canalicular membranes.
Increased Biliary Cholesterol Secretion in Response to Hepatic SR-BI Overexpression Is not Dependent on Microtubular Function.
We next investigated whether transcytosis pathways are involved in the SR-BI–mediated increase in biliary cholesterol secretion in vivo. AdSR-BI–administered mice were injected with either vehicle or colchicine to block microtubule function prior to bile cannulation for 90 minutes divided into three periods of 30 minutes each. During the entire cannulation period of 90 minutes, biliary secretion of bile acids, phospholipids, and cholesterol did not change significantly, and there was also no significant difference between mice that received colchicine and controls for each of these parameters at any time point (Supporting Table 4). After 90 minutes, livers were dissected out and efficient depolymerization of hepatic microtubules was confirmed by immunohistochemistry (data not shown). In addition, we ensured that colchicine treatment blocked transcytosis by demonstrating biliary secretion of horseradish peroxidase injected into the portal vein in mice administered phosphate-buffered saline, but not in colchicine-treated mice (Supporting Fig. 2A,B). These results indicate that transcytosis pathways might not be involved in the SR-BI–mediated increase in biliary cholesterol secretion.
The results of this study are consistent with a model that SR-BI has the capability to mediate bidirectional cholesterol transport in hepatocytes and might contribute to the previously identified Abcg5/Abcg8-independent fraction of biliary cholesterol excretion.
Previously, a link between hepatic SR-BI expression and biliary cholesterol levels and secretion rates has been described. SR-BI knockout mice show a 45% decrease in biliary cholesterol secretion rates,16 whereas SR-BI overexpression increases biliary cholesterol content.8, 9, 25 However, thus far the obligate heterodimeric ATP-binding cassette transporters Abcg5/Abcg8 have been characterized to represent the major and supposedly rate-controlling transport system for cholesterol excretion into bile. The role of Abcg5/Abcg8 has been substantiated by showing that overexpression of these transporters results in increased biliary cholesterol secretion19 and that knockouts for Abcg5 or Abcg8 have a decrease in biliary cholesterol secretion rates by about 75%.18, 21, 26 However, there is residual biliary cholesterol secretion in Abcg5 or Abcg8 knockout mice,18, 21, 26 and previous studies involving our group have shown that Abcg5/Abcg8-independent biliary cholesterol secretion occurs.21–23 Based on the results of this study, we would propose that SR-BI acts as a key candidate for mediating this Abcg5/Abcg8-independent fraction of biliary cholesterol secretion. As evidenced by the correction of the biliary cholesterol secretion deficit of Abcg5 knockout mice by hepatic SR-BI overexpression, SR-BI can drive biliary cholesterol secretion independent of Abcg5. Thereby, increased SR-BI protein expression at the canalicular membrane coincides with increased biliary cholesterol secretion. SR-BI can mediate bidirectional flux of cholesterol, i.e., influx and efflux depending on a cholesterol gradient.3–5 We speculate that SR-BI at the canalicular membrane might fulfill a similar function, whereby the gradient is provided through constant transport by the biliary secretion process. However, we cannot exclude that SR-BI expression at the canalicular membrane is resulting in more efficient micellization due to an increased cholesterol content of the canalicular membrane.
Formally, we also cannot rule out the possibility that SR-BI is mediating increased hepatic cholesterol uptake and that the increase in biliary cholesterol secretion occurs then via a different transporter. However, under certain (patho)physiological conditions, SR-BI might be rate-limiting for biliary cholesterol secretion. Previous studies from our group in transgenic mice indicated that modifying the main SR-BI ligand, HDL, through the action of secretory phospholipase A2 (sPLA2) results in a significant increase in the mass flux of cholesterol through SR-BI into the liver and subsequently an increased hepatic cholesterol content in the face of unchanged hepatic SR-BI expression and localization.27 Under these conditions, biliary cholesterol secretion remains unchanged in sPLA2 transgenic mice. However, if the hepatic expression level of SR-BI is increased by means of a recombinant adenovirus, biliary cholesterol secretion increases in transgenic mice even beyond levels observed in C57BL/6J controls (Supporting Fig. 5). We would interpret these data to indicate (1) that, at least in this model under conditions of increased cholesterol influx into the liver, SR-BI expression might be rate-limiting for biliary cholesterol secretion and (2) that this biological effect of SR-BI is not taken over by a different transporter.
Previous work localized SR-BI to the basolateral and canalicular membranes.28, 29 Here, by immunogold electron microscopy as well as western blot, we demonstrate SR-BI protein expression at the bile canaliculus and, in addition, show that canalicular SR-BI increases in response to SR-BI overexpression. In agreement with our results, a previous study also reported that hepatic overexpression of SR-BI by means of a recombinant adenovirus resulted in increased basolateral as well as canalicular SR-BI expression.8 However, these authors did not detect a significant change in phospholipid mass levels in gallbladder bile in response to AdSR-BI injection,8 although notably levels were 30% higher. This difference to our results might conceivably be due to the different method employed in our study, namely, continuous collection of hepatic bile by means of cannulation. In addition, the positive effect of SR-BI expression on biliary phospholipid secretion is not a consistent phenomenon in our study; importantly, it only occurred in C57BL/6J mice and not in mice of mixed or different genetic backgrounds. Although the underlying basis of this effect is not entirely clear, it might be related to the genetic background of the experimental mice.
Another interesting finding of our study is the lack of an effect of increased hepatic cholesterol levels in response to SR-BI overexpression on the expression of the Lxr target genes Abcg5/Abcg8. Given data suggesting a link between SR-BI activity and Abcg5/Abcg8 expression,24 this result was unexpected. However, the data of our current study are in line with previous results in a model with increased influx via SR-BI into the liver in the face of unchanged hepatic SR-BI mRNA and protein expression levels.27 In this model, an increase in HDL cholesterol catabolism by overexpression of sPLA2 resulted in increased hepatic cholesterol content, but without any change in the expression of Lxr target genes. Consistent with our current study, the expression of the Srebp2 target genes Ldlr as well as HMGCoA reductase was also decreased.27 Taken together, these data suggest that cholesterol derived from HDL and entering the liver via SR-BI is partitioned into compartments accessible toward Srebp2 but not toward Lxr sensing. Identification of the nature of these compartments will be the subject of future work.
Although some in vitro studies showed that SR-BI–mediated selective uptake does not require endocytosis30, 31 and SR-BI is able to mediate selective uptake into phospholipid vesicles in the absence of a cellular environment,32 other data suggested that SR-BI is an endocytic receptor.33 Vesicular trafficking of SR-BI together with HDL through the early endosome system to subapical compartments, where cholesteryl ester hydrolysis likely takes place, has been demonstrated.33 This is followed by recycling of the cholesterol-depleted HDL together with SR-BI to the basolateral membrane, while the free cholesterol traffics toward the canalicular membrane.33 Interestingly, unesterified cholesterol taken up via SR-BI moves by a nonvesicular pathway from basolateral to apical membranes.34 In addition, a transcytotic route for SR-BI has been characterized, which mediates the movement of the receptor from basolateral to apical membranes in a microtubule-dependent fashion.35, 36 Although transcytosis of SR-BI was not formally investigated in our study, our data indicate that transcytotic pathways do not contribute to the SR-BI–mediated increase in biliary cholesterol secretion, because cholesterol output into bile was not reduced in response to colchicine, which blocks microtubule function.
Taken together, our data are consistent with a model that SR-BI within the hepatocyte assumes a certain distribution ratio between basolateral and canalicular membranes, which depends on its mRNA expression level and possibly also on additional signals such as the cholesterol content of cellular membranes or the cellular environment. SR-BI mediates cholesterol uptake at the basolateral side; however, transcytosis pathways are not required for the subsequent increase in biliary cholesterol secretion in response to SR-BI. SR-BI might exert its functionality on-site at the canalicular membrane, facilitating cholesterol secretion into bile independent of Abcg5/Abcg8. Given the therapeutic potential of SR-BI in the areas of atherosclerotic cardiovascular disease as well as gallstone disease, further research will be required to delineate the specific in vivo signals that underlie a preferential localization of SR-BI to the canalicular membrane.
We are grateful to Rick Havinga for expert technical assistance with the animal experiments and to Dr. Henk Wolters (University Medical Center Groningen) and Dr. Coen Paulusma (Academic Medical Center Amsterdam) for advice on canalicular membrane isolations. We thank Dr. Karen Kozarsky (GlaxoSmithKline, King of Prussia, PA) for providing the SR-BI–expressing adenovirus used in this study, as well as Dr. Han van der Want and Freark Dijk (University Medical Center Groningen) for their invaluable help with electron microscopy and immunogold labeling. We are shocked and deeply grieved by the untimely death of the first author of this study and dedicate this manuscript in his memory.