SEMINARS IN LIVER DISEASE—VOL. 20, NO. 3, 2000
Mechanisms of Biliary Lipid Secretion and Their Role in
Lipid Homeostasis
RONALD P.J. OUDE ELFERINK, Ph.D. and ALBERT K. GROEN, Ph.D.
ABSTRACT: Bile secretion serves different important functions. First, it is one of the main mechanisms for the disposition of many endogenous and exogenous amphipatic compounds, including drugs, toxins, and waste products.
Second, it supplies bile salts to the intestine, which is of crucial importance for the emulsification of dietary lipids. In
the last decade considerable progress has been achieved in the elucidation of the process of bile formation. Several
key transporters in the canalicular membrane have been identified and characterized. This also holds for the mechanism of biliary lipid secretion, where the lipid translocating function of a P-glycoprotein was found to be indispensable for phospholipid secretion. Concomitantly, it became clear that bile salt-induced lipid secretion is an extremely
complex process, in which several steps remain elusive. The production of mice with a specific defect in bilary lipid
secretion and the identification of an analogous inherited human disease have made it possible to study the integrated function of biliary lipid secretion in whole body lipid homeostasis. In this review we discuss our current understanding of hepatocanalicular lipid secretion in this context. The pathologic consequences of defects in biliary
lipid secretion are discussed in another review in this issue.
KEY WORDS: bile, phospholipid, cholesterol
Bile salts, the main solutes in bile, are present in
concentrations that exceed the plasma concentration by
several orders of magnitude. This also holds for many
drugs and waste products that are secreted in bile. It has
become clear in the last decade that a set of ATPbinding cassette transporters (ABC transporters) in the
canalicular membrane is responsible for the uphill
transport of these compunds across the canalicular
membrane. Bile salts are transported via the bile salt
export pump (BSEP, formerly also called sister of
P-glycoprotein). The gene for this transporter has been
cloned, and it was demonstrated that mutations in this
gene lead to a strong or complete defect in biliary bile
salt secretion. The concentrative transport of organic
Objectives
Upon completion of this article, the reader should be able to 1) summarize the main steps involved in biliary secretion of phospholipids and
cholesterol, 2) understand the consequences of a defect in canalicular lipid secretion; and 3) understand the role of biliary lipid secretion in
whole body lipid homeostasis.
Accreditation
The Indiana University School of Medicine is accredited by the Accreditation Council for Continuing Medical Education to provide
continuing medical education for physicians.
Credit
The Indiana University School of Medicine designates this educational activity for a maximum of 1.0 hours credit toward the AMA Physicians
Recognition Award in category one. Each physician should claim only those hours of credit that he/she actually spent in the educational activity.
Disclosure
Statements have been obtained regarding the authors’ relationships with financial supporters of this activity. There is no apparent conflict of
interest related to the context of participation of the authors of this article.
From the Laboratory for Experimental Hepatology, Academic Medical Center, Amsterdam, The Netherlands
Reprint requests: R. Oude Elferink, Lab. Exp. Hepatology, Academic Medical Center F0–116, Meibergdreef 9, 1105 AZ Amsterdam, The
Netherlands. E-mail: [email protected]
Copyright © 2000 by Thieme Medical Publishers, Inc., 333 Seventh Avenue, New York, NY 10001, USA. Tel.: +1(212) 584-4662.
0272-8087,p;2000,20,03,293,306,ftx,en;sld00067x
293
294
SEMINARS IN LIVER DISEASE—VOL. 20, NO. 3, 2000
anions is mediated by another ABC transporter, multidrug resistance protein 2 (MRP2). A third ABC transporter, MDR3 in humans and the orthologue Mdr2 in rodents, is responsible for the secretion of phospholipid.
The production of mice with a disrupted mdr2 gene
highlighted the importance of Mdr2 P-glycoprotein for
biliary phospholipid secretion. This protein appeared to
be absolutely essential for lipid secretion by its function
as a phospholipid flippase. The human homolog MDR3
P-glycoprotein serves an identical function, and it was
discovered that mutations in this gene lead to a form
of progressive familial intrahepatic cholestasis (PFIC
type 3).
Biliary secretion of phospholipid and cholesterol
represent major lipid fluxes, but at the same time it can
be questioned whether this is of importance for whole
body lipid homeostasis. Both phospholipids and cholesterol are reabsorbed to a large extent in the intestine.
Although cholestasis (e.g., by bile duct obstruction)
leads to hypercholesterolemia, a specific defect in biliary lipid secretion appears to have limited effects on
plasma and tissue lipid concentrations, both in mice and
in humans. This suggests that both liver and intestine
can adapt quite well to a situation in which biliary lipid
secretion is absent. In contrast, in humans extensive
cholesterol secretion frequently leads to the formation
of gallstones, a condition that on the long term induces
chronic inflammation of the gallbladder and thereby
represents an important health problem. Why then do
we secrete large amounts of lipids into bile? The recent
recognition of extensive liver pathology associated with
a defect in biliary lipid secretion has demonstrated that
biliary lipids are of crucial importance for the protection
of hepatocytes and bile duct epithelial cells against the
cytotoxicity of detergent bile salts. In this review we
also discuss the recent advances in our knowledge of the
integrated proces of biliary lipid secretion and its integration in lipid homeostasis.
SOURCES OF BILIARY LIPIDS
Lipids Derived from Plasma Lipoproteins
The hepatocyte is a major site of lipoprotein synthesis and release into the plasma, mainly in the form of
very-low-density lipoprotein (VLDL). This lipoprotein
mainly contains triglycerides that are assembled with
apolipoprotein B in the endoplasmic reticulum and released in to the blood.1 VLDL assembly requires a relatively small amount of cholesterol and phosphoplipid
derived from de novo synthesis. The availability of
triglycerides (fatty acids) is a major regulatory factor in
VLDL synthesis. It has been suggested that, at least in
rodents, VLDL release into plasma and biliary lipid secretion are reciprocally regulated.2,3 Indeed, several
conditions known to increase biliary cholesterol secretion, such as feeding2,3 diosgenin or fish oil,4,5 lead to
reduced VLDL release. The hepatocyte also receives
lipids from the plasma compartment, mainly in the form
of low-density (LDL) and high-density lipopoprotein
(HDL). These lipoproteins contain relatively high
amounts of cholesterol(ester) and are therefore key
players in cholesterol homeostasis. Of these two
lipoproteins, HDL is the more important source of biliary lipid, especially of cholesterol.6–8 LDL particles
arise from VLDL upon extensive modification by
lipoprotein lipase in the plasma. Nascent HDL particles
are synthesized by the liver and by the intestine. They
acquire free cholesterol from the plasma membrane of
cells in peripheral tissues that is converted into cholesterolester by the action of lecithin cholesterol acyltransferase. Cholesterol(ester) from HDL can subsequently
be delivered to the liver (reverse cholesterol transport)
or can be transferred to LDL by the action of cholesterol
ester transfer protein.
The mechanism of delivery of LDL and HDL to the
liver is fundamentally different. LDL particles are endocytosed after binding to the hepatic LDL receptor, transported to the lysosome, where they are degraded.9,10
Delivery of HDL cholesterol probably does not require
endocytosis of the particle. The liver was found to take
up more cholesterol than apo A-I from HDL, suggesting
that the particles are not endocytosed and/or metabolized as such.11,12 More recently, scavenger receptor BI
(SR-BI), the receptor responsible for binding of HDL,
was cloned and characterized.13 This receptor is highly
expressed in the liver and in the steroidogenic cells of
the ovaries, adrenal cells, and testes, the main sites of
selective uptake of HDL cholesterol in vivo. The current
model is that after binding of the HDL particle, this receptor facilitates the molecular uptake of cholesterol(ester).14 SR-BI turns out to be a key player in HDL homeostasis; mice with a disrupted Sr-bI gene have
significantly increased HDL cholesterol levels.15
Furthermore, polymorphisms in this gene associate with
alterations in plasma lipid levels and body mass index.16
Conversely, both stable overexpression of SR-BI in
transgenic mice17 or transient overexpression by transduction with an appropriate adenoviral vector18 leads to
a dramatic reduction of plasma HDL cholesterol. In addition, and more importantly, in the context of biliary
lipid secretion, the expression level of SR-BI also regulates the level of biliary cholesterol secretion. Knockout
mice for the Sr-bI gene have reduced biliary cholesterol
secretion rates,15 whereas overexpressing mice have increased secretion rates.18,19 The strong overexpression
that was reached with adenoviral transduction led to a
virtual disappearance of HDL and a simultaneous doubling of biliary cholesterol secretion. An important observation in these studies was that the changes in biliary
cholesterol secretion were independent from biliary
BILIARY LIPID SECRETION—OUDE ELFERINK, GROEN
phospholipid secretion demonstrating that Sr-bI is not a
major factor in the supply of phospholipids.
295
cretion, suggesting that either HDL is not a major source
or increased de novo synthesis can compensate for this situation. Within the hepatocyte, PC molecules can be extensively remodeled with regard to their fatty acyl chains.27
De Novo Synthesis in the Hepatocyte
The above-described data fit with the existing idea
that most biliary cholesterol is derived from plasma
lipoproteins. Several studies have provided estimations
of the contribution of de novo synthesis to biliary secretion and these amount to about 5% for cholesterol.20 The
contribution of de novo phospholipid synthesis seems to
be higher than that of cholesterol. Phosphatidylcholine
(PC) synthesis in the hepatocyte occurs mainly via the
cytidyldi-phosphate (CDP)-choline pathway, which
uses phosphocholine and diacylglycerol as precursors
(Kennedy pathway). The rate-controlling step in this
pathway is the conversion of CDP-choline to phosphocholine by the enzyme CTP: phosphocholine cytidylyltransferase (CT), which has been characterized at the
gene and protein level.21 A second minor pathway involves methylation of phosphatidylethanolamine
(PE).22 The enzyme responsible for this step, phosphatidylethanolamine N-methyltransferase, has been
cloned and characterized.23 The contribution of this
pathway to PC synthesis in the liver has been estimated
as 20–40%.24 Walker et al.25 directly assessed the role of
pemt in phospholipid synthesis by the production of
pemt/ mice. These mice have no obvious endogenous phenotype and their liver content of PC is normal.
This does not mean that pemt is irrelevant to PC synthesis in the liver. Walkey et al. observed a 60% increase in
the activity of CT in livers from pemt/ mice, indicating that these animals can compersate for the loss of
pemt activity by induction of the Kennedy pathway. The
importance of pemt-mediated PE methylation in the
liver became clear when they fed the pemt/ mice a
choline-deficient diet, thereby minimizing PC synthesis
via the Kennedy pathway.26 Under this condition the animals did develop a phenotype; the liver PC content was
reduced by 50% and within 3 days the animals developed severe liver pathology with strong fatty vacuolization. Concomitantly, there was a dramatic fall in plasma
VLDL and HDL levels, suggesting that the liver is not
capable to mobilize sufficient phospholipid for lipoprotein synthesis. Data for biliary lipid secretion have not
been reported yet in choline-deficient pemt/ mice;
although it may be expected that biliary lipid secretion
is reduced, it will be interesting to see to what extent
this condition affects biliary lipid secretion in comparison with lipoprotein synthesis.
Apart from de novo synthesis, PC may be derived
from extrahepatic sources. PC from the coat of HDL may
contribute,7 but the recent experiments in which HDL uptake was modulated by overexpression or deficiency of
the HDL receptor, SR-BI, showed little effect on PC se-
FEATURES OF HEPATOCANALICULAR
LIPID SECRETION (FIG. 1)
A tight coupling between lipid and bile salt secretion has been demonstrated in numerous animal
species,28,29 although the ratio of lipid to bile salt, as
well as the types of secreted bile salts, varies considerably between species. The number of hydroxyl groups
and the stereospecific position in the steroid nucleus of
these molecules determine the hydrophobicity of the
molecule and its detergent properties. There is a hyperbolic increase in phospholipid secretion with increasing
bile salt output, and the slope of this relation is steeper
during secretion of hydrophobic bile salts compared
with more hydrophilic bile salts.30 These observations
suggest that bile salts in some way solubilize phospholipid (and cholesterol) from the canalicular membrane.
Indeed, it could be demonstrated that retrograde injection of bile salt into the biliary tree elicits a burst of
lipids, indicating that the mere contact of detergent with
the canalicular membrane is sufficient to drive lipid
secretion.31
Despite this concept, there is a striking difference
between the composition of biliary phospholipids and
those in the canalicular membrane. Biliary phospholipids consist almost exclusively of PC (>95%),
whereas the canalicular membrane also contains sphingomyelin (±20%), phosphatidylethanolamine (±20%),
and phosphatidylserine (±20%).32 Furthermore, PC in
bile is more hydrophilic than the PC species in the
canalicular membrane.33 Biliary PC predominantly contains palmitate (16:0) at the sn1 position and oleate
(18:1) or linoleate (18:2) at the sn2 position, whereas the
sn2 position of PC from the membrane contains more
arachidonate (20:4) (for review, see ref. 28).
A third specific feature of biliary lipid secretion is
that it is crucially dependent on the function of the
phopsholipid translocating P-glycoprotein, MDR3 Pglyprotein in humans, and Mdr2 P-glycoprotein in rodents. These transporters are members of the ABC family. This was found by the absence of lipid secretion in
mice in which the mdr2 gene had been disrupted.34 Bile
from mdr2/ mice turned out to be almost devoid of
both phospholipid and cholesterol, whereas bile salt secretion was normal. Bile of mice heterozygous for mdr2
gene disruption (mdr2+/ mice that have 50% expression of mdr2 P-glycoprotein compared with control
mice) had a 40% decreased phospholipid but normal
cholesterol content.35 This strongly suggested that mdr2
P-glycoprotein catalyzes a rate-controlling step in the
296
SEMINARS IN LIVER DISEASE—VOL. 20, NO. 3, 2000
FIG. 1. Lipid fluxes in the hepatocyte.
biliary secretion of phospholipids. Furthermore, it
demonstrates that the absent phospholipid secretion in
mdr2/ mice is the primary defect caused by mdr2
gene disruption, whereas the absent cholesterol secretion is a secondary effect. When mdr2/ mice were
intravenously infused with taurodeoxycholate, a bile
salt that is more hydrophobic than the endogenous bile
salts of the mouse (±70% muricholate and 30%
cholate), there was still no significant phospholipid secretion, whereas this condition did induce cholesterol
secretion.36
A fourth feature of biliary lipid secretion is that it
most likely occurs in the form of vesicles. Vesicles containing PC and cholesterol, besides mixed micelles of
PC, cholesterol, and bile salts, have been unequivocally
demonstrated in human bile samples collected downstream in the biliary tree.37–39 Hence, its composition
does not necessarily reflect that of primary bile in the
canaliculus. It could be that modification of bile during
its passage through the biliary tree leads to the formation
of vesicles. Indeed, water is added in the bile ductules,
and the resulting decrease in total lipid concentration
could induce a transition from the micellar to the vesicular phase.38,39 Two morphologic studies have, however,
suggested that lipid vesicles also exist at the canalicular
level. In an electronmicroscopic study, Ulloa et al.40 observed lipid vesicles within the canalicular lumen of human liver specimens. Crawford et al.41 used an ultrarapid
freezing technique to study the morphology of the
canalicular lipid secretion process. They observed characteristic electron-lucent vesicular structures that were
clearly distinct from microvilli and attached to the
canalicular membrane by an electron-dense base. The
presence of these vesicular structures depended on ongoing bile salt secretion during the freezing of the tissue.
Furthermore, these vesicular structures were not observed during infusion of the nonmicelle forming bile
salt, taurodehydrocholate, which is incapable of driving
lipid secretion. The fact that these “stalked” vesicles are
significantly reduced in liver specimens from mdr2
(/) mice42 strongly suggests that they represent intermediate structures of lipid secretion.
Mdr2 and MDR3 P-glycoproteins
are PC Flippases
Several studies have subsequently demonstrated
that Mdr2 and MDR3 P-glycoprotein are able to translocate phospholipid from the inner leaflet of the cell membrane to the outer leaflet.43–45 Most of these studies have
used fluorescently labeled phospholipid analogues as
probes to monitor translocation. Ruetz and Gros43 have
analyzed the function of the mouse mdr2 P-glycoprotein
expressed in yeast. They were able to show net translocation of 7-nitro-2,l,3-benzoxadiazol-4-yl-phosphatidyl
choline (NBD-PC) in mdr2 P-glycoprotein-containing
vesicles in the presence of ATP. Using a very similar assay, Nies et al.46 found similar results in canalicular
membranes from rat and showed that this could be stimulated by the presence of bile salt. Smith et al.44 demonstrated enhanced translocation of natural PC in fibro-
297
BILIARY LIPID SECRETION—OUDE ELFERINK, GROEN
blasts from mice overexpressing MDR3 P-glycoprotein.
In an elegant study with polarized epithelial cells that
were transfected with MDR3 P-glycoprotein, van
Helvoort et al.45 showed that this P-glycoprotein transports NBD-labeled phosphatidylcholine but not the
NBD-labeled analogues of other main membrane phospholipids, sphingomyelin, phosphatidylethanolamine,
and phosphatidylserine. This observation might in part
explain the selectivity of biliary phospholipid secretion.
Although the specificity of MDR3 or mdr2 P-glycoprotein for subspecies of PC with various acyl chains
has not been studied in any detail, van Helvoort et al.45
showed that MDR3 P-glycoprotein is not able to
translocate PC with two (short) C6 acyl chains.
the presence of cholesterol, sphingomyelin molecules
with shorter fatty acyl chains (C16:0) were preferentially solubilized over sphingomyelin molecules with
long fatty acyl chains (>C22:0). Thus, this study
clearly demonstrated that it is the combination of cholesterol and sphingomyelin that makes membranes
specifically resistant toward bile salts. A model in
which the combination of cholesterol and sphingomyelin forms bile salt-resistant domains fits with the
observation that cholesterol has a high affinity for
sphingomyelin49 and these lipids may actually be laterally segregated in membranes.50,51 The canalicular
membrane is enriched in both cholesterol52,53 and
sphingomyelin of long saturated acyl chain composition,54 and this may explain the particular resistance of
this membrane toward bile salts that is required to prevent aspecific and uncontrolled solubilization.
Composition of the Canalicular Membrane
The observation that mdr2/ mice do not secrete any lipid into bile whereas (normal) high bile salt
secretion are reached in the canalicular lumen leads to
the assumption that in the absence of this flippase
function (as in the case of the mdr2/ mice and PFIC
type 3 patients), only detergent-resistant lipids are
thought to reside in the outer leaflet of the membrane,
because in this situation no lipids are extracted from
the membrane. This resistance may be achieved by a
high content of sphingomyelin, cholesterol, and possibly also glycolipids. In model systems, these lipids
were found to be quite detergent resistant. It was observed that sheep erythrocytes are more resistant toward bile salts than red blood cells from pigs or humans. The main difference in lipid composition
between these species is the higher sphingomyelin content of the sheep erythrocytes.28 In addition, Velardi et
al.47 observed that the incorporation of cholesterol in
erythrocytes membranes considerably increased their
resistance toward bile salts. In a recent elegant study
by Eckhardt et al.48 this aspect was studied in detail.
These authors prepared small unilamellar vesicles of
different phospholipid composition (PC and sphingomyelin) and with different cholesterol to phospholipid ratios. They incubated these vesicles with taurocholate and analyzed the distribution of phospholipid
between the micellar and vesicular phase. When the
vesicles contained little cholesterol, both PC and
sphingomyelin were well solubilized in bile salt micelles and sphingomyelin-containing vesicles were
actually less resistant toward taurocholate. This solubilization decreased drastically, however, when increasing amounts of cholesterol were added to the vesicles
(up to a cholesterol to phospholipid ratio of 0.4). Under
this condition, PC was preferentially solubilized in the
micellar phase, whereas sphingomyelin remained in
the vesicular phase. In addition, it was observed that in
Hypothetical Model for
Canalicular Lipid Secretion (Fig. 2)
The above-mentioned features of biliary lipid secretion have led to a model that was centered around the
experimental data suggesting that mdr2/MDR3
P-glycoprotein function as flippases for phosphatidylcholine (see below). Through flipping of PC from the
inner to the outer leaflet of the canalicular membrane,
the phospholipid becomes available for bile saltinduced liberation into the canalicular lumen. Upon
translocation, PC molecules may reside in microdomains in the outer leaflet that otherwise consists
of more rigid detergent-resistant lipids. The active
pumping by Mdr2 P-glycoprotein may cause a phospholipid excess in the outer leaflet that destabilizes these
PC-containing microdomains. This is, however, not sufficient to induce PC secretion in the absence of bile
salts. The simultaneous secretion of bile salts via Bsep
leads to high supramicellar concentrations in the lumen,
which will further destabilize the microdomains, possibly by the formation of inverted micelles in the membrane. Through the ongoing translocation of PC, the domains grow into vesicular structures that can pinch off
to yield biliary lipid vesicles. According to this model,
all lipids that are translocated are also secreted into the
lumen as vesicles; therefore, the selective secretion of
bile-type PC must be determined by the substrate specificity of Mdr2 P-glycoprotein in this model. Currently,
it remains unclear whether other proteins within, or attached to, the canalicular membrane play a role in this
process. In view of its complexity and the importance to
protect the cell against membrane leakage during this
process, it may well be that other proteins are important
for the vesiculation itself and/or for the stabilization of
the membrane during this process.
298
SEMINARS IN LIVER DISEASE—VOL. 20, NO. 3, 2000
FIG. 2. A hypothetical model for canalicular lipid secretion.
INTRACELLULAR SUPPLY OF LIPIDS TO
THE CANALICULAR MEMBRANE
Are Cytosolic Proteins Involved?
The function of P-glycoprotein as a flippase in the
canalicular membrane requires that PC is delivered to
this membrane in an asymmetric fashion. If supply occurred by means of vesicles that contain PC both in the
inner and in the outer leaflet, 50% of the PC inserted in
the canalicular membrane would already be in the outer
leaflet and therefore available for extraction. The complete absence of PC secretion in the mdr2/ mice
shows that this is not the case. Therefore, it must be assumed that supply does not occur via symmetric vesicles. Cohen et al.55,56 performed a series of elegant experiments that strongly suggested that the cytosolic
PC-transfer protein (PC-TP) might be capable of performing this function. PC-TP is a 25-kDa cytosolic protein that was demonstrated initially to exchange phosphatidylcholines between membrane vesicles.57,58
Cohen et al.55 showed that the protein is also capable of
net transport of PC from PC-rich membranes to membranes that are devoid of PC. They further showed that
this process is stimulated by the presence of submicellar
concentrations of bile salts and that the efficiency of this
induction increased with increasing bile salt hydrophobicity. The function of PC-TP in PC supply to the
canalicular membrane would be an attractive mechanism to explain the necessary asymmetric supply, because it would exclusively donate phospholipid molecules to the inner leaflet of the membrane. This
suggestion was further strengthened by the observation
of LaMorte et al.59 that the affinity of different molecular phosphatidylcholine species correlated well with the
the species found in bile. To test a possible function of
Pc-tp in overall biliary phospholipid secretion, van
Helvoort et al.60 produced a knockout mouse for the Pctp gene. These mice were found to lack any endogenous
phenotype and had a normal phospholipid and cholesterol secretion into bile. Even when phospholipid secretion rates were pushed to a very high rate, by infusion of
tauroursodeoxycholate (TUDC), there was no difference between normal and Pc-tp/ mice. This excluded that Pc-tp plays a crucial role in the intracellular
supply of phospholipid. This study showed also for the
first time that Pc-tp is not primarily an hepatic protein
but is much higher expressed in epididymis, testis, and
bone marrow-derived mast cells. Furthermore, hepatic
expression of Pc-tp was only high until 1 week after
birth and then dropped to much lower levels. Currently,
the physiologic function of Pc-tp remains elusive.
Sterol carrier protein 2 (SCP-2) is another candidate for intracellular lipid transfer. This is a 13-kDa protein that was originally characterized as a cytosolic factor required for the conversion of 7-dehydrocholesterol
to cholesterol.61 It appeared to be identical to the nonspecific lipid transfer protein that catalyzes the exchange of a variety of phospholipids.62 More recently,
purified SCP-2 was demonstrated to have a higher affinity for fatty acids and their CoA-esters than for sterols.63
SCP-2 is translated as a larger precursor from which
also the peroxisomal thiolase is derived.64 The localization of this protein within the cell is a matter of debate;
although originally described as a cytosolic protein, it
bears a C-terminal serine-lysine-leucine (SKL) peroxisomal targeting signal and has been localized to the peroxisomes by immunocytochemistry.65,66 If SCP-2 would
have an exclusive peroxisomal localization, it is obviously unlikely to play a role in cytosolic lipid transfer.
BILIARY LIPID SECRETION—OUDE ELFERINK, GROEN
Fuchs et al.67 observed that Scp-2 expression is elevated
in C57L mice that hypersecrete cholesterol and are susceptible to gallstone formation upon cholesterol feeding
as opposed to insensitive AKR mice. These observations suggest that Scp-2 might play a role in overall biliary lipid secretion.
Puglielli et al.68 suggested a role for Scp-2 in the
supply of newly synthesized cholesterol to the canalicular membrane. They treated rats with an antisense
oligonucleotide for Scp-2 and observed a significant decrease in biliary cholesterol secretion. In addition, they
observed that Scp-2 was overexpressed in rats that fed
diosgenin, which is known to specifically increase biliary cholesterol secretion.
Seedorf et al.64 produced knockout mice for the
Scp-2 gene and found no abnormalities in plasma or
liver cholesterol levels (although the liver cholesterol
ester level was reduced by 50%). In contrast, their experiments revealed that catabolism of methyl-branched
fatty acyl-CoAs, which is a peroxisomal process, was
strongly impaired.
Role of Vesicular Transport
Besides molecular supply of lipids to the canaliculus via cytosolic proteins, a second route of supply may
be by vesicular transport. It is clear that there is extensive vesicular transport to the canalicular membrane.
Most attention has been given to this aspect with regard
to the delivery of canalicular proteins and transcytosis
of IgA, but unfortunately, little information can be
drawn from this work with respect to lipid supply.
Various studies have used microtubule inhibitors to assess the contribution of vesicular transport to biliary
lipid secretion, and these have provided different results
(for review, see Ref. 69). The bottom line is that lipid
secretion is partly sensitive to compounds like
colchicine and monensine.69–71 As mentioned above, the
current paradigm of Mdr2/MDR3 P-glycoprotein as
phospholipid translocators requires asymmetric delivery of PC to the canalicular membrane. Thus, vesicular
supply requires an asymmetric distribution of lipids
with PC mainly in the cytoplasmic and detergent-resistant lipids (sphingolipids and cholesterol) mainly in the
exoplasmic leaflet of these vesicles. Studies from different groups using fluorescent lipid analogues strongly
suggest that sphingolipids are indeed specifically targeted to the apical membrane of epithelial cells in general and hepatic cells in particular (for extensive review
of this issue, see Hoekstra et al.).
A third route of lipid supply has to be considered
within the context of asymmetric insertion of PC in the
inner leaflet of the canalicular membrane, and this is the
delivery of lipids via lateral diffusion from the basolateral membrane to the canalicular membrane. Tight junc-
299
tions are impermeable to lateral diffusion in the outer
leaflet, but they are permeable for lateral movement of
lipids in the cytoplasmic leaflet.72 Recent work by
Robins and Fasulo73 suggest that such a pathway is certainly possible for cholesterol (analogues). They analyzed the biliary appearance of sitostanol, a plant sterol
that is secreted into bile with the same kinetics and in
direct proportion to cholesterol. By pulsing the perfused
liver with this sterol, which is normally not present,
they could define the velocity with which it transverses
the hepatocyte. This and various of their previous studies revealed several important aspects. First, delivery of
sitostanol to the liver via HDL was much more efficient
than with other lipoproteins with regard to biliary secretion. Second, already within 2–4 minutes after addition
of HDL-sitostanol the sterol was detected in bile.
Importantly, the ratio of sitostanol to cholesterol in bile
was much higher than that in the liver, strongly suggesting that the plant sterol circumvented the hepatocyte
pool of cholesterol. These results strongly suggest that
the sterol is directly delivered to the canalicular membrane via lateral diffusion along the plasma membrane.
Although no data are available on such a mechanism for
phospholipid, there is no reason to exclude the possibility that PC might take a similar route. If so, all traffic of
vesicles, containing either newly synthesized or remodeled PC and cholesterol, to the basolateral membrane
would represent a direct source of biliary lipid via lateral diffusion from the inner basolateral to the inner
canalicular leaflet of the plasma membrane. Combined
with the selective inner leaflet permeability of the tight
junctions, this would represent an efficient asymmetric
delivery of lipids to the canalicular membrane.
RELATION BETWEEN P-GLYCOPROTEINS
P-glycoproteins were initially characterized as extrusion pumps that confer resistance against cytotoxic
amphipathic compounds by the extrusion of these molecules from the cell. The drug transporting P-glycoproteins (class I/II P-glycoproteins) include MDR1 in humans and mdr1a and mdr1b in the mouse. A prominent
feature of these P-glycoproteins is that they can adapt a
wide variety of substances, and although extensive studies have been performed, it is only partly understood
which features are of importance for the affinity for the
active site of the protein. With the characterization of
mdr2 P-glycoprotein and MDR3 P-glycoprotein (class
III P-glycoproteins) as phospholipid translocators, the
situation became only more complex; on the one hand,
the class I/II P-glycoproteins appeared to be highly unspecific, whereas the class III P-glycoproteins manifest
themselves as specific for phospholipids and in particular PC. Indeed, class III P-glycoproteins were found not
to confer multidrug resistance. This notion is difficult to
300
reconcile with the high extent of amino acid sequence
identity (78%) between MDR1 and MDR3 P-glycoprotein. The close functional similarity of these two proteins has been further substantiated by the fact that considerable pieces of MDR1 sequence can be exchanged
with the corresponding MDR3 sequence without loss of
the drug transporting function.74–78
Because the broad specificity of MDR1 P-glycoprotein has been settled beyond any doubt, the question
rises how specific MDR3 P-glycoprotein really is for
PC. This question was partly addressed in an elegant
study by van Helvoort et al., who studied translocation
of short chain lipid analogues that carried fluorescent
NBD moieties, in polarized porcine kidney epithelial
cells (LLC-PK1) cells transfected with the various Pglycoproteins. They observed that MDR3 P-glycoprotein does not transport C6-NBD-sphingomyelin nor
C6-NBD-phosphatidylethanolamine. The first lipid analogue has the same headgroup as PC, whereas the latter
has a similar backbone (apart from the fatty acid that
was replaced by the C6-NBD group). This suggested
that the specificity of MDR3 P-glycoprotein is determined by both headgroup and lipid backbone. This was
further strengthened by the observation that C8C8-PC
without an NBD moiety was also not transported by
MDR3 P-glycoprotein. In contrast to MDR3, MDR1 Pglycoprotein translocated all the short chain lipid analogues, including short chain analogues of PC. Thus, the
conclusion from this study was that MDR3 P-glycoprotein is a lipid translocator that is highly specific for PC,
whereas MDR1 P-glycoprotein (as well as the murine
mdr1a P-glycoprotein) has a very broad substrate specificity. The latter does, however, not include natural PC,
because the high expression of mdr1a/b P-glycoprotein
in the hepatocanalicular membrane of mdr2/ mice
does not lead to any phospholipid secretion into bile. A
few other studies, however, suggest that MDR3 P-glycoprotein is not only specific for PC. First, in a few
studies a correlation was observed between MDR3 overexpression in B-cell leukemias and transport of
daunorubicin, suggesting that MDR3 P-glycoprotein
might be able to transport this amphipath. Kino et al.
transfected yeast with MDR3 cDNA and observed low
level resistance for aureobasidin A, an antifungal amphipathic peptide. These studies, however, only provide indirect suggestions that MDR3 P-glycoprotein
might be capable of transporting other compounds than
PC. A more direct study was performed by Smith.79 He
used the LLC-PK1 cells again, transfected with MDR1
and MDR3 cDNAs, and studied the vectorial transport
of various drugs that are known as MDR1 P-glycoprotein substrates. Surprisingly, MDR3 P-glycoproteinmediated transport in the apical direction was observed
for digoxin, paclitaxel vinblastine, and ivermectine.
The transport was less efficient than MDR1-transfected
cells with similar expression levels but could be inhib-
SEMINARS IN LIVER DISEASE—VOL. 20, NO. 3, 2000
ited by typical MDR1 reversal agents such as cyclosporin A, verapamil, and PSC 833. These compounds were also able to inhibit transport of C6-NBDPC. These results strongly suggest that MDR3
P-glycoprotein is not specific for PC but is able to
translocate various typical MDR1 substrates as well. In
contrast to this, we have never been able to inhibit biliary PC secretion in intact mice or perfused mouse livers with verapamil, which would be expected if verapamil is an inhibitor of mdr2 P-glycoprotein (R. Oude
Elferink, unpublished observations, 1997). A possible
explanation for this might be that only the human
MDR3 is able to transport these drugs, whereas this is
not the case for the mouse protein. Alternatively, the
concentration of PC in the inner leaflet of the canalicular membrane may be higher than that in the apical
membrane of LLC-PK1 cells, so that mdr2 in the
canalicular membrane may be more or less saturated
with PC, whereas this may not be the case in LLC-PK1
cells. Clearly, the issue of the genuine substrate specificity of the class III P-glycoproteins has not been settled yet. Probably, MDR3 P-glycoprotein will have to
be reconstituted in membranes devoid of PC to analyze
which drugs are transported and to what extent compared with MDR1 P-glycoprotein.
REGULATION OF mdr2/MDR3
EXPRESSION
As described above, there is an intimate relation between bile salt and lipid secretion into bile. Not only is
lipid secretion directly driven by the canalicular secretion
of bile salts, but lipid secretion is also crucial for the inactivation of the cytotoxic detergent action of bile salts.
This intimate relation raised the question whether the expression and/or activity of mdr2/MDR3 P-glycoprotein
would be regulated by bile salts. Frijters et al.80 addressed
this question by feeding normal mice a diet with 0.1%
cholate, which enlarges the bile salt pool and increases
hepatic bile salt concentrations. This condition was found
to increase mdr2 mRNA levels in mice and, correspondingly, biliary lipid secretion increased to the same extent.
Interestingly, feeding these mice a fivefold higher
amount of ursodeoxycholate was without effect, suggesting that the induction of mdr2 depends on the type of bile
salt in the pool. The extent of induction by cholate feeding was, however, limited. Only a 42% increase in mdr2
mRNA was observed. It was hypothesized that under
normal conditions, mdr2 is already largely induced, and
this raised the possibility that reduction of the bile salt
pool would lead to a substantial decrease in mdr2 expression. Indeed, chronic diversion of bile in rats (for 4
and 8 days) led to a fivefold reduction in hepatic mdr2
mRNA levels with a corresponding decrease in phospholipid secretion capacity (C.M.G. Frijter, P.J. Bosma, R.J.
BILIARY LIPID SECRETION—OUDE ELFERINK, GROEN
Fryters, A. Groen, R. Oude Elferink, unpublished data,
2000). The data from these experiments confirmed the
hypothesis that mdr2 expression is regulated by the prevailing bile salt concentrations in the liver.
Carrella et al81 observed that chronic treatment of
rats with pravastatin led to an increase in mdr2 mRNA
level, but also to increased expression of the key enzyme
in PC synthesis, CT. They hypothesized that increased
PC synthesis might represent a signal for increased expression of mdr2. Alternatively, the reverse may also be
true; in a pull mechanism, increased expression of mdr2
P-glycoprotein may require increased PC synthesis and
therefore lead to induction of expresion of cytidylyltransferase. Hooiveld et al.82 also studied the effect of
pravastatin on mdr2; they found very similar results and
hypothesized that the effect of pravastatin might be mediated through its inhibition of cholesterol synthesis.
They proposed that mdr2 expression may be under control of sterol regulatory element binding proteins
(SRBEPs). Indeed, a consensus sequence for SREBP is
present in the 5 sequence of the rat mdr2 gene.
IS BILIARY SECRETION IMPORTANT FOR
LIPID HOMEOSTASIS?
It is often stated that biliary lipid secretion is important because it is the only pathway by which cholesterol can leave the body. Cholesterol is secreted as such,
but because it is the precursor molecule for bile salt synthesis, bile salt secretion also contributes to the elimination of cholesterol. The idea that biliary secretion is important for whole body cholesterol elimination needs
some adjustment, however. First, a substantial amount
of the secreted cholesterol is reabsorbed in the intestine.
Although this has been subject of extensive studies, the
process by which cholesterol is taken up in the gut is
largely unknown. Whatever the mechanism is, the reabsorption step constitutes a potentially important regulatory point in whole body homeostasis. In this context, it
is important to note that in the intestine, biliary cholesterol is mixed with two other pools of cholesterol, (i.e.,
dietary cholesterol and cholesterol derived from
sloughed intestinal cells). Because the turnover of intestinal cells is very high, this cholesterol pool is substantial. It has been estimated that dietary cholesterol on
a Western diet amounts to 300–500 mg/day83 and the
cholesterol derived from intestinal cells, 250–400
mg/day. Bile releases a daily amount of 800–1,200 mg
cholesterol to the intestine. This means that somewhat
less than half of the cholesterol that enters the intestine
originates from bile. The question what the biliary contribution to cholesterol homeostasis is can be approached by looking at the effect of complete absence
of biliary cholesterol secretion. One could look at the
301
effect of complete cholestasis, such as bile duct ligation
in experimental animals, but this does not give a proper
answer because cholestasis has many more effects than
the absence of biliary cholesterol secretion. During
cholestasis, there is also an absence of bile salts in the
intestinal lumen, and it is known that bile salts are of
crucial improtance for intestinal uptake of cholesterol.84
To get more specific insight in the importance of biliary
cholesterol and phospholipid secretion on lipid homeostasis, Voshol et al.85 studied lipoprotein levels in the
mdr2/ mouse. In contrast to experimental cholestasis, bile salt secretion is normal in these animals, but
phospholipid and cholesterol secretion are completely
absent. Although the hepatic contents of phospholipid
and cholesterol were not different from control mice of
the same strain, a significant decrease in plasma cholesterol (by 65%) was observed, and this could be attributed to a decrease in HDL levels (and to a less extent in
VLDL levels). The activity of HMG-CoA reductase in
the liver of mdr2/ mice was threefold higher than
in control mice, but the plasma cholesterol decay, as
measured by injection of radiolabeled cholesterol, was
not different, suggesting that the rate of cholesterol synthesis was not influenced by the absence of biliary cholesterol secretion. Surprisingly, a significantly increased
fecal secretion of neutral sterols was found. The discrepancy between an unchanged cholesterol synthesis
and an increased fecal sterol loss was explained by assuming an increased turnover of intestinal cells. The
reason for this might be that lipid free bile salts might
not only be cytotoxic in the biliary tree but also in the
intestine. In a subsequent study, Voshol et al.86 found
that the absorption of triglycerides was delayed in
mdr2/ mice, and this was accompanied by lipid accumulation in the small intestinal wall. The delayed
lipid absorption could be partly restored by duodenal infusion of bile. This suggested that biliary phospholipid
is necessary for proper intestinal assembly of chylomicrons. It was, however, also observed that quantitative
long-term lipid absorption was not affected in mdr2
/ mice. A possible explanation for these seemingly
discrepant results may be that by passage of lipids to a
more distal part of the intestine, sufficient capacity is
mobilized to ensure quantitative absorption.
Unfortunately, very little is known about the
plasma lipoprotein pattern in patients with PFIC type 3
(MDR3 deficiency). This would be much more informative with regard to the role of biliary lipid secretion in
human lipid homeostasis. The available data suggest,
however, that there are no major clinical or biochemical
signs of disturbed lipid homeostasis (see also next section). Thus, the specific absence of biliary lipid secretion appears to have a limited effect on plasma and liver
lipid homeostasis, probably because the body can sufficiently adapt to this situation. This adaptation probably occurs both in the liver and in the intestine.
302
SEMINARS IN LIVER DISEASE—VOL. 20, NO. 3, 2000
LIPID HOMEOSTASIS DURING
CHOLESTASIS
Whitington et al.87,88 performed a study into the
clinical, histologic and biochemical features of patients
with PFIC. At that time, this group was not divided in
subtypes, because the genetic background of these types
was not yet known. However, because the serum GGT was normal in these patients, it may be assumed
that they were predominantly type 1 and type 2 patients,
that is, patients with mutations in the bile salt transporter BSEP (type 2) or mutations in the FIC-1 gene
(type 1; see Jansen, p. 245, this issue, for an extensive
review on this subject). As a consequence of their mutation, these two patient groups have a defect in biliary
bile salt secretion and secondary to this defect they also
have very low biliary lipid levels. PFIC patients appear
to differ from patients with obstructive cholestasis by a
normal plasma cholesterol level.87,88 In this study, a
group of Alagille patients, who have a congenital defect
in the formation of intrahepatic bile ducts, was used as a
control group with obstructive cholestasis. The latter
group of patients did suffer from a chronic hypercholesterolemia. This is also a common feature of adult patients with obstructive cholestasis. The plasma lipoprotein profile from such patients reveals an abnormal
lipoprotein fraction in the LDL region (when analyzed
by ultracentrifugation) or in the pre- region when analyzed by agar electrophoresis. This abnormal “cholestatic” lipoprotein was characterized as a unilamellar
vesicle with an aqeous lumen and was designated
lipoprotein X (LpX). The characteristic association of
the appearance of LpX in association with obstructive
cholestasis and the resemblance of this vesicular
lipoprotein with biliary lipid vesicles gave rise to the
hypothesis that this lipoprotein actually represents biliary vesicles that are regurgitated into the plasma in the
absence of bile flow. To shed light on the origin and the
nature of LpX, we analyzed the appearance of this
lipoprotein in bile duct ligated normal and mdr2/
mice. We observed that normal mice also develop hypercholesterolemia upon bile duct ligation, and this is
associated with the presence of massive amounts of
LpX in the plasma. Not only was this phenomenon completely absent in mdr2/ mice, but the extent of hypercholesterolemia during bile duct ligation was completely dependent on the expression level of mdr2
P-glycoprotein or the human orthologue MDR3 P-glycoprotein. This could be assessed by the use of a set of
transgenic mice with different expression levels of these
P-glycoproteins. Apparently, during cholestasis, the formation of biliary vesicles continues and the release of
these vesicles is redirected to the plasma compartment.
Indeed, it was shown that during bile duct ligation, the
expression of bsep and mdr2 P-glycoprotein are not
downregulated and that these proteins are redistributed
from the canalicular membrane to intracellular vesicles,
where they may continue to function in the formation of
biliary vesicles. At the same time, several studies89–92
demonstrated that cholesterol synthesis is increased
during cholestasis. Similarly, it has been reported that
phospholipid synthesis is also increased during
cholestasis.93 Hence, the formation and release of LpX,
which is poorly taken up by the liver again,94 may act as
a “pull” mechanism to withdraw cholesterol and phospholipid from the hepatic pool and thereby stimulate
their hepatic synthesis. LpX from the plasma seems to
be taken up by peripheral tissues via nonspecific fluid
phase endocytosis.94 In patients with chronic cholestasis
(PBC patients) this situation leads to the accumulation
of cholesterol in xanthomata.95
The conclusion from these data is that during obstructive cholestasis, the continued formation of lipid
vesicles by Mdr2/MDR3 P-glycoprotein and the release
into plasma leads to a dysregulation of cholesterol
homeostasis. In patients with specific defects in this
process (PFIC patients), this process is absent and therefore these patients have normal plasma cholesterol
levels.
CONCLUSIONS
Present data show that the process of biliary lipid secretion is complex and that several key steps are yet
poorly understood. The function of the lipid translocating MDR3 P-glycoprotein (and mdr2 P-glycoprotein in
the mouse) plays a rate-controlling step in this process,
but additional proteins may be found that are indispensable. This may be particularly be the case because the
process involves solubilization of lipids from the membrane, which, if not controlled tightly, will lead to leakage and thus to cell death. By driving the enterohepatic
cycle of cholesterol and phospholipid, biliary lipid secretion is heavily integrated in whole body lipid homeostasis. However, its absence does not seem to have dramatic
consequence for proper lipid homeostasis, suggesting
that the body can adapt quite well to this situation.
ABBREVIATIONS
ABC
AKR
BSEP
CDP
CETP
CT
HDL
LDL
LLC-PK
LpX
ATP-binding cassette
bile salt export protein
cytidine diphosphate
cholesterol ester transfer protein
CTP:phosphocholine cytidylyltransferase
high density lipotrotein
low density lipoprotein
porcine kidney epithelial cells
lipoprotein X
BILIARY LIPID SECRETION—OUDE ELFERINK, GROEN
MDR
MRP
PC
PC-TP
PFIC
Pgp
SCP-2
SKL
SR-BI
SREBP
TUDC
VLDL
multidrug resistance
multidrug resistance protein
phosphatidylcholine
phosphatidylcholine transfer protein
progressive familial intrahepatic
cholestasis
P-glycoprotein
sterol carrier protein 2
serine-lysine-leucine
scavenger receptor BI
sterol regulatory element binding protein
tauroursodeoxycholate
very low density lipoprotein
REFERENCES
1. Cooper A. Role of the enterohepatic circulation of bile salts in
lipoprotein metabolism. In: Ad C, ed. Bile salts: metabolic,
pathologic and therapeutic considerations. Philadelphia: W.B.
Saunders, 1999;211–229
2. Stone BG, Eridison SK, Graig WY, et al. Regulation of rat biliary cholesterol secretion by agents that alter intrahepatic cholesterol metabolism. J Clin Invest 1985;76:1773–1781
3. Nervi F, Marinovic I, Rigotti A, et al. Regulation of biliary cholesterol secretion. Functional relationship between the canalicular and sinusoidal cholesterol secretory pathways in the rat. J
Clin Invest 1988;82:1818–1825
4. Smit MJ, Temmerman AM, Wolters H, et al. Dietary fish oilinduced changes in intrahepatic cholesterol transport and bile
acid synthesis in rats. J Clin Invest 1991;88:943–951
5. Smit MJ, Verkade HJ, Havinga R, et al. Dietary fish oil potentiates bile acid-induced cholesterol secretion into bile in rats. J
Lipid Res 1994;35:301–310
6. Robins SJ, Fasulo JM, Leduc R, et al. The transport of lipoprotein cholesterol into bile: A reassessment of kinetic studies in the
experimental animal. Biochim Biophys Acta 1989;1004:
327–331
7. Portal I, Clerc T, Sbarra V, et al. Importance of high-density
lipoprotein-phosphatidylcholine in secretion of phospholipid
and cholesterol in bile. Am J Physiol 1993;264:G1052–G1056
8. Robins SJ, Fasulo JM. High density lipoproteins, but not other
lipoproteins, provide a vehicle for sterol transport to bile. J Clin
Invest 1997;99:380–384
9. Brown MS, Goldstein JL. A receptor-mediated pathway for cholesterol homeostasis. Science 1986;232:34–47
10. Havel RJ, Hamilton RL. Hepatocytic lipoprotein receptors and intracellular lipoprotein catabolism. Hepatology 1988;8:1689–1704
11. Glass C, Pittman RC, Weinstein DB, et al. Dissociation of tissue
uptake of cholesterol ester from that of apoprotein A-1 of rat
plasma high density lipoprotein: Selective delivery of cholesterol ester to liver, adrenal, and gonad. Proc Nat Acad Sci USA
1983;80:5435–5439
12. Glass C, Pittman RC, Civen M, et al. Uptake of high-density
lipoprotein-associated apoprotein A-1 and cholesterol esters by
16 tissues of the rat in vivo and by adrenal cells and hepatocytes
in vitro. J Biol Chem 1985;260:744–750
13. Acton S, Rigotti A, Landschulz KT, et al. Identification of scavenger receptor sr-bi as a high density lipoprotein receptor [see
comments]. Science 1996;271:518–520
14. Rinninger F, Brundert M, Jackle S, et al. Selective uptake of
high-density lipoprotein-associated cholesteryl esters by human
hepatocytes in primary culture. Hepatology 1994;19:1100–
1114
303
15. Trigatti B, Rayburn H, Vinals M, et al. Influence of the high density lipoprotein receptor SR-BI on reproductive and cardiovascular pathophysiology. Proc Natl Acad Sci USA 1999;96:
9322–9327
16. Acton S, Osgood D, Donoghue M, et al. Association of polymorphisms at the SR-BI gene locus with plasma lipid levels and
body mass index in a white population. Arterioscler Thromb
Vasc Biol 1999;19:1734–1743
17. Arai T, Wang N, Bezouevski M, et al. Decreased atherosclerosis
in heterozygous low density lipoprotein receptor-deficient mice
expressing the scavenger receptor BI transgene. J Biol Chem
1999;274:2366–2371
18. Kozarsky KF, Donahee MH, Rigotti A, et al. Overexpression of
the HDL receptor SR-BI alters plasma HDL and bile cholesterol
levels. Nature 1997;387:414–417
19. Sehayek E, Ono JG, Shefer S, et al. Biliary cholesterol excretion: A novel mechanism that regulates dietary cholesterol absorption. Proc Natl Acad Sci USA 1998;95:10194–10199
20. Empen K, Lange K, Stange EF, et al. Newly synthesized cholesterol in human bile and plasma: Quantitation by mass isotopomer distribution analysis. Am J Physiol 1997;272:
G367–G373
21. Kent C. CTP:phosphocholine cytidylyltransferase. Biochim
Biophys Acta 1997;1348:79–90
22. Vance DE, Walkey CJ, Cui Z. Phosphatidylethanolamine Nmethyltransferase from liver. Biochim Biophys Acta 1997;1348:
142–150
23. Cui Z, Vance JE, Chen MH, et al. Cloning and expression of a
novel phosphatidylethanolamine N-methyltransferase. A specific biochemical and cytological marker for a unique membrane
fraction in rat liver. J Biol Chem 1993;268:16655–16663
24. Sundler R, Akesson B. Regulation of phospholipid biosynthesis
in isolated rat hepatocytes. Effect of different substrates. J Biol
Chem 1975;250:3359–3367
25. Walkey CJ, Donohue LR, Bronson R, et al. Disruption of the
murine gene encoding phosphatidylethanolamine N-methyltransferase. Proc Natl Acad Sci USA 1997;94:12880–12885
26. Walkey CJ, Yu L, Agellon LB, et al. Biochemical and evolutionary significance of phospholipid methylation. J Biol Chem
1998;273–27043–27046
27. Patton GM, Fasulo JM, Robins SJ. Hepatic phosphatidylcholines: Evidence for synthesis in the rat by extensive reutilization of endogenous acylglycerides. J Lipid Res 1994;35:
1211–1221
28. Coleman R, Rahman K. Lipid flow in bile formation. Biochim
Biophys Acta 1992;1125:113–133
29. Mazer NA, Carey MC. Mathematical model of biliary lipid secretion: A quantitative analysis of physiological and biochemical
data from man and other species. J Lipid Res 1984;25:932–953
30. Elferink RP, Tytgat GN, Groen AK. Hepatic canalicular membrane 1: The role of mdr2 P-glycoprotein in hepatobiliary lipid
transport. FASEB J 1997;11:19–28
31. Coleman R, Rahman K, Kan KS, et al. Retrograde intrabiliary
injection of amphipathic materials causes phospholipid secretion
into bile. Biochem J 1989;258:17–22
32. Evens WH. A biochemical dissection of the functional polarity
of the plasma membrane of the hepatocyte. Biochim Biophys
Acta 1980;604:27–64
33. Yousef IM, Barnwell S, Gratton F, et al. Liver cell membrane
solubilization may control maximum secretory rate of cholic
acid in the rat. Am J Physiol 1987;252:G84–G91.
34. Smit JJM, Schinkel AH, Oude Elferink RPJ, et al. Homozygous
disruption of the murine mdr2 P-glycoprotein gene leads to a
complete absence of phospholipid from bile and to liver disease.
Cell 1993;75:451–462
35. Oude Elferink RPJ, Smit JJM, Schinkel AH, et al. The physiological role of mdr2 P-glycoprotein in hepatobiliary phospho-
304
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
SEMINARS IN LIVER DISEASE—VOL. 20, NO. 3, 2000
lipid transport. In: Keppler D, Jungermann K, eds. Transport in
the liver. Dordrecht: Kluwer, 1994;204–213
Oude Elferink RPJ, Ottenhoff R, van Wijland M, et al. Uncoupling of biliary phospholipid and cholesterol secretion in mice
with reduced expression of mdr2 P-glycoprotein. J Lipid Res
1996;37:1065–1075
Somjen GJ, Gilat T. A non-micellar mode of cholesterol transport in human bile. FEBS Lett 1983;61:265–268
Pattinson NR. Solubilisation of cholesterol in human bile. FEBS
Lett 1985;181:339–342
Lee SP, Park HZ, Madani H, et al. Partial characterization of a
nonmicellar system of cholesterol solubilization in bile. Am J
Physiol 1987;252:G374–G384
Ulloa N, Garrido J, Nervi F. Ultracentrifugal isolation of vesicular carriers of biliary chlesterol in native human and rat bile. Hepatology 1987;7:235–244
Crawford JM, Möckel G-M, Crawford AR, et al. Imaging biliary
lipid secretion in the rat: Ultrastructural evidence for vesiculation of the hepatocyte canalicular membrane. J Lipid Res
1995;36:2147–2163
Crawford AR, Smith AJ, Hatch VC, et al. Hepatic secretion of
phospholipid vesicles in the mouse critically depends on mdr2 or
MDR3 P-glycoprotein expression. Visualization by electron microscopy. J Clin Invest 1997;100:2562–2567
Ruetz S, Gros P. Phosphatidylcholine translocase: A physiological role for the mdr2 gene. Cell 1994;77:1071–1082
Smith AJ, Timmermans-Hereijgers JLPM, Roelofsen B, et al.
The human MDR3 P-glycoprotein promotes translocation of
phosphatidylcholine through the plasma membrane of fibroblasts from transgenic mice. FEBS Lett 1994;354:263–266
van Helvoort A, Smith AJ, Sprong H, et al. MDR1 P-glycoprotein
is a lipid translocase of broad specificity, while MDR3 P-gp specifically translocates phosphatidylcholine. Cell 1996;87:507–517
Nies AT, Gatmaitan Z, Arias IM. Atp-dependent phosphatidylcholine translocation in rat liver canalicular plasma membrane
vesicles. J Lipid Res 1996;37:1125–1136
Velardi ALM, Groen AK, Oude Elferink RPJ, et al. Cell typedependent effect of phospholipid and cholesterol on bile salt cytotoxicity. Gastroenterology 1991;101:457–464
Eckhardt ERM, Moschetta A, Renooij W, et al. Asymmetric distribution of phosphatidylcholine and sphingomyelin between
micellar and vesicular phases: Potential implications for canalicular bile formation. J Lipid Res 1999;40:2022–2033
Demel RA, Jansen JW, van Dijck PW, et al. The preferential interaction of cholesterol with different classes of phospholipids.
Biochim Biophys Acta 1977;465:1–10
Slotte JP. Lateral domain formation in mixed monolayers containing cholesterol and dipalmitoylphosphatidylcholine or Npalmitoylsphingomyelin. Biochim Biophys Acta 1995;1235:
419–427
Mattjus P, Bittman R, Vilcheze C, et al. Lateral domain formation in cholesterol/phospholipid monolayers as affected by the
sterol side chain conformation. Biochim Biophys Acta
1995;1240:237–247
Evans W, Kremmer T, Culvenor J. Role of membranes in bile
formation. Comparison of the composition of bile and a livercanalicular plasma membrane subfraction. Biochem J 1976;
154:589–595
Meier PJ, Sztul ES, Reuben A, et al. Structural and functional
polarity of canalicular and basolateral plasma membrane vesicles isolated in high yield from rat liver. J Cell Biol 1984;
98:991–1000
Nibbering CP, Carey MC. Sphingomyelins of rat liver: Biliary
enrichment with molecular species containing 16:0 fatty acids as
compared to canalicular-enriched plasma membranes. J Membr
Biol 1999;167:165–171
Cohen DE, Leonard MR, Carey MC. In vitro evidence that phos-
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
pholipid secretion into bile may be coordinated intracellularly
by the combined actions of bile salts and the specific phosphatidylcholine transfer protein of liver. Biochemistry 1994;33:
9975–9980
Cohen DE, Green RM. Cloning and characterization of a cDNA
encoding the specific phosphatidylcholine transfer protein from
bovine liver. Gene 1995;163:327–328
Wirtz KWA, Kamp HH, van Deenen LLM. Isolation of a protein
from beef liver which specifically stimulates the exchange of
phosphatidylcholine. Biochim Biophys Acta 1972;274:606–617
Wirtz KWA. Phospholipid transfer proteins revisited [review].
Biochem J 1997;324:353–360
LaMorte WW, Booker ML, Kay S. Determinants of the selection
of phosphatidylcholine molecular species of secretion into bile
in the rat. Hepatology 1998;28:631–637
van Helvoort A, de Brouwer A, Ottenhoff R, et al. Mice without
phosphatidylcholine transfer protein have no defects in the secretion of phosphatidylcholine into bile or into lung airspaces.
Proc Natl Acad Sci USA 1999;96:11501–11506
Noland BJ, Arebalo RE, Hansbury E, et al. Purification ad properties of sterol carrier protein 2. J Biol Chem 1980;255:
4282–4289
Bloj B, Zilversmit DB. Rat liver proteins capable of transferring
phosphatidylethanollamine. Purification and transfer activity for
other phospholipids and cholesterol. J Biol Chem 1977;252:
1613–1619
Stolowich NJ, Frolov A, Atshaves B, et al. The sterol carrier
protein-2 fatty acid binding site: An NMR, circular dichroic, and
fluorescence spectroscopic determination. Biochemistry 1997;
36:1719–1729
Seedorf U, Raabe M, Ellinghaus P, et al. Defective peroxisomal
catabolism of branched fatty acyl coenzyme A in mice lacking
the sterol carrier protein-2/sterol carrier protein-x gene function.
Genes Dev 1998;12:1189–1201
Keller GA, Scallen TJ, Clarke D, et al. Subcellular localization
of sterol carrier protein-2 in rat hepatocytes: Its primary localization to peroxisomes. J Cell Biol 1989;108:1353–1361
Ossendorp BC, Wirtz KW. The non-specific lipid-transfer protein (sterol carrier protein 2) and its relationship to peroxisomes.
Biochimie 1993;75:191–200
Fuchs M, Lammert F, Wang DQ, et al. Sterol carrier protein 2
participates in hypersecretion of biliary cholesterol during gallstone formation in genetically gallstone-susceptible mice.
Biochem J 1998;336:33–37
Puglielli L, Rigotti A, Amigo L, et al. Modulation of intrahepatic
cholesterol trafficking—evidence by in vivo antisense treatment
for the involvement of sterol carrier protein-2 in newly synthesized cholesterol transport into rat bile. Biochem J 1996;
317:681–687
Crawford JM, Berken CA, Gollan JL. Role of the hepatocyte microtubular system in the excretion of bile salts and biliary lipid:
Implications for intracellular vesicular transport. J Lipid Res
1988;29:144–156
Casu A, Camogliano L. Glycerophospholipids and cholesterol
composition of bile in bile-fistula rats treated with monensin.
Biochim Biophys Acta 1990;1043:113–115
Reynier MO, Abouhashieh I, Crotte C, et al. Monensin action on
the Golgi complex in perfused rat liver evidence against bile salt
vesicular transport. Gastroenterology 1992;102:2024–2032
Van Meer G, Simons K. The function of tight junctions in maintaining differences in lipid composition between the apical and
the basolateral cell surface domains of MDCK cells. EMBO J
1986;5:1455–1464
Robins SJ, Fasulo JM. Delineation of a novel hepatic route for
the selective transfer of unesterified sterols from high-density
lipoproteins to bile: Studies using the perfused rat liver. Hepatology 1999;29:1541–1548
BILIARY LIPID SECRETION—OUDE ELFERINK, GROEN
74. Zhang X, Collins KI, Greenberger LM. Functional evidence that
transmembrane 12 and the loop between transmembrane 11 and
12 form part of the drug-binding domain in P-glycoprotein encoded by MDR1. J Biol Chem 1995;270:5441–5448.
75. Currier SJ, Kane SE, Willingham MC, et al. Identification of
residues in the first cytoplasmic loop of P-glycoprotein involved
in the function of chimeric human MDR1-MDR2 transporters. J
Biol Chem 1992;267:25153–25159
76. Buschman E, Gros P. Functional analysis of chimeric genes obtained by exchanging homologous domains of the mouse mdr1
and mdr2 genes. Mol Cell Biol 1991;11:595–603
77. Zhou Y, Gottesman MM, Pastan I. Domain exchangeability between the multidrug transporter (MDR1) and phosphatidylcholine flippase (MDR2). Mol Pharmacol 1999;56:997–1004
78. Zhou Y, Gottesman MM, Pastan I. Studies of human MDR1MDR2 chimeras demonstrate the functional exchangeability of a
major transmembrane segment of the multidrug transporter and
phosphatidylcholine flippase. Mol Cell Biol 1999;19:
1450–1459
79. Smith AJ. The substrate specificites of the human MDR1 and
MDR3 P-glycoproteins. Amsterdam: Netherlands Cancer Inst,
1998:133
80. Frijters CMG, Ottenyhoff R, Vanwijland MJA, et al. Regulation
of mdr2 p-glycoprotein expression by bile salts. Biochem J
1997;321:389–395
81. Carrella M, Feldman D, Cogoi S, et al. Enhancement of mdr2
gene transcription mediates the biliary transfer of phosphatidylcholine supplied by an increased biosynthesis in the pravastatintreated rat. Hepatology 1999;29:1825–1832
82. Hooiveld GJEJ, Vos TA, Scheffer GL, et al. 3-hydroxy-3methylglutaryl-coenzyme a reductase inhibitors (statins) induce
haptic expression of the phospholipid translocase mdr2 in rats.
Gastroenterology 1999;117:678–687
83. Grundy SM. Absorption and metabolism of dietary cholesterol.
Annu Rev Nutr 1983;3:71–96
84. Westergaard H, Dietschy JM. The mechanism whereby bile acid
micelles increase the rate of fatty acid and cholesterol uptake
into the intestinal mucosal cell. J Clin Invest 1976;58:97–108
305
85. Voshol PJ, Havinga R, Wolters H, et al. Reduced plasma cholesterol and increased fecal sterol loss in multidrug resistance gene
2 p-glycoprotein-deficient mice. Gastroenterology 1998;114:
1024–1034
86. Voshol PJ, Minich DM, Havinga R, et al. Postprandial chylomicron formation and fat absorption in multidrug resistance gene 2
P-glycoprotein-deficient mice. Gastroenterology 2000;118:
173–182
87. Whitington PF, Freese DK, Alonso EM, et al. Clinical and biochemical findings in progressive familial intrahepatic cholestasis. J Pediatr Gastroenterol Nutr 1994;18:134–141
88. Alonso EM, Snover DC, Montag A, et al. Histologic pathology
of the liver in progressive familial intrahepatic cholestasis. J Pediatr Gastroenterol Nutr 1994;18:128–133
89. Kattermann R, Creutzfeldt W. The effect of experimental
cholestasis on the negative feedback regulation of cholesterol
synthesis in rat liver. Scand J Gastroenterol 1970;5:337–342
90. Harry DS, Dini M, McIntyre N. Effect of cholesterol feeding and
biliary obstruction on hepatic cholesterol biosynthesis in the rat.
Biochim Biophys Acta 1973;296:209–220
91. Cooper AD, Ockner RD. Studies of hepatic cholesterol synthesis
in experimental acute biliary obstruction. Gastroenterology
1974;66:586–595
92. Dueland S, Reichen J, Everson GT, et al. Regulation of cholesterol and bile acid homoeostasis in bile-obstructed rats. Biochem
J 1991;280:373–377
93. McIntyre N, Harry DS, Pearson AJ. The hypercholesterolaemia
of obstructive jaundice. Gut 1975;16:379–391
94. Walli AK, Seidel D. Role of lipoprotein-X in the pathogenesis of
cholestatic hypercholesterolemia. Uptake of lipoprotein-X and
its effect on 3-hydroxy-3-methylglutaryl coenzyme A reductase
and chylomicron remnant removal in human fibroblasts, lymphocytes, and in the rat. J Clin Invest 1984;74:867–879
95. Ahrens EH, Kunkel HG. The relationship between serum lipids
and skin xanthomata in eighteen patients with primary biliary
cirrhosis. J Clin Invest 1949;28:1565–1574