Tài liệu Báo cáo khoa học: Hypoxia downregulates farnesoid X receptor via a hypoxia-inducible factor-independent but p38 mitogen-activated protein kinase-dependent pathway doc

Hypoxia downregulates farnesoid X receptor via a
hypoxia-inducible factor-independent but p38
mitogen-activated protein kinase-dependent pathway
Tomofumi Fujino
1
, Kaori Murakami
1
, Issei Ozawa
1
, Yoshie Minegishi
1
, Ryo Kashimura
1
, Toshihiro
Akita
1
, Susumu Saitou
1
, Takehisa Atsumi
1
, Takashi Sato
1
, Ken Ando
1
, Shuntaro Hara
2
, Kiyomi
Kikugawa
1
and Makio Hayakawa
1

1 School of Pharmacy, Tokyo University of Pharmacy and Life Science, Japan
2 School of Pharmaceutical Sciences, Showa University, Tokyo, Japan
Nuclear receptors (NRs) are ligand-activated transcrip-
tional factors that belong to a large superfamily
consisting of over 150 different members [1]. Farne-
soid X receptor (FXR), a member of the NR super-
family, was first isolated from a rat liver cDNA
library, and named after its weak activation by supra-
physiological concentrations of farnesol, an intermedi-
ate in the mevalonate biosynthetic pathway [2].
Subsequent studies have revealed that certain types of
bile acids act as physiological ligands for FXR, leading
to its activation [3–5].
FXR expression is limited to very few tissues; it is
highly expressed in the liver, intestine, kidney, and
adrenal gland [2,6–8], whereas lower levels of expres-
sion are reported in adipose tissues and heart [9–11].
FXR forms a heterodimer with retinoid X receptor
Keywords
cholestasis; farnesoid X receptor; hypoxia-
inducible factor; ischemia; mitogen-activated
protein kinase
Correspondence
M. Hayakawa, School of Pharmacy, Tokyo
University of Pharmacy and Life Science,
1432-1 Horinouchi, Hachioji, Tokyo
192-0392, Japan
Fax: +81 42 676 4508
Tel: +81 42 676 4513
E-mail:

(Received 18 September 2008, revised 28
November 2008, accepted 19 December
2008)
doi:10.1111/j.1742-4658.2009.06867.x
Farnesoid X receptor (FXR), a member of the nuclear receptor superfam-
ily, has been shown to play pivotal roles in bile acid homeostasis by regu-
lating the biosynthesis, conjugation, secretion and absorption of bile acids.
Accumulating data suggest that FXR signaling is involved in the pathogen-
esis of liver and metabolic disorders. Here we show that FXR expression is
significantly suppressed in HepG2 cells exposed to hypoxia. Concomitantly,
the expression of the bile salt export pump, known as an FXR target gene
product and responsible for the excretion of bile acids from the liver, is
also decreased under hypoxia. Overexpression of hypoxia-inducible factor
(HIF)-1a does not mimic the suppressive effect of hypoxia on FXR expres-
sion. Furthermore, simultaneous knockdown of HIF-1a, HIF-2a and
HIF-3a fails to restore the FXR expression level under hypoxia, indicating
that HIF is not involved in hypoxia-evoked FXR downregulation. Instead,
we demonstrate that p38 mitogen-activated protein kinase is an indispens-
able factor for FXR downregulation under hypoxia. Thus, we propose a
novel liver disorder model in which two signaling molecules, p38 mitogen-
activated protein kinase and FXR, may contribute to the linkage of two
pathogenic conditions, i.e. ischemia, a condition accompanying hypoxia,
and cholestasis, a condition with intrahepatic accumulation of cytotoxic
bile acids.
Abbreviations
BSEP, bile salt export pump; CDCA, chenodeoxycholic acid; ERK, extracellular signal-regulated kinase; FXR, farnesoid X receptor; FXRE,
farnesoid X receptor response element; GLUT-1, glucose transporter-1; HCC, hepatocellular carcinoma; HIF, hypoxia-inducible factor; IL,
interleukin; JNK, c-Jun N-terminal kinase; LRH-1, liver receptor homolog 1; MAPK, mitogen-activated protein kinase; MRP, multidrug
resistant-associated protein; NF-jB, nuclear factor kappaB; NR, nuclear receptor; SD, standard deviation; SHP, small heterodimer partner;
siRNA, small interfering RNA; TNF, tumor necrosis factor; VEGF, vascular endothelial growth factor.

FEBS Journal 276 (2009) 1319–1332 ª 2009 The Authors Journal compilation ª 2009 FEBS 1319
and binds to specific response elements (FXREs) on
the promoters of its targeted genes [12,13]. Upon
ligand binding, FXR undergoes conformational
changes to release corepressors and recruit coactiva-
tors, resulting in the induction of target gene expres-
sion [12,13].
Activation of hepatic FXR in response to bile acids
increases the expression of the FXR target gene, small
heterodimer partner (SHP). SHP in turn binds to and
inactivates liver receptor homolog 1 (LRH-1), a tran-
scription factor that is critical for CYP7A1 expression.
Thus, FXR activation leads to the repression of
CYP7A1 transcription in an indirect manner [13].
Activated FXR also increases the transcription of three
hepatic transporters that function to export bile acids
out of hepatocytes; the three proteins bile salt export
pump (BSEP), multidrug resistant-associated protein
(MRP)2 (ABCC2) and multidrug resistance P-glyco-
protein 3 (MDR3, ABCB4) are localized on the mem-
brane of hepatocytes, and secrete bile acids from the
hepatocytes into the bile canaliculi [13]. Bile, containing
bile acids, phospholipids, cholesterol, and proteins, is
then stored in the gall bladder. Bile acids excreted into
bile undergo intensive enterohepatic circulation. They
are reabsorbed in the intestine, taken up again by the
liver, and re-excreted into the bile, with this cycle being
repeated several times before they are eliminated with
the feces [14]. In enterocytes, FXR has been shown to
regulate apical sodium-dependent bile acid transporter

and ileum bile acid-binding protein [13]. Thus, it is
clear that FXR is a key sensor for bile acids and has a
central role in maintaining bile acid homeostasis.
Bile acid homeostasis is the result of a balance
between bile acid uptake, efflux, and biosynthesis.
Maintenance of this balance is essential, as most bile
acids are cytotoxic. Cholestasis, a medical condition
characterized by the impairment of normal bile flow,
results in intrahepatic accumulation of cytotoxic bile
acids, which cause liver injury, ultimately leading to
biliary fibrosis and cirrhosis [14]. As the expression of
transporters responsible for bile acid export at the can-
alicular membrane are regulated by FXR, FXR has
been thought to be a possible pharmaceutical target
for the treatment of cholestasis [13].
In different types of clinical situations, liver ischemia
may occur and cause or contribute to hepatobiliary
dysfunction, which is most often of the cholestatic type
[15,16]. Several lines of evidence demonstrate that the
development of biliary cirrhosis is associated with the
occurrence of hepatocellular hypoxia and the induction
of hepatic angiogenesis [17–19]. Low levels of O
2
(hypoxia) are encountered by cells within rapidly
growing tissues, such as developing embryos or solid
tumors. Most vertebrates respond to this hypoxic
stress by activating the expression of a large number
of genes involved in glycolysis, angiogenesis, and
hematopoiesis [20]. This hypoxic transcriptional
response is mediated primarily by hypoxia-inducible

factor (HIF), a key transcriptional regulator composed
of an oxygen-regulated HIF-a subunit and the ubiqui-
tous HIF-b (also called the aryl hydrocarbon receptor
nuclear translocator) partner protein [20]. HIF-a pro-
tein turnover in normoxia is very rapid, owing to the
inhibitory action of the HIF-a prolyl hydroxylases.
These oxygen-dependent enzymes hydroxylate two con-
served prolyl residues within a central oxygen-depen-
dent degradation domain of the HIF-a proteins,
promoting binding of the von Hippel–Lindau tumor
suppressor protein, which acts as the ubiquitin ligase
for HIF-a proteins, leading to the degradation by the
proteasome [20]. Any HIF-a escaping this normoxic
degradation is also subjected to hydroxylation of a
conserved asparagine residue within the C-terminal
transactivating domain, which represses activity via
abrogation of CREB-binding protein (CBP) ⁄ p300 coac-
tivator recruitment [21]. There are three HIF-a iso-
forms, i.e. HIF-1a, HIF-2a, and HIF-3a. HIF-2a is
very similar to HIF-1a in both structure and function,
but exhibits more restricted tissue-specific expression,
and may also be differentially regulated by nuclear
translocation [20]. HIF-3a also exhibits conservation
with HIF-1a and HIF-2a in the oxygen-dependent deg-
radation domain; however, in contrast to HIF-1a and
HIF-2a, HIF-3a does not possess a hypoxia-inducible
transactivation domain, instead, having a novel C-ter-
minus with additional uncharacterized transactivation
properties [22]. Several researchers have already
reported that hepatocellular hypoxia causes liver angio-

genesis and fibrogenesis through the inducible expres-
sion of vascular endothelial growth factor (VEGF), one
of the most representative HIF target gene products
[17–19]. In a recent report, Fouassier et al. demon-
strated that the expression of BSEP and FXR was
impaired in the ischemic rat liver or cultured hepato-
cytes exposed to hypoxia, whereas VEGF expression
was elevated under the same conditions [23]. However,
the molecular mechanism by which hepatocellular
hypoxia caused the reduced expression of the FXR and
BSEP genes has not been elucidated yet.
Bile acids are now recognized as important regula-
tory molecules, not only for their own synthesis, but
also for cholesterol synthesis, gluconeogenesis, glyco-
genesis, and apoptosis [24–26]. Among various signal-
ing pathways, mitogen-activated protein kinases
(MAPKs) play important roles in transducing or
modulating the bile acid-regulated cellular responses.
Hypoxia downregulates FXR via p38 MAPK T. Fujino et al.
1320 FEBS Journal 276 (2009) 1319–1332 ª 2009 The Authors Journal compilation ª 2009 FEBS
In mammals, four distinct subgroups of MAPKs have
been identified [27]. These include: (a) extracellular sig-
nal-regulated kinases (ERKs); (b) c-Jun N-terminal
kinases (JNKs); (c) p38 group MAPKs; and (d)
ERK5 ⁄ big MAP kinase 1 [27]. JNKs act as negative
regulators of bile acid synthesis by repressing CYP7A1
expression in SHP-independent ways [14]. Bile acid-
activated FXR induces the expression of the FGF19
gene, resulting in the suppression of CYP7A1 through
a JNK-dependent pathway [28]. Alternatively, JNK

activation induced by protein kinase C [29] or tumor
necrosis factor (TNF) [30] also leads to the suppression
of CYP7A1 expression. Bile acids are also known to
induce oxidative stress in mitochondria, leading to the
activation of ERKs through the inactivation of protein
tyrosine phosphatases [31]. In contrast to JNK, p38
MAPK is reported to activate bile acid synthesis
through the induction of CYP7A1 expression [32].
Thus, the individual MAPK family members play dif-
ferent roles in maintaining bile acid homeostasis, and
their imbalanced activation may cause impaired bile
acid metabolism, leading to cholestasis-mediated
cirrhosis.
In this study, we have demonstrated that FXR is
significantly downregulated in HepG2 cells exposed to
hypoxia. Interestingly, HIFs, known as the master reg-
ulators in hypoxic responses, do not participate in the
hypoxia-induced downregulation of FXR. In contrast,
p38 MAPK, which is also known to be activated in
response to hypoxia [33–35], is responsible for the
downregulation of FXR under hypoxic conditions.
Results
Cholestatic disorder is known to arise from liver ische-
mia [15,16]. As FXR, a member of the NR super-
family, plays a key role in maintaining bile acid
metabolism, we first examined whether FXR activity is
impaired in cultured hepatocellular cells exposed to
hypoxia. Whereas FXRE-driven luciferase activity was
significantly increased in HepG2 cells treated with the
FXR ligand chenodeoxycholic acid (CDCA) under

normoxia, cells cultured under hypoxia exhibited mar-
ginal activation induced by CDCA (Fig. 1A). Similar
results were obtained when Huh7 cells were used
instead of HepG2 cells (Fig. 1B).
In order to verify whether the expression level of
BSEP, a target gene of FXR, is indeed lowered under
hypoxia, BSEP mRNA levels in cells under normoxia
or hypoxia were compared. As shown in Fig. 1C,
CDCA-induced elevation of BSEP mRNA was clearly
demonstrated under normoxia. In contrast, BSEP
induction by CDCA was greatly reduced under
hypoxia, indicating that hypoxia impaired the activity
of FXR, resulting in the lowered expression of BSEP.
Under the same hypoxic conditions, the level of
nuclear factor kappaB (NF-jB) activation induced by
TNF was comparable to that in cells under normoxia
(Fig. 1D), indicating that HepG2 cells exposed to
hypoxia maintained the physiological response in terms
of NF-jB activation. Furthermore, when viable cell
numbers were measured by 3-(4,5-dimethylthiazol-2-
yl)-2,5-diphenyl-tetrazolium bromide assay, there was
no difference between hypoxia and normoxia (data not
shown). These results suggest that the hypoxia-induced
downregulation of FXR activity is a physiologically
regulated cellular response rather than a nonspecific
result reflecting lowered cell viability.
Next we examined whether or not FXR expression
itself is lowered in cells exposed to hypoxia. As shown
in Fig. 2A, the level of FXR protein detected by
antibody against FXR was significantly decreased in

cells exposed to hypoxia as compared with that in
cells under normoxia, where comparable levels of
actin expression were observed under hypoxia and
normoxia.
By measurement of the level of FXR mRNA, the
lowered expression was also confirmed under hypoxia
(Fig. 2B). Besides FXR, LRH-1 is also known to regu-
late BSEP expression [36]. However, under the same
hypoxic conditions where the expression level of FXR
was significantly lowered, LRH-1 expression was not
changed at all (Fig. 2C), suggesting that FXR but not
LRH-1 is involved in the downregulation of BSEP
under hypoxia. It should be noted that the expression
level of SHP, another FXR target gene, was also
lowered under the same hypoxic conditions (Fig. 2D).
The lowered FXR mRNA level may reflect two pos-
sibilities: one is the decreased stability of FXR mRNA
under hypoxia, and the other is the decreased tran-
scription of the FXR gene in response to hypoxia.
Therefore, we compared FXR mRNA stability under
hypoxia with that under normoxia by treating HepG2
cells with actinomycin D. As shown in Fig. 2E, there
was no difference in the kinetics of FXR mRNA deg-
radation between hypoxia and normoxia, suggesting
that hypoxia downregulates FXR by lowering its
transcription.
As the master regulator of hypoxic responses, HIF, a
heterodimeric transcriptional factor consisting of HIF-a
and HIF-b subunits, plays important roles by inducing
the expression of genes required for survival of cells

exposed to hypoxic environments [37]. To elucidate
whether or not HIF is involved in the hypoxia-evoked
downregulation of FXR expression, we examined the
effect of ectopically overexpressed HIF-a isoforms on
T. Fujino et al. Hypoxia downregulates FXR via p38 MAPK
FEBS Journal 276 (2009) 1319–1332 ª 2009 The Authors Journal compilation ª 2009 FEBS 1321
FXR expression in HepG2 cells. HIF-a consists of three
isoforms, HIF-1a, HIF-2a, and HIF-3a [38]. Among
three isoforms, HIF-1a is ubiquitously expressed and
has been suggested to play a primary role in hypoxic
responses. When Pro402, Pro564 and Asn803 are
replaced by alanine residues, the HIF-1a mutant
becomes constitutively active even under normoxia,
escaping from degradation through the ubiquitin–pro-
teasome system. Indeed, we detected a significant level
of constitutively active form of HIF-a (HIF-1a)CA
expression in HepG2 cells under normoxia (Fig. 3A,
top panel). For HIF-2a and HIF-3a, overexpression
was achieved by the use of wild-type cDNAs without
Pro ⁄ Ala substitutions (Fig. 3A, middle and bottom
panels). We first examined whether or not ectopically
overexpressed HIF-a isoforms do indeed function in
HepG2 cells under normoxia. As shown in Fig. 3B,
HIF-1a CA and HIF-3a significantly induced glucose
transporter-1 (GLUT-1), a well-known HIF target gene
[39], whereas ectopically overexpressed HIF-2a failed to
elevate its level. As HIF-a isoforms are known to func-
tion in cell type-specific or target gene-specific ways [39],
HIF-2a may not be active in terms of GLUT-1 regula-
tion in HepG2 cells. Although at least HIF-1a and

HIF-3a were shown to be functional, neither of them
lowered the FXR protein expression level (Fig. 3C).
Furthermore, when FXR mRNA levels were examined,
none of the HIF-a isoforms could mimic the effect
of hypoxia, whereas slight but significant increases in
FXR mRNA were observed in cells overexpressing
HIF-1aCA and HIF-3a (Fig. 3D).
To confirm that HIF is not involved in FXR down-
regulation under hypoxia, the expression level of FXR
was examined in HepG2 cells in which endogenous
HIF-a isoforms were simultaneously knocked down.
A
B
C
D
Fig. 1. Hypoxia-dependent suppression of FXR activity. (A) HepG2
cells were cotransfected with FXRE-driven luciferase reporter vec-
tor and R. reniformis luciferase expression vector phRL–TK. After
24 h, CDCA (100 l
M) or dimethylsulfoxide (DMSO) (vehicle control)
was added to the culture, and cells were further cultured under
normoxia or hypoxia for 24 h. Cellular FXR activities were then
determined as described in Experimental procedures. Results are
calculated as changes relative to the FXR activity in cells cultured
under normoxia in the absence of CDCA. Data are shown as the
mean ± standard deviation (SD) of four determinations. Similar
results were obtained in four separate experiments. (B) Huh7 cells
were were cotransfected with FXRE-driven luciferase reporter vec-
tor and R. reniformis luciferase expression vector phRL–TK, and
cellular FXR activities were then determined as in (A). (C) HepG2

cells were cultured under normoxia or hypoxia for 24 h in the pres-
ence of CDCA (100 l
M) or dimethylsulfoxide as vehicle control. The
amounts of BSEP mRNA were quantified by real-time PCR as
described in Experimental procedures. Results are calculated as
changes relative to the amount of BSEP mRNA from cells cultured
under normoxia in the absence of CDCA. Data are shown as the
mean ± SD of four determinations. Similar results were obtained in
five separate experiments. (D) HepG2 cells were cotransfected
with 3 · jB–Luc luciferase reporter vector and R. reniformis lucifer-
ase expression vector phRL–TK. After 24 h, cells were stimulated
with TNF and further cultured under normoxia or hypoxia for 24 h.
Results are calculated as changes relative to the NF-jB activity in
cells cultured under normoxia in the absence of TNF. Data are
shown as the mean ± SD of four determinations. Similar results
were obtained in three separate experiments.
Hypoxia downregulates FXR via p38 MAPK T. Fujino et al.
1322 FEBS Journal 276 (2009) 1319–1332 ª 2009 The Authors Journal compilation ª 2009 FEBS
As shown in Fig. 4A, the expression levels of endoge-
nous HIF-1a and HIF-2a were significantly elevated in
cells exposed to hypoxia. When cells were transfected
with the combined mixture of small interfering RNAs
(siRNAs) against HIF-1a, HIF-2a, and HIF-3a, the
levels of HIF-1a and HIF-2a in cells under hypoxia
were drastically decreased (Fig. 4A). In the case of
HIF-3a, endogenous HIF-3a was not significantly ele-
vated under hypoxia; therefore, knockdown efficiency
was evaluated using HepG2 cells that ectopically over-
expressed HIF-3a. As shown in Fig. 4B, the combined
transfection of three siRNAs against HIF-1a, HIF-2a

and HIF-3a also effectively reduced the expression
level of HIF-3a. However, under the condition where
three isoforms of HIF-a were significantly knocked
down, the lowered expression of FXR protein in cells
under hypoxia was not restored at all (Fig. 4C). Simi-
larly, at the level of mRNA, hypoxia-dependent down-
regulation of FXR was not restored by the triple
knockdown of three HIF-a isoforms (Fig. 4D). These
results clearly demonstrated that HIFs do not partici-
pate in the downregulation of FXR in cells under
hypoxia.
Next, we focused on MAPK families as candidates
for upstream signaling molecules responsible for the
downregulation of FXR in response to hypoxia, as bile
acid homeostasis is regulated in many aspects by each
subfamily of MAPKs [14,32] and, in addition, several
A
B
C
D
E
Fig. 2. Hypoxia suppresses the transcription of FXR. (A) HepG2
cells cultured under normoxia or hypoxia for 24 h were lysed in
SDS ⁄ PAGE sample buffer, and the resultant cell lysates were
subjected to immunoblotting analyses to detect FXR or to detect
b-actin levels as a loading control. Similar results were obtained in
three separate experiments. (B) HepG2 cells were cultured under
normoxia or hypoxia for 24 h, and total RNAs were then prepared.
The amounts of FXR mRNA were quantified by real-time PCR as
described in Experimental procedures. Results are calculated as

changes relative to the amount of FXR mRNA from cells cultured
under normoxia. Data are shown as the mean ± SD of four deter-
minations. Similar results were obtained in three separate experi-
ments. (C) HepG2 cells were cultured under normoxia or hypoxia
for 24 h, and the amounts of LRH-1 mRNA were quantified by real-
time PCR as in (B). Results are calculated as changes relative to
the amount of LRH-1 mRNA from cells cultured under normoxia.
Data are shown as the mean ± SD of four determinations. Similar
results were obtained in three separate experiments. (D) HepG2
cells were cultured under normoxia or hypoxia for 24 h, and the
amounts of SHP mRNA were quantified by real-time PCR as in (B).
Results are calculated as changes relative to the amount of SHP
mRNA from cells cultured under normoxia. Data are shown as the
mean ± SD of four determinations. Similar results were obtained in
three separate experiments. (E) HepG2 cells were treated with acti-
nomycin D (5 l
M) and then cultured under normoxia or hypoxia for
indicated times. The amounts of FXR mRNA were quantified by
real-time PCR as described in Experimental procedures. Results are
calculated as changes relative to the amount of FXR mRNA at 0 h.
Data are shown as the mean ± SD of four determinations. Similar
results were obtained in three separate experiments.
T. Fujino et al. Hypoxia downregulates FXR via p38 MAPK
FEBS Journal 276 (2009) 1319–1332 ª 2009 The Authors Journal compilation ª 2009 FEBS 1323
lines of evidence indicate that hypoxia induces the acti-
vation of MAPK families [33–35,40,41]. The activation
of each MAPK was examined by the use of antibodies
that detect phosphorylated forms of MAPKs (Fig. 5).
Before cells were exposed to hypoxia, the phosphory-
lated form of p38 was retained at a low level; however,

its level was significantly increased during hypoxia,
becoming more than 10 times higher after 8 h of
A
B
C
D
Fig. 3. Ectopically overexpressed HIF-a isoforms fail to suppress
FXR expression. (A) HepG2 cells were transfected with pcDNA3.1
as a vehicle control or with pcDNA3.1–HIF-1a CA, pcDNA3.1–HIF-
2a, or pcDNA3.1–HIF-3a. After 24 h, cell lysates were prepared,
and immunoblotting analyses to detect HIF-1a, HIF-2a and HIF-3a
then performed. (B) HepG2 cells were transfected with pcDNA3.1
as a vehicle control or with pcDNA3.1–HIF-1a CA, pcDNA3.1–HIF-
2a, or pcDNA3.1–HIF-3a. After 24 h, total RNAs were prepared,
and real-time PCR analysis to detect GLUT-1 mRNA was then per-
formed as described in Experimental procedures. (C) As described
in (A), cell lysates were prepared, and immunoblotting analyses
were then performed to detect FXR or to detect b-actin levels as a
loading control. (D) HepG2 cells were transfected with pcDNA3.1
as a vehicle control or with pcDNA3.1–HIF-1a CA, pcDNA3.1–
HIF-2a, or pcDNA3.1–HIF-3a. After 24 h, total RNAs were pre-
pared, and real-time PCR analysis was then performed to detect
FXR mRNA, as described in Experimental procedures.
A
B
C
D
Fig. 4. Knockdown of endogenous HIF isoforms does not restore
FXR expression under hypoxia. (A) HepG2 cells were transfected
with control siRNA (#1022076) or with the combined mixture of

siRNAs against HIF-1a, HIF-2a, and HIF-3a, and then cultured under
normoxia or hypoxia for 24 h. Cell lysates were then prepared, and
immunoblotting analyses were then performed to detect HIF-1a
and HIF-2a. (B) In order to estimate the knockdown efficiency of
HIF-3a, HepG2 cells transfected with pcDNA3.1–HIF-3a were then
treated with the combined mixture of siRNAs against HIF-1a,
HIF-2a, and HIF-3a, and then cultured under normoxia for 24 h. Cell
lysates were then prepared, and this was followed by immunoblot-
ting analysis to detect HIF-3a. (C) As described in (A), cell lysates
were prepared, and then immunoblotting analyses were performed
to detect FXR or to detect b-actin levels as a loading control. (D)
HepG2 cells transfected with the combined mixture of siRNAs
against HIF-1a, HIF-2a and HIF-3a were then cultured under
hypoxia for 24 h. After 24 h, total RNAs were prepared, and then
real-time PCR analysis was performed to detect FXR mRNA as
described in Experimental procedures.
Hypoxia downregulates FXR via p38 MAPK T. Fujino et al.
1324 FEBS Journal 276 (2009) 1319–1332 ª 2009 The Authors Journal compilation ª 2009 FEBS
hypoxia treatment (Fig. 5). On the other hand, phos-
phorylation of ERK2 was observed before hypoxia
treatment. Hypoxia treatment for 8 h resulted in a
two-fold elevation of phospho-ERK2 level, whereas
only slight activation of ERK1 was observed under
hypoxia (Fig. 5). During this hypoxia treatment, the
expression levels of FXR were inversely lowered in a
time-dependent manner (Fig. 5). It should be noted
that we could not detect the activation of JNK using
anti-phospho-JNK antibodies in HepG2 cells exposed
to hypoxia, whereas the same antibodies can detect the
activation of JNK in L929 cells treated with TNF

(data not shown). These results suggest that p38 or
ERK1 ⁄ 2 but not JNK may participate in the downre-
gulation of FXR during hypoxia.
In many studies concerning MAPK signaling, phar-
macological inhibitors specific for each subfamily of
MAPKs have been widely used. However, the action
of these compounds must be considered with caution,
as they sometimes affect other signaling pathways in
unexpected ways. Indeed, for unknown reasons,
SB203580, which is widely used as a p38-specific inhib-
itor, lowered FXR expression level by itself but
increased phosphorylated forms of p38 in HepG2 cells
(data not shown). There are many reports indicating
that SB203580 inhibits different types of kinases, such
as Akt ⁄ protein kinase B [42] and RIP ⁄ RICK [43],
independently of p38 inhibition. Furthermore,
SB203580 was shown to activate ERK in human
hepatocytes and HepG2 cells [44,45]. Therefore, we
carried out RNA interference experiments in order to
verify the role of p38 and ERK1 ⁄ 2, instead of using
pharmacological inhibitors.
The expression level of p38 was examined by immu-
noblotting using antibody against p38 followed by
densitometric analysis (Fig. 6A). The amount of p38
protein derived from cells treated with the p38a
siRNA under hypoxia was reduced to 61% when
expressed as a percentage of the value of the ‘normox-
ia’ control (Fig. 6A, middle panel). It should be noted
that the increase in the concentration of p38a siRNA
did not improve the knockdown efficiency of p38, but

rather resulted in impairment of cell viability (data not
shown). In spite of this partial knockdown efficiency,
p38a siRNA treatment sufficiently reversed the
hypoxia-dependent downregulation of FXR. As shown
in the top panel of Fig. 6A, the amount of FXR pro-
tein under hypoxia was 27% of the value of the ‘norm-
oxia’ control, and the treatment of cells with p38a
siRNA increased the FXR protein level to 42% even
under hypoxia (Fig. 6A, top panel). The restorative
effect of p38 siRNA was also confirmed by the quanti-
fication of FXR mRNA. As shown in Fig. 6B, the
lowered FXR mRNA level under hypoxia was signifi-
cantly elevated by treatment of cells with p38a siRNA.
The reduced BSEP mRNA expression under hypoxia
was also restored by the same treatment with p38a
siRNA (Fig. 6C). As a stress-activated MAPK, p38 is
known to be activated by various stimuli, including
proinflammatory cytokines. As shown in Fig. 6D,
interleukin (IL)-1b induced strong p38 activation in
HepG2 cells. Interestingly, the FXR level was reduced
by treatment with IL-1b (Fig. 6D, top panel). These
results suggest that p38 acts as an upstream signaling
molecule that responds to various environmental
stresses, including hypoxia, and downregulates FXR
transcription.
We next examined the role of ERK1 ⁄ 2 in hypoxia-
dependent downregulation of FXR. As shown in
Fig. 6E, simultaneous knockdown of ERK1 and
ERK2 reduced the ERK2 protein level to 44% of the
value of the ‘normoxia’ control (Fig. 6E, middle

panel). ERK1 expression was also decreased in cells
treated with ERK1⁄ 2 siRNAs. Similar to the case of
p38a siRNA, higher concentrations of siRNA against
ERK1 ⁄ 2 did not provide good knockdown efficiency
(data not shown). Under the condition described
above, the lowered expression of FXR protein in cells
exposed to hypoxia (48% of the value of the ‘normox-
ia’ control) was not elevated at all by the treatment
with ERK1 ⁄ 2 siRNAs (41%), as shown in the top
panel of Fig. 6E. As we could not achieve the com-
plete knockdown of ERK1 ⁄ 2, we could not rule out
the involvement of ERK1 ⁄ 2 in the hypoxia-dependent
downregulation of FXR. However, we can highlight
p38 MAPK as the key molecule responsible for the
Fig. 5. Activation of p38 and ERK1 ⁄ 2 MAPKs in response to
hypoxia occurs in inverse relation to the downregulation of FXR.
HepG2 cells were cultured under hypoxia for the indicated times.
Cell lysates were then prepared, and this was followed by immuno-
blotting analyses to detect FXR, phosphorylated forms of p38 and
ERK1 ⁄ 2orb-actin levels as a loading control. Quantification of the
bands was done by densitometric analysis (
IMAGE GAUGE 4.0). Similar
results were obtained in three separate experiments.
T. Fujino et al. Hypoxia downregulates FXR via p38 MAPK
FEBS Journal 276 (2009) 1319–1332 ª 2009 The Authors Journal compilation ª 2009 FEBS 1325
negative regulation of FXR transcription, as its activa-
tion occurs more drastically than that of ERK1 ⁄ 2
under hypoxia (Fig. 5), its partial knockdown is effec-
tive enough to restore the FXR function impaired by
the hypoxia treatment (Fig. 6A–C), and its activation

by IL-1b also leads to the downregulation of FXR
(Fig. 6D).
Discussion
In the current study, we have demonstrated that FXR, a
key transcription factor that regulates bile acid meta-
bolism, is downregulated under hypoxia through a p38
MAPK-dependent mechanism. The experimental model
shown here may give an explanation of how chronic
ischemia impairs liver function by attenuating the bile
acid homeostasis regulated by FXR and possibly leads
to progressive liver disorders such as primary biliary
cirrhosis and primary sclerosing cholangitis. Indeed, an
ischemia-induced low-oxygen condition, i.e. hypoxia,
has been thought of as a possible cause of bile duct
injury, in particular after liver transplantation, hepatic
surgery, and intra-arterial chemotherapy [23,46,47].
Hypoxia is a serious stress for living organs, because
of the need to make the massive change from oxygen-
dependent to oxygen-independent energy production.
Therefore, the ability to sense and respond to changes
in oxygen is essential for the survival of multicellular
organisms. HIF is the key transcription factor for sens-
ing and responding to lowered oxygen, acting by
inducing the transcription of various genes required
C
B
D
E
A
Fig. 6. p38 but not ERK1 ⁄ 2 MAPK is involved in the hypoxia-

dependent downregulation of FXR. (A) HepG2 cells were transfect-
ed with control siRNA or with the siRNA against p38a, and then
cultured under hypoxia for 24 h. Cell lysates were then prepared,
and this was followed by immunoblotting analyses to detect FXR
and p38 or b-actin levels as a loading control. Quantification of the
bands were done by densitometric analysis (
IMAGE GAUGE 4.0). Simi-
lar results were obtained in three separate experiments. (B) HepG2
cells transfected with control siRNA or with the siRNA against
p38a were cultured under hypoxia for 24 h. Total RNAs were then
prepared, and real-time PCR analysis to detect FXR mRNA was car-
ried out as described in Experimental procedures. Similar results
were obtained in three separate experiments. (C) HepG2 cells
transfected with control siRNA or with the siRNA against p38a
were cultured under hypoxia for 24 h. Total RNAs were then pre-
pared, and real-time PCR analysis to detect BSEP mRNA was per-
formed as described in Experimental procedures. Similar results
were obtained in three separate experiments. (D) HepG2 cells cul-
tured under normoxia were stimulated with 3 ngÆmL
)1
IL-1b for
1 h. Cells were then lysed in SDS ⁄ PAGE sample buffer, and the
resultant cell lysates were subjected to immunoblotting analyses to
detect FXR, phosphorylated forms of p38 or to detect b-actin levels
as a loading control. Quantification of the bands was done by densi-
tometric analysis (
IMAGE GAUGE 4.0). Similar results were obtained in
three separate experiments. (E) HepG2 cells were transfected with
control siRNA (#1022076) or with the combined mixture of siRNAs
against ERK1 and ERK2 then cultured under hypoxia for 24 h. Cell

lysates were then prepared followed by the immunoblotting analy-
ses to detect FXR and ERK1 ⁄ 2orb-actin levels as a loading
control. Quantification of the bands was done by densitometric
analysis (
IMAGE GAUGE 4.0). Similar results were obtained in three
separate experiments.
Hypoxia downregulates FXR via p38 MAPK T. Fujino et al.
1326 FEBS Journal 276 (2009) 1319–1332 ª 2009 The Authors Journal compilation ª 2009 FEBS
for survival under conditions where the oxygen supply
is limited. VEGF is one such HIF-targeted gene, and
is known to be upregulated in the cirrhotic liver [17].
During the development of biliary cirrhosis, hepatocel-
lular hypoxia and hepatic angiogenesis induced by
VEGF are known to contribute to the progression of
liver fibrosis [18]. Moreover, enhanced proliferation of
liver tumor cells leads to local hypoxia in hepato-
cellular carcinoma (HCC) [48], and in turn, hypoxia-
induced expression of angiogenic factors such as
VEGF results in the hypervascularity of HCC.
Undoubtedly, HIF and VEGF play central roles dur-
ing new vessel formation in HCC [48]. In contrast, our
present study has revealed that HIF is not involved in
hypoxia-dependent downregulation of FXR. Thus,
hypoxia may affect liver disorders by activating two
signaling pathways: one is the HIF-dependent pathway
that leads to the induction of VEGF, resulting in
hypervascularity in the liver; and the other is the
HIF-independent but p38-mediated pathway that
causes down-egulation of FXR and BSEP, resulting in
a decreased capacity for bile acid excretion from the

liver through the biliary tracts.
However, we should not rush to the conclusion that
the downregulated FXR expression under hypoxia
should be primarily restored to avoid the progress of
cholestatic liver disorders, as cholestasis results either
from an impairment of bile secretion or from an
obstruction of the bile duct [49]. As suggested by
Fiorucci et al., activation of canalicular transporters,
including BSEP, could negatively impact on intrahe-
patic bile duct pressure in patients with severely
obstructed bile flow [12]. This concept is supported by
the observation that FXR
) ⁄ )
mice develop less hepatic
injury in response to bile duct ligation than wild-type
mice [50,51].
Besides BSEP, the MRPs are known to be involved
in the export of bile acids from the liver [52]. Among
six MRP subfamilies (MRP1–6), four (MRP1–4) are
expressed in the liver [53]. MRP3 and MRP4 localize
in the basolateral membrane of the hepatocyte,
whereas MRP2 is expressed on the canalicular mem-
brane in a similar manner as BSEP. Interestingly, bile
duct obstruction in FXR
) ⁄ )
mice causes robust induc-
tion of MRP4 in the basolateral membrane [50]. As
this effect is not observed in wild-type mice, FXR
might negatively regulate the activation of basolateral
efflux of bile acid from hepatocytes [50,51]. Moreover,

the hepatic expression of human MRP3 is usually very
low; however, it is induced in patients with cholestasis
and cirrhosis [54]. Thus, the basolateral MRPs upregu-
lated during severe cholestasis may act as an alterna-
tive export system to eliminate bile acids from the liver
by elevating bile acid efflux across the basolateral
membrane of the hepatocyte, instead of their being
excreted through the bililary tract using transporters
located on the canalicular membrane.
Therefore, hypoxia-evoked downregulation of FXR
and BSEP may have opposite effects on the progress of
cholestasis, depending on whether or not bile flow is
severely obstructed. Under conditions where biliary
tract is not yet injured, the decreased bile flow that
results from the downregulation of FXR ⁄ BSEP in
response to hypoxia will lead to the accumulation of
toxic bile acids in hepatocytes, and cholestasis will con-
sequently be promoted. It should be noted that genetic
defects of BSEP cause a severe liver disease in humans
called progressive familial intrahepatic cholestasis
type 2, which leads to irreversible liver damage, owing
to intrahepatic bile acid accumulation [55]. In addition,
decreased FXR expression and activity is known to be
associated with FIC1 mutations, suggesting that FXR
itself may play an important role in the pathogenesis of
progressive familial intrahepatic cholestasis type 1
[56,57]. As we have shown in our present study, the
expression level of FXR is dynamically changed in cells
in response to extracellular stimuli, such as hypoxic
stress. Thus, at the early stage of ischemia, when bile

flow is not yet severely obstructed, hypoxia-evoked
FXR ⁄ BSEP downregulation may lead directly to chole-
stasis. In contrast, under conditions where the bile duct
is severely obstructed, the use of ‘retrograde’ alternative
and basolateral transporters instead of ‘orthograde’
canalicular transporters will be beneficial for the liver.
In this context, at the late stage of chronic ischemia, the
p38 MAPK signaling pathway may constitute the
molecular switch system that turns off the FXR-regu-
lated canalicular transporters in response to hypoxia,
and turns on the alternative basolateral transporters in
order to minimize the intrahepatic accumulation of bile
acids, although direct evidence for this has not yet been
provided. Interestingly, Kubitz et al. reported that the
treatment of HepG2 cells with cycloheximide, an inhibi-
tor of protein translation, induced trafficking of BSEP
from the Golgi to the canalicular membrane in a p38
MAPK-dependent manner [58]. It is likely that p38
MAPK activated in response to hypoxia allows BSEP
to be expressed in the canalicular membrane at a
minimal level while it turns off the FXR-regulated
transcription of BSEP.
Bile acid metabolism is tightly regulated via a com-
plex network of signaling pathways. Among them,
FXR and MAPKs have major roles, and imbalanced
outputs of their signals will lead to liver disorders.
There have been several reports showing cros-
stalk between FXR-regulated pathways and MAPK
T. Fujino et al. Hypoxia downregulates FXR via p38 MAPK
FEBS Journal 276 (2009) 1319–1332 ª 2009 The Authors Journal compilation ª 2009 FEBS 1327

signaling pathways [14,28–32]. In this study, we have
highlighted p38 MAPK and FXR as the signaling mol-
ecules that may play key roles in the pathogenesis of
the liver disorders accompanying ischemia and chole-
stasis. Although the molecules that act downstream of
p38 MAPK to suppress the function of FXR remain
to be elucidated, we believe that our present study con-
tributes to the understanding of the molecular basis of
cholestasis progressing under ischemia. Further
insights into the crosstalk between p38 MAPK and
FXR will be useful in identifying a novel therapeutic
target for this type of liver disorder.
Experimental procedures
Antibodies
Antibodies specific for b-actin (C-2), FXR (D-3) and
HIF-3a (H-170) were purchased from Santa Cruz Biotech-
nology (Santa Cruz, CA, USA). Antibodies against p38,
phospho-p38 (Thr180 ⁄ Tyr182) and phospho-ERK1 ⁄ 2 were
obtained from Cell Signalling Technology (Danvers, MA,
USA). Antibody against HIF-1a (clone 54) was a product
of BD Biosciences Pharmingen (Franklin Lakes, NJ, USA).
Antibody against HIF-2a (EP-190b) was obtained from
Novus Biologicals (Littleton, CO, USA). ECL anti-mouse
IgG, horseradish peroxidase-linked whole antibody (from
sheep), and ECL anti-rabbit IgG, horseradish peroxidase-
linked whole antibody (from donkey), were purchased from
GE Healthcare (Little Chalfont, UK).
Cell culture
Human hepatocellular carcinoma cell lines HepG2 and Huh7
were cultured in DMEM containing 10% fetal bovine serum,

50 unitsÆmL
)1
penicillin and 50 lgÆmL
)1
streptomycin in a
humidified atmosphere of 8.5% CO
2
at 37 °C.
Hypoxia experiments
Hypoxia (< 3%) was obtained in a workstation with O
2
and CO
2
and temperature control (Ikemoto Rika Co.,
Tokyo, Japan). For hypoxic harvesting, cells were lysed in
the workstation with buffers equilibrated in the hypoxic
condition. Cell viability during hypoxia was assessed by
3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide
assay [59].
Measurement of FXR activity
Cellular FXR activity was measured by dual luciferase
assay (Promega, Madison, WI, USA). HepG2 cells were
seeded on a 24-well culture plate at a density of
2 · 10
5
cells per well and cultured for 24 h. FXRE-driven
firefly luciferase reporter vector pGL4–FXREx4 (0.4 lg)
containing four copies of the FXRE consensus sequence
(GGGTCAGTGACCC) and Renilla reniformis luciferase
expression vector phRL–TK (0.04 lg) were cotransfected

into cells using Lipofectamine 2000 transfection reagent
(Invitrogen, Carlsbad, CA, USA) according to the manu-
facturer’s instructions. After 24 h of incubation, cells were
treated with 100 lm CDCA, and then further incubated in
the hypoxia workstation or left under the regular culture
condition (normoxia) for 24 h. Cells were then harvested,
and cellular firefly and Renilla luciferase activities were
measured using a chemiluminescense photometer. Firefly
luciferase activity was normalized with that of Renilla
luciferase. Data were analyzed by Student’s t-test.
NF-jB activity assay
TNF-induced NF-jB activation in HepG2 cells exposed to
hypoxia was assessed by dual luciferase assay (Promega).
HepG2 cells were seeded on a 24-well culture plate at a
density of 2 · 10
5
cells per well and cultured for 24 h, and
3 · j B–Luc luciferase reporter vector (0.4 lg) and phRL–
TK (0.04 lg) were cotransfected into cells using Lipofecta-
mine 2000 transfection reagent (Invitrogen). After 24 h of
incubation, cells were treated with 1 or 10 ngÆmL
)1
TNF,
and then further incubated in the hypoxia workstation or
left under the regular culture condition (normoxia) for
24 h. Cells were then harvested, and cellular firefly and
Renilla luciferase activities were measured using a chemilu-
minescense photometer. Firefly luciferase activity was
normalized with that of Renilla luciferase.
Immunoblotting

Cells were washed with NaCl ⁄ P
i
, and cell extracts were
prepared using SDS sample buffer. After normalization of
protein content by the protein assay, samples were sepa-
rated by SDS ⁄ PAGE and subjected to immunoblotting
analysis. For the detection of FXR, p38, phospho-p38, and
phospho-ERK1 ⁄ 2, poly(vinylidene difluoride) membranes
were incubated with the primary antibody for 2 h at room
temperature, and this was followed by incubation for 16 h
at 4 °C. For the detection of ERK1 ⁄ 2, HIF-1a, HIF-2a,
HIF-3a, and b-actin, the membranes were incubated with
the primary antibody for 2 h. Immunocomplexes on the
poly(vinylidene difluoride) membranes were visualized with
enhanced chemiluminescence western blotting detection
reagents (GE Healthcare Biosciences).
Quantification of mRNAs
The amounts of mRNAs were quantified by real-time PCR.
Briefly, 5 lg of total RNAs were reverse-transcribed by the
use of Ready-to-Go-You-Prime First-Strand Beads (GE
Hypoxia downregulates FXR via p38 MAPK T. Fujino et al.
1328 FEBS Journal 276 (2009) 1319–1332 ª 2009 The Authors Journal compilation ª 2009 FEBS
Healthcare Biosciences) according to the manufacturer’s
instructions, and the resultant cDNAs were then subjected to
real-time PCR analysis using a TaqMan Gene Expression
Assay kit (Applied Biosystems, Foster City, CA, USA). For
the detection of FXR, BSEP, LRH-1, SHP, GLUT-1 and
b-actin mRNAs, TaqMan assay mixtures Hs00231968,
Hs00184824, Hs00187067, Hs0022677, Hs00892681 and
4310881E were used, respectively. TaqMan assay mixture

Hs00231968 can detect all of the four isoforms of FXR [60].
Amplification and quantification were done with the
PRISM 7000 Real-Time PCR System (Applied Biosystems).
FXR, BSEP, LRH-1, SHP and GLUT-1 mRNA levels
were normalized to the levels of b-actin mRNA as an internal
control. Data were analyzed by Student’s t-test.
Construction of the constitutively active HIF-1a
expression vector
We have previously used the expression vector encoding the
mutant of HIF-1a, in which Pro564 and Asn803 are
replaced by alanine [61], as this mutant was expected to act
as a constitutively active version of HIF-1a [21]. In addi-
tion to Pro564, Pro402 in HIF-1a is now recognized as the
target residue of prolyl hydroxylase; therefore, we made an
additional mutation, replacing Pro402 with the alanine resi-
due. In brief, we performed site-directed mutagenesis by
PCR as follows. The pBluescript SK vector (Stratagene, La
Jolla, CA, USA) encompassing an HIF-1a cDNA in which
Pro564 and Asn808 were replaced with alanines was used
as a template. Two sets of primers [set 1, 5¢-TAGTCCAG
TGTGGTGGAATTCTGC-3¢ (sense) and 5¢-AAAGCATC
AGGTTCCTTCTTAAG-3¢ (antisense); set 2, 5¢-AACTTT
GCTGGCCGCCGCCGCTGG-3¢ (sense) and 5¢-GGCAAC
TAGAAGGCACAGTCGAGG-3¢ (anti-sense)] were used
to generate the substitution of Pro402 for the alanine resi-
due. The resultant cDNA was subcloned into pcDNA3.1
vector (Invitrogen) and termed pcDNA3.1–HIF-1a CA.
Transfection of HIF-a isoform
HepG2 cells were seeded on 35 mm dishes at a density of
2 · 10

5
cells per dish and cultured for 24 h. Cells were then
transfected with pcDNA3.1–HIF-1a CA, pcDNA3.1–HIF-
2a or pcDNA3.1–HIF-3a by the use of Lipofectamine 2000
transfection reagent (Invitrogen). After 24 h, protein
extracts or total RNAs were prepared, and then immuno-
blotting or real-time PCR analyses were carried out.
RNA interference experiments
The custom, HP or HP-validated siRNAs purchased from
Qiagen (Valencia, CA, USA) were used to knock down the
expression of the following target molecules as described
below.
According to the report by Sowter et al. [62], custom
siRNAs against HIF-1a and HIF-2a were prepared as
follows. The mixture of the sense (5¢-CUGAUGACCA
GCAACUUGAdTdT-3¢) and antisense (5¢-UCAAGUUGC
UGGUCAUCAGdTdT-3¢) oligonucleotides was denatured
at 90 °C, cooled for annealing, and used to knock down
HIF-1a. Similarly, the siRNA against HIF-2a was prepared
using the sense (5¢-CAGCAUCUUUGAUAGCAGUdTdT-
3¢) and antisense (5¢-ACUGCUAUCAAAGAUGCUGdT
dT-3¢) oligonucleotides. To knock down HIF-3a,HP
siRNA (Hs_HIF3A_1_HP siRNA) was used. In order to
knock down endogenous HIF-1a, HIF-2a and HIF-3a
simultaneously, HepG2 cells were seeded on 35 mm dishes
at a density of 2 · 10
5
cells per dish and cultured for 24 h,
and this was followed by transfection with the combined
mixture of siRNAs against HIF-1a, HIF2-a, and HIF3a

(150 nm each), using Oligofectamine Reagent (Invitrogen)
according to the manufacturer’s instructions. On the day
after the first transfection, a second transfection was
performed similarly to the first [63]. Cells were then
cultured under normoxic or hypoxic conditions for 24 h,
and protein extracts or total RNAs were prepared for
immunoblotting or real-time PCR analyses. To knock
down endogenous p38a or ERK1 ⁄ 2 expression, we used the
following validated siRNAs: Hs_MAPK14_6_HP validated
siRNA against human p38a; Hs_MAPK3_7_HP validated
siRNA against human ERK1; and Hs_MAPK1_10_HP
validated siRNA against human ERK2. HepG2 cells were
seeded on 35 mm dishes at a density of 5 · 10
5
cells per
dish. Immediately after seeding, cells were transfected with
p38a siRNA (10 nm), or with the mixture of ERK1 siRNA
and ERK2 siRNA (10 nm each), using HiPerfect Transfec-
tion Reagent (Qiagen) according to the manufacturer’s
instructions. Cells were then cultured under hypoxic condi-
tions for 24 h, and protein extracts or total RNAs were
prepared for immunoblotting or real-time PCR analyses. In
the series of RNA interference experiments, ‘nonsilencing
control’ siRNA (#1022076) from Qiagen was used as a
control.
Acknowledgements
We thank R. Sato, T. Nishimaki-Mogami and H. Hay-
ashi for helpful advice and discussions. We also thank
S. Miyamoto for providing us with the 3 · jB–Luc
luciferase reporter vector. This work was supported in

part by a grant from the Japan Private School Promo-
tion Foundation.
References
1 Mangelsdorf DJ, Thummel C, Beato M, Herrlich P,
Schu
¨
tz G, Umesono K, Blumberg B, Kastner P, Mark
T. Fujino et al. Hypoxia downregulates FXR via p38 MAPK
FEBS Journal 276 (2009) 1319–1332 ª 2009 The Authors Journal compilation ª 2009 FEBS 1329
M, Chambon P et al. (1995) The nuclear receptor
superfamily: the second decade. Cell 83, 835–839.
2 Forman BM, Goode E, Chen J, Oro AE, Bradley DJ,
Perlmann T, Noonan DJ, Burka LT, McMorris T,
Lamph WW et al. (1995) Identification of a nuclear
receptor that is activated by farnesol metabolites. Cell
81, 687–693.
3 Makishima M, Okamoto AY, Repa JJ, Tu H, Learned
RM, Luk A, Hull MV, Lustig KD, Mangelsdorf DJ &
Shan B (1999) Identification of a nuclear receptor for
bile acids. Science 284, 1362–1365.
4 Parks DJ, Blanchard SG, Bledsoe RK, Chandra G,
Consler TG, Kliewer SA, Stimmel JB, Willson TM,
Zavacki AM, Moore DD et al. (1999) Bile acids: natu-
ral ligands for an orphan nuclear receptor. Science 284,
1365–1368.
5 Wang H, Chen J, Hollister K, Sowers LC & Forman
BM (1999) Endogenous bile acids are ligands for the
nuclear receptor FXR ⁄ BAR. Mol Cell 3, 543–553.
6 Seol W, Choi HS & Moore DD (1995) Isolation of
proteins that interact specifically with the retinoid X

receptor: two novel orphan receptors. Mol Endocrinol 9,
72–85.
7 Glass CK (1994) Differential recognition of target genes
by nuclear receptor monomers, dimers, and heterodi-
mers. Endocr Rev 15, 391–407.
8 Huber RM, Murphy K, Miao B, Link JR, Cunningham
MR, Rupar MJ, Gunyuzlu PL, Haws TF, Kassam A,
Powell F et al. (2002) Generation of multiple farnesoid-
X-receptor isoforms through the use of alternative
promoters. Gene 290, 35–43.
9 Zhang Y, Kast-Woelbern HR & Edwards PA (2003)
Natural structural variants of the nuclear receptor
farnesoid X receptor affect transcriptional activation.
J Biol Chem 278, 104–110.
10 Rizzo G, Disante M, Mencarelli A, Renga B, Gioiello
A, Pellicciari R & Fiorucci S (2006) The farnesoid X
receptor promotes adipocyte differentiation and regu-
lates adipose cell function in vivo. Mol Pharmacol 70,
1164–1173.
11 Cariou B, van Harmelen K, Duran-Sandoval D, van
Dijk TH, Grefhorst A, Abdelkarim M, Caron S, Torpi-
er G, Fruchart JC, Gonzalez FJ et al. (2006) The farne-
soid X receptor modulates adiposity and peripheral
insulin sensitivity in mice. J Biol Chem 281, 11039–
11049.
12 Fiorucci S, Rizzo G, Donini A, Distrutti E & Santucci
L (2007) Targeting farnesoid X receptor for liver and
metabolic disorders. Trends Mol Med 13, 298–309.
13 Zhang Y & Edwards PA (2008) FXR signaling in meta-
bolic disease. FEBS Lett 582, 10–18.

14 Zollner G, Marschall HU, Wagner M & Trauner M
(2006) Role of nuclear receptors in the adaptive
response to bile acids and cholestasis: pathogenetic and
therapeutic considerations. Mol Pharm 3, 231–251.
15 Cutrin JC, Cantino D, Biasi F, Chiarpotto E, Salizzoni
M, Andorno E, Massano G, Lanfranco G, Rizzetto M,
Boveris A et al. (1996) Reperfusion damage to the bile
canaliculi in transplanted human liver. Hepatology 24,
1053–1057.
16 Henley KS, Lucey MR, Appelman HD, Baliga P,
Brown KA, Burtch GD, Campbell DA Jr, Ham JM,
Merion RM & Turcotte JG (1992) Biochemical and his-
topathological correlation in liver transplant: the first
180 days. Hepatology 16, 688–693.
17 Rosmorduc O, Wendum D, Corpechot C, Galy B, Seb-
bagh N, Raleigh J, Housset C & Poupon R (1999)
Hepatocellular hypoxia-induced vascular endothelial
growth factor expression and angiogenesis in experi-
mental biliary cirrhosis. Am J Pathol 155, 1065–1073.
18 Corpechot C, Barbu V, Wendum D, Kinnman N, Rey
C, Poupon R, Housset C & Rosmorduc O (2002)
Hypoxia-induced VEGF and collagen I expressions are
associated with angiogenesis and fibrogenesis in experi-
mental cirrhosis. Hepatology 35, 1010–1021.
19 Bozova S & Elpek GO (2007) Hypoxia-inducible factor-
1a expression in experimental cirrhosis: correlation with
vascular endothelial growth factor expression and
angiogenesis. APMIS 115, 795–801.
20 Giaccia AJ, Simon MC & Johnson R (2004) The biol-
ogy of hypoxia: the role of oxygen sensing in develop-

ment, normal function, and disease. Genes Dev 18,
2183–2194.
21 Lando D, Peet DJ, Whelan DA, Gorman JJ & White-
law ML (2002) Asparagine hydroxylation of the HIF
transactivation domain a hypoxic switch. Science 295,
858–861.
22 Maynard MA, Qi H, Chung J, Lee EH, Kondo Y,
Hara S, Conaway RC, Conaway JW & Ohh M (2003)
Multiple splice variants of the human HIF-3 alpha
locus are targets of the von Hippel-Lindau E3 ubiquitin
ligase complex. J Biol Chem 278, 11032–11040.
23 Fouassier L, Beaussier M, Schiffer E, Rey C, Barbu V,
Mergey M, Wendum D, Callard P, Scoazec JY, Lasnier
E et al. (2007) Hypoxia-induced changes in the expres-
sion of rat hepatobiliary transporter genes. Am J Phys-
iol Gastrointest Liver Physiol 293, G25–G35.
24 Datta S, Wang L, Moore DD & Osborne TF (2006)
Regulation of 3-hydroxy-3-methylglutaryl coenzyme A
reductase promoter by nuclear receptors liver receptor
homologue-1 and small heterodimer partner: a mecha-
nism for differential regulation of cholesterol synthesis
and uptake. J Biol Chem 281, 807–812.
25 De Fabiani E, Mitro N, Gilardi F, Caruso D, Galli G
& Crestani M (2003) Coordinated control of cholesterol
catabolism to bile acids and of gluconeogenesis via a
novel mechanism of transcription regulation linked to
the fasted-to-fed cycle. J Biol Chem 278, 39124–39132.
26 Han SI, Studer E, Gupta S, Fang Y, Qiao L, Li W,
Grant S, Hylemon PB & Dent P (2004) Bile acids
Hypoxia downregulates FXR via p38 MAPK T. Fujino et al.

1330 FEBS Journal 276 (2009) 1319–1332 ª 2009 The Authors Journal compilation ª 2009 FEBS
enhance the activity of the insulin receptor and glycogen
synthase in primary rodent hepatocytes. Hepatology 39,
456–463.
27 Ono K & Han J (2000) The p38 signal transduction
pathway: activation and function. Cell Signal 12, 1–13.
28 Holt JA, Luo G, Billin AN, Bisi J, McNeill YY, Kozar-
sky KF, Donahee M, Wang DY, Mansfield TA, Klie-
wer SA et al. (2003) Definition of a novel growth
factor-dependent signal cascade for the suppression of
bile acid biosynthesis. Genes Dev 17, 1581–1591.
29 Stravitz RT, Rao YP, Vlahcevic ZR, Gurley EC, Jarvis
WD & Hylemon PB (1996) Hepatocellular protein kina-
se C activation by bile acids: implications for regulation
of cholesterol 7 alpha-hydroxylase. Am J Physiol 271,
G293–G303.
30 De Fabiani E, Mitro N, Anzulovich AC, Pinelli A,
Galli G & Crestani M (2001) The negative effects of
bile acids and tumor necrosis factor-a on the transcrip-
tion of cholesterol 7 a-hydroxylase gene (CYP7A1) con-
verge to hepatic nuclear factor-4: a novel mechanism of
feedback regulation of bile acid synthesis mediated by
nuclear receptors. J Biol Chem 276, 30708–30716.
31 Fang Y, Han SI, Mitchell C, Gupta S, Studer E, Grant
S, Hylemon PB & Dent P (2004) Bile acids induce mito-
chondrial ROS, which promote activation of receptor
tyrosine kinases and signaling pathways in rat hepato-
cytes. Hepatology 40 , 961–971.
32 Xu Z, Tavares-Sanchez OL, Li Q, Fernando J, Rodri-
guez CM, Studer EJ, Pandak WM, Hylemon PB & Gil

G (2007) Activation of bile acid biosynthesis by the p38
mitogen-activated protein kinase (MAPK): hepatocyte
nuclear factor-4a phosphorylation by the p38 MAPK is
required for cholesterol 7a-hydroxylase expression.
J Biol Chem 282, 24607–24614.
33 Kulisz A, Chen N, Chandel NS, Shao Z & Schumacker
PT (2002) Mitochondrial ROS initiate phosphorylation
of p38 MAP kinase during hypoxia in cardiomyocytes.
Am J Physiol Lung Cell Mol Physiol 282, L1324–L1329.
34 Xu L, Pathak PS & Fukumura D (2004) Hypoxia-
induced activation of p38 mitogen-activated protein
kinase and phosphatidylinositol 3¢-kinase signaling
pathways contributes to expression of interleukin 8 in
human ovarian carcinoma cells. Clin Cancer Res 10,
701–707.
35 Eguchi R, Suzuki A, Miyakaze S, Kaji K & Ohta T
(2007) Hypoxia induces apoptosis of HUVECs in an
in vitro capillary model by activating proapoptotic sig-
nal p38 through suppression of ERK1 ⁄ 2. Cell Signal
19, 1121–1131.
36 Song X, Kaimal R, Yan B & Deng R (2008) Liver recep-
tor homolog 1 transcriptionally regulates human bile salt
export pump expression. J Lipid Res 49, 973–984.
37 Zagorska A & Dulak J (2004) HIF-1: the knowns and
unknowns of hypoxia sensing. Acta Biochim Pol 51,
563–585.
38 Bruegge K, Jelkmann W & Metzen E (2007) Hydroxyl-
ation of hypoxia-inducible transcription factors and
chemical compounds targeting the HIF-alpha hydroxy-
lases. Curr Med Chem 14, 1853–1862.

39 Maynard MA & Ohh M (2007) The role of hypoxia-
inducible factors in cancer. Cell Mol Life Sci 64, 2170–
2180.
40 Conrad PW, Rust RT, Han J, Millhorn DE & Beitner-
Johnson D (1999) Selective activation of p38a and p38
c
by hypoxia. Role in regulation of cyclin D1 by hypoxia
in PC12 cells. J Biol Chem 274, 23570–23576.
41 Cowan KJ & Storey KB (2003) Mitogen-activated pro-
tein kinases: new signaling pathways functioning in cel-
lular responses to environmental stress. J Exp Biol 206,
1107–1115.
42 Lali FV, Hunt AE, Turner SJ & Foxwell BM (2000)
The pyridinyl imidazole inhibitor SB203580 blocks
phosphoinositide-dependent protein kinase activity,
protein kinase B phosphorylation, and retinoblastoma
hyperphosphorylation in interleukin-2-stimulated T cells
independently of p38 mitogen-activated protein kinase.
J Biol Chem 275, 7395–7402.
43 Hollenbach E, Vieth M, Roessner A, Neumann M,
Malfertheiner P & Naumann M (2005) Inhibition of
RICK ⁄ nuclear factor-kappaB and p38 signaling attenu-
ates the inflammatory response in a murine model of
Crohn disease. J Biol Chem 280, 14981–14988.
44 Henklova P, Vrzal R, Papouskova B, Bednar P, Janco-
va P, Anzenbacherova E, Ulrichova J, Maurel P, Pavek
P & Dvorak Z (2008) SB203580, a pharmacological
inhibitor of p38 MAP kinase transduction pathway acti-
vates ERK and JNK MAP kinases in primary cultures
of human hepatocytes. Eur J Pharmacol 593, 16–23.

45 Birkenkamp KU, Tuyt LM, Lummen C, Wierenga AT,
Kruijer W & Vellenga E (2000) The p38 MAP kinase
inhibitor SB203580 enhances nuclear factor-kappa B
transcriptional activity by a non-specific effect upon the
ERK pathway. Br J Pharmacol 131, 99–107.
46 Batts KP (1998) Ischemic cholangitis. Mayo Clin Proc
73, 380–385.
47 Beaussier M, Wendum D, Fouassier L, Rey C, Barbu
V, Lasnier E, Lienhart A, Scoazec JY, Rosmorduc O &
Housset C (2005) Adaptative bile duct proliferative
response in experimental bile duct ischemia. J Hepatol
42, 257–265.
48 Kim KR, Moon HE & Kim KW (2002) Hypoxia-
induced angiogenesis in human hepatocellular carci-
noma. J Mol Med 80, 703–714.
49 Paumgartner G (2006) Medical treatment of cholestatic
liver diseases: from pathobiology to pharmacological
targets. World J Gastroenterol 12, 4445–4451.
50 Stedman C, Liddle C, Coulter S, Sonoda J, Alvarez JG,
Evans RM & Downes M (2006) Benefit of farnesoid X
receptor inhibition in obstructive cholestasis. Proc Natl
Acad Sci USA 103, 11323–11328.
T. Fujino et al. Hypoxia downregulates FXR via p38 MAPK
FEBS Journal 276 (2009) 1319–1332 ª 2009 The Authors Journal compilation ª 2009 FEBS 1331
51 Marschall HU, Wagner M, Bodin K, Zollner G, Fickert
P, Gumhold J, Silbert D, Fuchsbichler A, Sjo
¨
vall J &
Trauner M (2006) Fxr(– ⁄ –) mice adapt to biliary
obstruction by enhanced phase I detoxification and

renal elimination of bile acids. J Lipid Res 47, 582–592.
52 Alrefai WA & Gill RK (2007) Bile acid transporters:
structure, function, regulation and pathophysiological
implications. Pharm Res 24, 1803–1823.
53 Trauner M & Boyer JL (2003) Bile salt transporters:
molecular characterization, function, and regulation.
Physiol Rev 83, 633–671.
54 Ko
¨
nig J, Rost D, Cui Y & Keppler D (1999) Character-
ization of the human multidrug resistance protein
isoform MRP3 localized to the basolateral hepatocyte
membrane. Hepatology 29, 1156–1163.
55 Thompson R & Strautnieks S (2001) BSEP: function
and role in progressive familial intrahepatic cholestasis.
Semin Liver Dis 21, 545–550.
56 Chen F, Ananthanarayanan M, Emre S, Neimark E,
Bull LN, Knisely AS, Strautnieks SS, Thompson RJ,
Magid MS, Gordon R et al. (2004) Progressive familial
intrahepatic cholestasis, type 1, is associated with
decreased farnesoid X receptor activity. Gastroenterol-
ogy 126, 756–764.
57 Alvarez L, Jara P, Sa
´
nchez-Sabate
´
E, Hierro L, Larrau-
ri J, Dı
´
az MC, Camarena C, De la Vega A, Frauca E,

Lo
´
pez-Collazo E et al. (2004) Reduced hepatic expres-
sion of farnesoid X receptor in hereditary cholestasis
associated to mutation in ATP8B1. Hum Mol Genet 13,
2451–2460.
58 Kubitz R, Su
¨
tfels G, Ku
¨
hlkamp T, Ko
¨
lling R & Ha
¨
us-
singer D (2004) Trafficking of the bile salt export pump
from the Golgi to the canalicular membrane is regulated
by the p38 MAP kinase. Gastroenterology 126, 541–553.
59 Mosmann T (1983) Rapid colorimetric assay for cellular
growth and survival: application to proliferation and
cytotoxicity assays. J Immunol Methods 65, 55–63.
60 Zhang Y, Castellani LW, Sinal CJ, Gonzalez FJ &
Edwards PA (2004) Peroxisome proliferator-activated
receptor-gamma coactivator 1 a (PGC-1 a) regulates
triglyceride metabolism by activation of the nuclear
receptor FXR. Genes Dev 18, 157–169.
61 Horii K, Suzuki Y, Kondo Y, Akimoto M, Nishimura
T, Yamabe Y, Sakaue M, Sano T, Kitagawa T, Himeno
S et al. (2007) Androgen-dependent gene expression of
prostate-specific antigen is enhanced synergistically by

hypoxia in human prostate cancer cells. Mol Cancer
Res 5, 383–391.
62 Sowter HM, Raval RR, Moore JW, Ratcliffe PJ &
Harris AL (2003) Predominant role of hypoxia-induc-
ible transcription factor (Hif)-1 a versus Hif-2 a in
regulation of the transcriptional response to hypoxia.
Cancer Res 63, 6130–6134.
63 Elvidge GP, Glenny L, Appelhoff RJ, Ratcliffe PJ,
Ragoussis J & Gleadle JM (2006) Concordant regula-
tion of gene expression by hypoxia and 2-oxoglutarate-
dependent dioxygenase inhibition: the role of HIF-1a,
HIF-2a, and other pathways. J Biol Chem 281, 15215–
15226.
Hypoxia downregulates FXR via p38 MAPK T. Fujino et al.
1332 FEBS Journal 276 (2009) 1319–1332 ª 2009 The Authors Journal compilation ª 2009 FEBS