Berberine Improves Glucose Metabolism in Diabetic Rats
by Inhibition of Hepatic Gluconeogenesis
Xuan Xia1, Jinhua Yan1, Yunfeng Shen1, Kuanxiao Tang1, Jun Yin3, Yanhua Zhang1, Dongjie Yang2, Hua
Liang1, Jianping Ye3, Jianping Weng1*
1Department of Endocrinology, The Third Affiliated Hospital of Sun Yat-Sen University, Guangzhou, China, 2Department of gastrointestinal-pancreatic surgery, The First-
Affiliated Hospital, Sun Yat-Sen University, Guangzhou, China, 3Pennington Biomedical Research Center, Louisiana State University, Eunice, Louisiana, United States of
Berberine (BBR) is a compound originally identified in a Chinese herbal medicine Huanglian (Coptis chinensis French). It
improves glucose metabolism in type 2 diabetic patients. The mechanisms involve in activation of adenosine
monophosphate activated protein kinase (AMPK) and improvement of insulin sensitivity. However, it is not clear if BBR
reduces blood glucose through other mechanism. In this study, we addressed this issue by examining liver response to BBR
in diabetic rats, in which hyperglycemia was induced in Sprague-Dawley rats by high fat diet. We observed that BBR
decreased fasting glucose significantly. Gluconeogenic genes, Phosphoenolpyruvate carboxykinase (PEPCK) and Glucose-6-
phosphatase (G6Pase), were decreased in liver by BBR. Hepatic steatosis was also reduced by BBR and expression of fatty
acid synthase (FAS) was inhibited in liver. Activities of transcription factors including Forkhead transcription factor O1
(FoxO1), sterol regulatory element-binding protein 1c (SREBP1) and carbohydrate responsive element-binding protein
(ChREBP) were decreased. Insulin signaling pathway was not altered in the liver. In cultured hepatocytes, BBR inhibited
oxygen consumption and reduced intracellular adenosine triphosphate (ATP) level. The data suggest that BBR improves
fasting blood glucose by direct inhibition of gluconeogenesis in liver. This activity is not dependent on insulin action. The
gluconeogenic inhibition is likely a result of mitochondria inhibition by BBR. The observation supports that BBR improves
glucose metabolism through an insulin-independent pathway.
Citation: Xia X, Yan J, Shen Y, Tang K, Yin J, et al. (2011) Berberine Improves Glucose Metabolism in Diabetic Rats by Inhibition of Hepatic Gluconeogenesis. PLoS
ONE 6(2): e16556. doi:10.1371/journal.pone.0016556
Editor: Aimin Xu, University of Hong Kong, China
Received September 1, 2010; Accepted January 1, 2011; Published February 3, 2011
Copyright: ? 2011 Xia et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted
use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This study was supported by 973 Program of China 2006CB503902, Program for Changjiang Scholars and Innovative Research Team in University (985
project PCSIRT 0947), National Science Fund for Distinguished Young Scholars 81025005 (to Jianping Weng). The funders had no role in study design, data
collection and analysis, decision to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: firstname.lastname@example.org
Berberine (BBR, C20H18NO4) is an isoquinoline alkaloid
originally isolated from the Chinese herb Coptis chinensis
(Huanglian) . BBR is widely used to reduce blood glucose in
type 2 diabetes in China . The mechanism is related to
activation of AMPK (Adenosine monophosphate-activated protein
kinase), ACC (acetyl-CoA carboxylase) and improvement of fatty
acid oxidation . BBR may improve insulin sensitivity through
AMPK pathway or induction of insulin receptor expression
[2,3,4]. We reported that AMPK activation by BBR is dependent
on mitochondrial inhibition , which leads to an increase in the
AMP/ATP(adenosine mono-phosphate/adenosine triphosphate)
ratio in cells after BBR treatment. The AMP elevation is a result of
the complex I inhibition in the mitochondrial respiratory chain
. The complex I inhibition blocks AMP conversion into ATP in
the mitochondria. Therefore, BBR may reduce fasting blood
glucose through an insulin-independent signaling pathway, such as
AMPK activation and insulin receptor expression. This pathway is
important in the skeletal muscle and adipose tissue in the insulin-
induced glucose deposition. Hepatic gluconeogenesis contributes
to elevation of fasting glucose, and requires mitochondrial
production of ATP. We propose that the mitochondrial inhibition
by BBR may contribute to the fasting glucose reduction through
ATP depletion. In this study, we addressed this issue by studying
hepatic gluconeogenesis in diabetic rats.
Hyperglycemia in type 2 diabetes is characterized by enhanced
glucose production in the liver and kidney. Fasting blood glucose is
determined by de novo glucose production (in liver and kidney)
and glucose deposition in peripheral tissues. In the presence of
insulin resistance, enhanced glucose output by liver contributes to
hyperglycemia together with reduced glucose deposition in skeletal
muscle, heart and adipose tissue. Inhibition of hepatic glucose
production contributes to glycemic control in the diabetic patients
by insulin sensitizers. Insulin inhibits expression of phosphoenol-
pyruvate carboxykinase (PEPCK) and glucose-6-phosphatase
(G6Pase), two rate-limiting genes in gluconeogenesis. These genes
are induced by glucagon and reduced by insulin. Activation and
inhibition of transcription factors including forkhead transcription
factor O1 (FoxO1), hepatic nuclear factor 4 (HNF4), and
(PGC-1a) are the underlying mechanism of the hormone actions
[7,8,9]. The activities of these transcription factors are highly
dependent on ATP supply. In the transcriptional process,
transcriptional initiation, RNA elongation and RNA splicing all
requires ATP supply, which provides energy to enzymes at each
PLoS ONE | www.plosone.org1 February 2011 | Volume 6 | Issue 2 | e16556
step. When ATP is reduced in cells, such as under hypoxia, gene
transcription is suppressed in response to energy depletion .
Additionally, mRNA translation will be decreased leading to
suppression of protein synthesis. The reduction in gene transcrip-
tion and protein synthesis may lead to inhibition of hepatic
gluconeogenesis. Given that BBR inhibits mitochondrial function,
this possibility may represent a new mechanism for BBR in the
control of blood glucose.
In the present study, we examined the molecular mechanism of
hepatic gluconeogenesis in response to BBR in rats. We observed:
(1) BBR decreased fasting blood glucose and expression of key
gluconeogenic genes, such as PEPCK and G6Pase; (2) BBR
suppressed the transcription factors, FoxO1 and sterol regulatory
element-binding protein 1c (SREBP1); (3) BBR inhibited mito-
chondrial function in hepatocytes.
Materials and Methods
Type 2 diabetes rat model
The type 2 diabetic rat model was developed using male
Sprague-Dawley rats (7–8 weeks, 200,250 g) according to a
protocol published elsewhere . The rats were fed on high fat
diet (63.47% calorie in fat) for 8 weeks, and then followed by a
single low-dosage intraperitoneal injection of streptozotocin (STZ,
30 mg/kg, Sigma, St Louis, MO) after 12-hour fast. The non-
diabetic control rats were fed on chow diet. Diabetic rats were
divided into two subgroups: vehicle treated and BBR treated
(n=9). Non-diabetic rats were subgrouped in the same way.
Plasma glucose increased in diabetic rats to 16.7 mmol/L within 3
days after STZ treatment and remained at this level throughout
the experiment. These rats were used in the type 2 diabetic model.
All procedures were performed in accordance with the principles
of laboratory animal care (National Institutes of Health publica-
tion no. 85-23, revised 1985) and the animal protocol was
approved by the Sun Yat-Sen University, Institutional Animal
Care and Use Committee (IACUC, Approval ID:2008010102).
The BBR intervention was initiated on the third day after STZ
injection and conducted for five weeks. BBR solution was prepared
in PBS and delivered by oralgavage at dosage of 380 mg?kg21?d21.
The control group was given vehicle PBS.
Blood and liver collection
After five weeks’ BBR treatment, liver and blood were collected
from each rat under anesthetized. Blood was collected from
abdominal aorta using syringe puncture. Liver was removed and
divided into three pieces, which were subjected to 4% formalde-
hyde fixation, Oil Red O staining and protein/RNA extraction.
Tissue for protein/RNA preparation was stored in 280uC freezer
after snap-frozen in liquid nitrogen. The sample collection was
conducted in fasting (12-hour) and non-fasting conditions,
Fasting blood glucose was determined in the tail vein blood
using portable glucometer (Roche, Basel, Switzerland). Blood
insulin was measured in serum derived from the abdominal aorta
blood with an insulin radioimmunoassay kit, which was purchased
from Linco Research (Missouri, USA). Serum triglyceride and
total cholesterol, alanine aminotransferase (ALT) and aspartate
transaminase (AST) were determined using automatic biochem-
istry analyzer. Energy balance (diet intake, calorie intake, urine
volume, and body weight) was monitored in individual rat for 3
consecutive days using a ‘‘metabolic cage’’. The data were
collected every day.
Glucose tolerance and insulin tolerance tests
Oral glucose tolerance test (GTT) was performed 48 hours after
the five-week BBR intervention. Glucose (2.0 g?kg21body weight)
was administered through oral gavage after 12-hour fast. Blood
glucose was determined at 0, 30, 60, 90 and 120 minutes after
glucose challenge. Insulin tolerance test was carried out by
intraperitoneal injection with insulin (0.75 U.kg21body wt). Blood
glucose was checked at 0, 15, 30, 45, 60, and 90 minutes after
insulin injection. Food intake, urine volume, diet intake were
tested in each rat individually.
Insulin releasing test
Insulin releasing test (IRT) was performed during the glucose
tolerance test. Serum sample was collected at 0, 15, 30, 60, 90 and
120 minutes after glucose gavage. Serum insulin was determined
by radioimmunoassay kit (Linco Research, Missouri, USA).
Preparation for total, nuclear and cytoplasmic protein
To determine insulin signaling activity, we collected the liver
tissue at 30 minutes after intraperitoneal insulin injection
(0.75 U.kg21body wt) and stored the samples at 280uC. The
frozen liver sample was minced in ice cold phosphate buffered
saline and homogenized in cell lysis buffer (Cell Signaling
Technology Inc., Danvers, MA). The whole cell protein was
collected after centrifugation at 4uC. The protein concentration
was measured using the BCA protein assay kit (Sangon Biotech,
Shanghai, China). The nuclear and cytoplasm protein were
extracted from liver tissue using a Nuclear Extraction Kit
(Chemicon, Inc, USA).
Protein (30 mg) was resolved in 10% sodium dodecyl sulfate
polyacrylamide gel (SDS-PAGE) and transferred to a PVDF
membrane. The antibodies to Akt (Protein kinase B, PKB),
PEPCK, G6Pase, fatty acid synthase (FAS), FoxO1, SREBP1,
(ChREBP), TFIID (the TATA box binding protein, TBP) were
purchased from the Santa Cruz Biotechnology (CA, USA).
(GADPH) and b-actin were from the Cell Signaling Technology,
USA. The peroxide-conjugated anti-rabbit antibody (Boster,
Wuhan, China) was used to detect the protein signal with
enhanced chemiluminescence system (Pierce, USA). The band
intensity was quantified by densitometry (Calibrate densitometer,
RNA isolation and quantitative real-time reverse
transcriptase polymerase chain reaction (qRT-PCR)
Total RNA was extracted from the frozen liver tissues using the
TRIzol RNA isolation reagent (Invitrogen, Carlsbad, CA).
Reverse transcription of 1 mg RNA was carried out according to
the instructions of Prime ScriptTM1st Strand cDNA Synthesis Kit
(TaKaRa, Japan). The qRT-PCR reaction was conducted in 20 ml
(SYBRH Premix Ex TaqTM, TaKaRa Japan). The result was
normalized against b-actin mRNA signal. Primer sequences were
as follows: Rat PEPCK forward primer (59-CTCACCTCTGGC-
CAAGATTGGTA-39) and reverse primer (59-GTTGCAGGCC-
CAGTTGTTGA-39); Rat G6Pase forward primer (59-AACGT
CTGTCTGTCCCGGATCTAC-39) and reverse primer (59-
Berberine Improves Glucose Metabolism
PLoS ONE | www.plosone.org2 February 2011 | Volume 6 | Issue 2 | e16556
ACCTCTGGAGGCTGGCATTG-39); Rat FAS forward primer
(59-CACAGCATT CAGTCCTATCCACAGA-39) and reverse
primer (59-CACAGCCAACCAGATGCTTCA-39); Rat FoxO1
forward primer (59-TGCCAAACTCACTACACCACTT-39) and
reverse primer (59-ACGATCAGGTTCCGTCATTC-39); Rat b-
actin forward primer (59-GGAGATTACTGCCCTGGCTCC
TA-39) and reverse primer (59-GACTCA TCGTACTCCTGCTT
The liver sample fixed in 4% Para formaldehyde was paraffin-
embedded, sectioned at 8 mm in thickness and stained with
hematoxylin & eosin dye (HE) and immunohistochemistry (DAB
staining). The frozen samples were sectioned and stained with Oil
Red O and Sudan III. The immunohistostaining was performed
according to instruction of the SABC assay (Boster, Wuhan,
China). Paraffin-embedded sections were deparaffinized using
xylene and re-hydrated in a series of alcohols, and incubated in
antigen retrieval (0.01 M citrate buffered saline, pH 6.0) for 5 min
at 95uC and then for 20 min at room temperature. The sections
were then treated with 3% hydrogen peroxide for 5 min. The slide
was treated with 10% normal goat serum to prevent nonspecific
antibody binding. The sample was treated with the primary
antibody overnight at 4uC in PBS. Imaging was obtained using a
Nikon microscope and analyzed by software (HMIAS-2000,
Championimage Inc, China).
Oxygen consumption and AMP/ATP ratio
Oxygen consumption and AMP/ATP ratio were examined in
H4IIE hepatocytes using protocols reported elsewhere . The
cells were cultured in a plate embedded with an oxygen-sensitive
dye. After 6 hrs, BBR was added to the medium to treat cells for
12 hrs. The oxygen consumption was determined by the
fluorescence intensity in the oxygen-sensitive dye, and expressed
as relative fluorescence units (NRFU). The reading was normal-
ized with the background in a blank well. The cells were treated
with 20 mM BBR for 16 hrs in serum-free medium containing
0.25% BSA. The cells were lyzed in 0.6 N HClO4and neutralized
with 1 N KHCO3. Then AMP and ATP contents were measured
using high-performance liquid chromatography (HPLC).
The results were expressed as the means 6 SE. The significant
difference in two groups was statistically analyzed using the
Student’s t test. The significant difference in more than two groups
was statistically analyzed using ANOVA. The least-square
deconvolution (LSD) was performed to detect differences between
two groups. Statistical significance was set at P,0.05 (two tails).
Characteristics of the experimental animals treated by
To investigate BBR action in liver, we used diabetic rats, which
were generated after 8 weeks high fat diet feeding and a single
streptozotocin treatment. The rats were treated with BBR through
oral gavage for 5 weeks, and then subjected to variety of tests. BBR
reduced body weight, fasting blood glucose, total cholesterol, and
triglycerides in the diabetic rats (Table 1). Calorie intake, serum
insulin, urine volume, water intake, serum ALT, and AST were not
altered by BBR. These data suggest that the reduction in blood
glucose and lipids may be a result of enhanced energy expenditure in
the rats. These were observed with reduced body weight in the
absence of food intake reduction. The body weight was also
decreased in the non-diabetic rats (lean rats) by BBR (Table 1),
suggesting that BBR reduces body weight inboth obese and lean rats.
Systemic insulin sensitivity was evaluated by fasting glucose,
GTT and ITT. In the GTT test, the BBR group exhibited lower
blood glucose than the un-treated control at the basal level
(Fig. 1A). After oral glucose challenge, the untreated group had
little increase in blood glucose. The blood glucose only increased
modestly by the glucose challenge. However, the BBR group
exhibited a dramatic increase in blood glucose at 30 and 60
minutes. The glucose reached to the same level as the untreated
group at 90 and 120 minutes. In the test, blood insulin was
monitored at each time point to determine the pancreatic beta cell
function. The BBR group and untreated group had identical basal
insulin. However, the BBR group exhibited a significantly higher
insulin level than the untreated group between 15–60 minutes
after the glucose challenge (Fig. 1B). In the ITT test, both the BBR
group and untreated group exhibited similar change in percentage
Table 1. Characteristics of the experimental animals at the end of intervention study (n=9).
IndexDiab+ +BBRDiab NorNor+ +BBR
Body weight (g)462634.6a b
495629.6 489634.2 452621.8a b
Water intake (ml?g21?d21) 0.274960.0530b
Urine volume (ml?g21?d21) 0.256960.0281b
Diet intake (g?g21?d21)0.06760.004a
Calorie intake (kcal?g21?d21) 0.38360.02b
FBG (mmol/L) 7.3061.19a b
FINS (ng/mL) 0.13660.009 0.14160.0070.12160.0010.13260.002
TC (mmol/L) 1.0360.11a,b
ALT (U/L) 69.60618.7384.43631.14b
BBR, berberine; Diab+BBR, diabetic rats treated with berberine; Nor, normal lean rats; Nor+BBR, normal rats treated with berberine; FBG, fasting blood glucose; FINS,
fasting serum insulin; TC, total cholesterol; TG, triglycerides; ALT, alanine aminotransferase; AST, aspartate transaminase. Data are presented as mean 6 SE (n=9).
a: P,0.05, compared with diabetic rats;
b: P,0.05, compared with normal rats(Nor).
Berberine Improves Glucose Metabolism
PLoS ONE | www.plosone.org3 February 2011 | Volume 6 | Issue 2 | e16556
of blood glucose at 60–90 minutes, but the BBR group had less
response than the untreated group during 15–45 minutes (Fig. 1C),
suggesting a delayed response to insulin in the BBR group. Those
data demonstrate that BBR reduces fasting blood glucose without
altering blood insulin in the basal condition, suggesting an increase
in systemic insulin sensitivity by HOMA-IR. Though GTT and
ITT results do not support that systemic insulin sensitivity was
improved by BBR, the results may not be definitive. Hyperinsu-
linemic-euglycemic clamping is required to determine insulin
To test insulin action in liver, we examined the insulin signaling
activity by checking Akt phosphorylation status in response to
insulin. Liver was collected from rats in 30 minutes after insulin
challenge. Akt phosphorylation status was determined using a
phospho-specific antibody in a Western blot. The phosphorylation
was not significantly improved in the BBR group (Fig. 1D),
suggesting that BBR treatment does not improve insulin signaling
in liver. We also examined AMPK in the liver. AMPK protein and
its phosphorylation were both reduced in the diabetic rats (Fig. 1E).
In the BBR group, the reduction was not observed, suggesting that
BBR may protect the AMPK activity. AMPK improves glucose
uptake by stimulating GLUT4 translocation in adipocytes and
muscle cells. In hepatocytes, this mechanism is not working since
hepatocytes do not express GLUT4. The data suggest that BBR
does not alter insulin signaling pathway in liver.
Berberine inhibits the expression of gluconeogenic
genes in liver
To investigate hepatic gluconeogenesis, we examined expression
of PEPCK and G6Pase in liver. PEPCK and G6Pase are two rate-
limiting enzymes in the gluconeogenic process in hepatocytes. The
protein levels of PEPCK and G6Pase are primarily controlled by
gene transcription. In the liver of BBR group, both enzymes were
reduced significantly in proteins (Fig. 2A). The changes were
observed in both fasting and fed conditions. mRNA of PEPCK
and G6Pase was also decreased in the BBR group, suggesting that
protein reduction is a result of mRNA reduction (Fig. 2B and 2C).
The inhibition was observed in the fasting condition, when insulin
level is low in the blood (Table 1). Although insulin inhibits the two
enzyme expression, the data do not support a role of insulin. The
data suggest that BBR suppresses PEPCK and G6Pase expression
independently of insulin.
Berberine inhibits FoxO1 in liver
FoxO1 is an important transcription factor in the control of
gluconeogenesis in liver . FoxO1 induces PEPCK and G6Pase
gene transcription to initiate gluconeogenesis. This activity is
dominant in fasting condition in the absence of a high level of
insulin, which inactivates the transcriptional activity of FoxO1
through Akt-mediated nucleus exclusion and protein degradation
Figure 1. BBR decreased fasting blood glucose without improvement in insulin sensitivity. (A) Effects of BBR treatment on OGTT in
diabetic rats. Oral glucose tolerance test (OGTT) was conducted with glucose 2 g.kg21body wt after 5 week BBR treatment (380 mg.kg21day21)
(n=6). (B) Insulin release during the OGTT (n=6). (C) ITT test (n=6). The test was conducted after 8 hour fasting with insulin (0.75 U.kg21body wt).
Nor, normal rats without obesity; Diab+Veh: Diabetic rats treated with vehicle; Diab+BBR, diabetic rats treated with BBR. # P,0.05, compared with
Diab+Veh. (D) Insulin signaling. Phosphorylation of Akt (Ser473). Rats were challenged with insulin (0.75 U.kg21, intraperitoneal injection) and liver
was collected in 30 minutes. The Akt assay was performed in a Western blot. Loading control is GAPDH. (E) Expression and phosphorylation of AMPK
(Thr172). # P,0.05, compared with Diab+Veh group. * P,0.05, compared with normal group (Nor).
Berberine Improves Glucose Metabolism
PLoS ONE | www.plosone.org4February 2011 | Volume 6 | Issue 2 | e16556
of FoxO1. To understand the mechanism of PEPCK and G6Pase
inhibition by BBR, we examined FoxO1 protein and mRNA in
the liver. FoxO1 protein was higher in the liver of diabetic rats
relative to the non-diabetic control (Fig. 3A). The elevation may be
a result of reduced insulin activity from the low levels of insulin
and insulin sensitivity in the diabetic rats. The FoxO1 protein was
reduced by BBR. The reduction was observed in both cytoplasm
and nucleus. However, the nuclear reduction was much stronger
(Fig. 3A). The protein reduction was also observed in the liver
tissue by immunohistostaining (DAB dye) (Fig. 3B). FoxO1 mRNA
was reduced by BBR, and the inhibition degree was identical in
both fasting and fed conditions (Fig. 3C). Given that insulin is low
in the fasting condition, the data suggest that BBR inhibits FoxO1
expression independently of insulin.
Berberine inhibits lipid accumulation in liver
Lipid accumulation in liver impairs hepatocyte function leading
to hyperglycemia through hepatic glucose production. We
examined BBR impact on fat content in liver in the BBR-treated
rats. Liver tissue was analyzed under microscope after HE staining
and Oil Red O staining. Lipid accumulation was elevated in the
diabetic rats as indicated by the unstained area in hepatocytes
under HE staining (Fig.4A), and red color in hepatocytes under
the Oil Red O or Sudan III staining (Fig. 4, B and C). The
hepatocyte density and proliferation status were determined with
DNA staining in blue color by hematoxylin (Fig. 4, A and C).
Hepatocyte proliferation was enhanced in the diabetic condition as
indicated by the loosen chromatin DNA in the nucleus of
hepatocytes in the diabetic group (Fig. 4, A and C). The
unpacking process is required for DNA duplication and chromo-
some doubling. In the BBR group, the change in DNA was
inhibited, and DNA was packed tightly in the nucleus, suggesting a
reduction in cell proliferation. The cell density was not altered by
BBR. The data suggest that BBR inhibits lipid accumulation in
hepatocytes. This is not a result of liver toxicity since BBR did not
reduce cell density in liver. BBR may inhibit the metabolic activity
in hepatocytes as cell proliferation activity is reduced by BBR.
Berberine ameliorates fatty acid synthesis in liver by
inhibiting expression of SREBP1 and FAS
SREBP1 and ChREBP are transcription factors for lipogenic
gene expression, such as FAS [12,13]. Inhibition of SREBP1 by
Figure 2. Expression of gluconeogenic genes in liver. (A) Protein levels of PEPCK and G6Pase. Total protein was made from liver tissue and
used in a Western blot. (B) mRNA in fasting condition. Total RNA was extracted in from liver tissue and used in qRT-PCR. The signal was normalized
with actin mRNA. (C) mRNA in fed condition. The mRNA data are presented as mean 6 SEM (n=6). # P,0.05, compared with Diab+Veh group.
* P,0.05, compared with normal group (Nor).
Berberine Improves Glucose Metabolism
PLoS ONE | www.plosone.org5February 2011 | Volume 6 | Issue 2 | e16556
gene knockout was reported to decrease lipid accumulation in liver
. SREBP2 is an isoform of SREBP1 and has an activity in
induction of cholesterol synthetic genes. It was reported that BBR
reduces SREBP2 expression in hepatocytes . To understand
the molecular basis of lipid reduction in liver, we measured
expression of SREBP1, SREBP2, ChREBP and FAS. Three genes
were reduced in liver by BBR (Fig. 5). The SREBP1 total protein
was determined in the liver tissue lysate and was decreased by
BBR (Fig. 5A). A reduction was also observed in ChREBP.
SREBP2 was not changed in the same condition. FAS expression
was increased in liver under the fed condition (Fig. 5, B and C).
The increase was observed even in the non-diabetic rats. However,
in the diabetic rats, the increase was enhanced much more,
suggesting a more active fatty acid synthesis in response to feeding.
In response to the BBR treatment, the FAS increase was
attenuated (Fig. 5, B and C). The reduction was observed in the
diabetic rats as well as in the non-diabetic mice. These data
suggest that BBR inhibits FAS through suppression of SREBP1
and ChREBP expression. The inhibition represents a mechanism
of BBR prevention of fatty liver.
Inhibition of mitochondrial function in hepatocytes by
Mitochondria are the major subcellular organs in oxygen
consumption. In mitochondria, oxygen is used to oxidize glucose
and fatty acids in the production of ATP. The side products are
water and carbon dioxide in mitochondria. Although ATP is also
made in glycolysis, the efficiency of ATP production is much lower
in this mitochondria-independent pathway. Glycolysis becomes
dominant when mitochondrial function is inhibited for ATP
production. Mitochondrial inhibition leads to reduction in oxygen
consumption and an increase in the AMP/ATP ratio. Therefore,
oxygen consumption and AMP/ATP ratio are two major
indicators of mitochondrial function. In an early study, we
Figure 3. FoxO1 expression in liver. (A) FoxO1 protein in liver. Nuclear and cytoplasmic proteins were extracted from liver and analyzed in a
Western blot. TBP (TATA-binding protein) and GAPDH proteins are loading controls in the nuclear and cytoplasmic proteins. (B) Immunohistostaining
of FoxO1 protein in liver. DAB dye (brown) was used to indicate the FoxO1 protein signal; (C) mRNA expression of FoxO1. mRNA was quantified in real
time RT-PCR (n=6). # P,0.05, compared with Diab+Veh group; * P,0.05, compared with normal group (Nor).
Berberine Improves Glucose Metabolism
PLoS ONE | www.plosone.org6 February 2011 | Volume 6 | Issue 2 | e16556
Figure 4. Liver steatosis in histology. (A) Hematoxylin and eosin (H & E) staining. (B) Oil Red O staining. (C) Sudan III staining. Pictures were taken
under a microscopy with 620 object lenses.
Figure 5. Lipogenesis in liver. (A) Lipogenic transcription factor proteins. Liver total protein was used in the Western blot. SREBP1, SREBP2 and
ChREBP were detected in the fasted liver with specific antibodies. Beta-actin is a loading control. (B) FAS protein. The protein was detected in a
Western blot in the fasted and fed liver, respectively. (C) FAS mRNA. mRNA was detected by qRT-PCR and normalized with beta-actin mRNA (n=6).
# P,0.05, compared with Diab+Veh group; * P,0.05, compared with normal group (Nor).
Berberine Improves Glucose Metabolism
PLoS ONE | www.plosone.org7February 2011 | Volume 6 | Issue 2 | e16556
reported that BBR inhibited mitochondrial function in adipocytes
and myotubes . The effect was not tested in hepatocytes. To test
the role of mitochondrial inhibition in liver, we examined
mitochondrial function in hepatocytes after BBR treatment in cell
culture. BBR decreased oxygen consumption and increased AMP/
ATP ratio in H4IIE cells, a rat hepatoma cell line (Fig. 6, A and B).
The data suggest that the BBR suppresses mitochondrial function
in hepatocytes. This activity may account for the reduced gene
expression for gluconeogenesis and fatty acid synthesis in liver.
Our data suggest that BBR inhibits glucose production in liver.
The inhibition is related to suppression of PEPCK and G6Pase
expression. This is a new mechanism for BBR action in the
regulation of glucose metabolism. In literature, BBR is reported to
enhance insulin sensitivity in adipocytes and muscle cells.
Activation of the AMPK pathway is a mechanism of enhanced
glucose uptake in skeletal muscle [2,5]. However, the role of
AMPK pathway is controversial for the metabolic activities of
BBR. The AMPK activity was reported in muscle (L6) and
hepatocytes (HepG2) [16,17], but the activity was not observed in
3T3-L1 adipocytes . In HepG2 cells, AMPK was reported to
mediate BBR activity in the inhibition of cholesterol and
triglyceride synthesis . However, AMPK was not tested in
BBR action in the regulation of hepatic gluconeogenesis. In the
current study, we addressed this issue by examining AMPK and
gluconeogenesis in liver. We observed that BBR improved AMPK
activity and inhibited gluconeogenic gene expression at the same
time. In the diabetic rats, BBR treatment restored AMPK activity
to the level of non-diabetic rats. The change in AMPK activity is
associated with a reduction in PEPCK and G6Pase expression.
AMPK activation is proposed as a result of mitochondria
inhibition by BBR [5,6]. Here, our data suggest that mitochondria
inhibition is also a mechanism for suppression of hepatic
gluconeogenesis and lipogenesis. The inhibition leads to suppres-
sion of gene expression and catalytic inactivation of related
enzymes in liver through depletion of ATP. Recent literature
supports that BBR increases glucagon-like peptide-1 (GLP-1) level
in vivo and in vitro [19,20]. This activity of BBR may contribute
to AMPK activation by BBR as well.
We demonstrated that insulin signaling was not significantly
modified by BBR in liver. BBR was reported to enhance glucose
uptake through GLUT4 translocation in skeletal muscle [16,21].
However, this mechanism is controversial as the GLUT4 activity
was not confirmed in other studies [5,18]. Our data suggest that
even this AMPK-GLUT4 interaction is a mechanism for BBR
action in vivo, its contribution to glucose metabolism in whole
body is limited. Our data do not support the mechanism in vivo as
the insulin sensitivity was not significantly improved by BBR in the
diabetic rats. This conclusion is supported by a recent study, in
which BBR was shown to cause muscle atrophy through induction
of protein degradation . This report argues against the BBR
activity in insulin sensitization in muscle since insulin is known to
prevent muscle atrophy [23,24]. In the current study, we
examined the insulin signaling pathway in liver by determining
Akt phosphorylation status. We did not observe a significant
change in the insulin signaling pathway following the BBR
treatment. The current study suggests that BBR may regulate
glucose metabolism independently of insulin signaling pathway in
Inhibition of mitochondria ATP production may contribute to
the suppression of hepatic gluconeogenesis and hepatic steatosis.
We observed that oxygen consumption was reduced and the
AMP/ATP ratio was increased by BBR in hepatocytes. The data
provide evidence for this new action of BBR in mitochondria
suppression in hepatocytes. This observation is consistent with
BBR activity in the purified mitochondria, in which BBR inhibited
NAD-linked respiration in mitochondria [25,26,27]. In hepato-
cytes, we observed that BBR enhanced lactate release , a sign of
the mitochondria inhibition. Here, we demonstrated that oxygen
consumption was reduced in hepatocytes by BBR, and this
reduction was associated with an increase in AMP/ATP ratio.
Though ATP production pathway is switched from mitochondrial
oxidative phosphorylation to glycolysis in the presence of BBR, the
glycolysis-generated ATP is not sufficient to maintain the
intracellular ATP level in hepatocytes. The ATP depletion should
account for inhibition of PEPCK and G6Pase expression through
inactivation of transcription factors, whose activities are dependent
The mitochondrial inhibition is a mechanism for inhibition of
lipid accumulation in liver. We observed that BBR reduced
hepatic steatosis in the diabetic rats. The mechanism is likely
inhibition of lipogenesis in liver. BBR was shown to reduce plasma
lipid and reduce plasma hepatic enzymes (ALT, AST, GGT)
[28,29]. This BBR activity was confirmed in the current study. In
term of mechanism, BBR is reported to inhibit cholesterol
synthesis and induce LDLR expression [17,26]. In the current
Figure 6. Inhibition of mitochondrial function by BBR in hepatocytes. (A) BBR decreased oxygen consumption in H4IIE hepatocytes. The
cells were plated in DMEM culture medium supplemented with 10% FBS. Oxygen consumption was determined for 12 hrs after BBR (20 mM)
treatment. (B) AMP/ATP ratio in H4IIE cells. The cells were treated with BBR in serum-free medium for 16 hours. AMP and ATP levels were determined
using HPLC. The data are presented as mean 6 SEM (n=5). * P,0.05.
Berberine Improves Glucose Metabolism
PLoS ONE | www.plosone.org8 February 2011 | Volume 6 | Issue 2 | e16556
study, we observed that BBR inhibited fatty acid synthesis in liver.
BBR inhibited FAS gene expression. Expression of lipogenic
transcription factors (SREBP and ChREBP) was also reduced,
which provides a mechanism for the FAS inhibition. Given that
ATP is required for expression and function of the lipogenic
transcription factors, the ATP depletion provides an answer to the
lipogenic inhibition by BBR. The lipogenic inhibition may
contribute to body weight loss in the BBR-treated rats observed
in this study. BBR did not reduce food intake in our study. It has
been reported that BBR does not affect expression of the appetite-
regulating neuropeptides (POMC and NPY) in the hypothalamus
in rodents . Therefore, the reduction in hepatic steatosis is not a
consequence of alteration in calorie intake.
Pharmacokinetics of BBR is different among tissues . This
character explains the long lasting AMPK activation that is
observed in the current study. Recent studies suggest that: 1) BBR
has a long half life in the liver. The BBR clearance rate is quite
different among tissues, such as hippocampus VS plasma .
According to a recent report, there is a 70-fold difference in the
BBR clearance rate in liver VS plasma in rats . The studies
suggest that BBR may have a longer half life in liver for the AMPK
activation; 2) GLP-1 may mediate BBR’s long-standing effect.
BBR increases L cell activity in expression of GLP-1 in vivo
[33,34]. The improved L cell activity may remain in the absence of
BBR for the elevated GLP-1 expression. If this is the case, GLP-1
will mediate BBR activity in the AMPK activation in liver after
BBR withdrawn. In the future, we will test this possibility by
examining GLP-1 level in our models.
In summary, we conclude here that BBR reduces fasting blood
glucose, attenuates gluconeogenesis and hepatic steatosis in type 2
diabetic rats (Fig. 7). Suppression of hepatic gluconeogenesis
provides an excellent mechanism for the reduction in blood
glucose. Inhibition of PEPCK and G6Pase expression is associated
with decreased expression of transcription factors (FoxO1,
SREBP1 and ChREBP) in the liver. BBR suppresses mitochondria
function in hepatocytes to reduce ATP level. ATP depletion is a
potential mechanism for BBR action in the inhibition of
gluconeogenesis and lipogenesis in liver. The study suggests that
inhibition of mitochondrial function in liver is responsible for the
metabolic activities of BBR.
Conceived and designed the experiments: JW J. Ye. Performed the
experiments: XX YZ KT YS DY J. Ye. Analyzed the data: XX J. Yan.
Contributed reagents/materials/analysis tools: JW LH. Wrote the paper:
XX J. Ye J. Yan.
1. Yin J, Zhang H, Ye J (2008) Traditional chinese medicine in treatment of
metabolic syndrome. Endocr Metab Immune Disord Drug Targets 8: 99–111.
2. Lee YS, Kim WS, Kim KH, Yoon MJ, Cho HJ, et al. (2006) Berberine, a
Natural Plant Product, Activates AMP-Activated Protein Kinase With Beneficial
Metabolic Effects in Diabetic and Insulin-Resistant States. Diabetes 55:
3. Kong WJ, Zhang H, Song DQ, Xue R, Zhao W, et al. (2009) Berberine reduces
insulin resistance through protein kinase C-dependent up-regulation of insulin
receptor expression. Metabolism 58: 109–119.
4. Zhang H, Wei J, Xue R, Wu JD, Zhao W, et al. (2010) Berberine lowers blood
glucose in type 2 diabetes mellitus patients through increasing insulin receptor
expression. Metabolism 59: 285–292.
5. Yin J, Gao Z, Liu D, Liu Z, Ye J (2008) Berberine Improves Glucose
Metabolism through Induction of Glycolysis. Am J Physiol Endocrinol Metab
6. Turner N, Li JY, Gosby A, To SW, Cheng Z, et al. (2008) Berberine and its
more biologically available derivative, dihydroberberine, inhibit mitochondrial
respiratory complex I: a mechanism for the action of berberine to activate AMP-
activated protein kinase and improve insulin action. Diabetes 57: 1414–1418.
7. Roesler WJ, Vandenbark GR, Hanson RW (1989) Identification of multiple
protein binding domains in the promoter-regulatory region of the phospho-
enolpyruvate carboxykinase (GTP) gene. J Biol Chem 264: 9657–9664.
8. Chakravarty K, Cassuto H, Reshef L, Hanson RW (2005) Factors that control
the tissue-specific transcription of the gene for phosphoenolpyruvate carbox-
ykinase-C. Crit Rev Biochem Mol Biol 40: 129–154.
9. Postic C, Dentin R, Girard J (2004) Role of the liver in the control of
carbohydrate and lipid homeostasis. Diabetes Metab 30: 398–408.
10. Cummins EP, Taylor CT (2005) Hypoxia-responsive transcription factors.
Pflugers Arch 450: 363–371.
11. Bi Y, Cai M, Liang H, Sun W, Li X, et al. (2009) Increased carnitine palmitoyl
transferase 1 expression and decreased sterol regulatory element-binding protein
1c expression are associated with reduced intramuscular triglyceride accumu-
lation after insulin therapy in high-fat-diet and streptozotocin-induced diabetic
rats. Metabolism 58: 779–786.
12. Uyeda K, Repa JJ (2006) Carbohydrate response element binding protein,
ChREBP, a transcription factor coupling hepatic glucose utilization and lipid
synthesis. Cell Metab 4: 107–110.
13. Eberle D, Hegarty B, Bossard P, Ferre P, Foufelle F (2004) SREBP transcription
factors: master regulators of lipid homeostasis. Biochimie 86: 839–848.
14. Yahagi N, Shimano H, Hasty AH, Matsuzaka T, Ide T, et al. (2002) Absence of
sterol regulatory element-binding protein-1 (SREBP-1) ameliorates fatty livers
but not obesity or insulin resistance in Lep(ob)/Lep(ob) mice. J Biol Chem 277:
15. Li H, Dong B, Park SW, Lee HS, Chen W, et al. (2009) Hepatocyte nuclear
factor 1alpha plays a critical role in PCSK9 gene transcription and regulation by
the natural hypocholesterolemic compound berberine. J Biol Chem 284:
16. Cheng Z, Pang T, Gu M, Gao AH, Xie CM, et al. (2006) Berberine-stimulated
glucose uptake in L6 myotubes involves both AMPK and p38 MAPK. Biochim
Biophys Acta 1760: 1682–1689.
17. Brusq JM, Ancellin N, Grondin P, Guillard R, Martin S, et al. (2006) Inhibition
of lipid synthesis through activation of AMP kinase: an additional mechanism for
the hypolipidemic effects of berberine. J Lipid Res 47: 1281–1288.
18. Zhou L, Yang Y, Wang X, Liu S, Shang W, et al. (2007) Berberine stimulates
glucose transport through a mechanism distinct from insulin. Metabolism 56:
19. Gupta NA, Mells J, Dunham RM, Grakoui A, Handy J, et al. (2010) Glucagon-
like peptide-1 receptor is present on human hepatocytes and has a direct role in
Figure 7. Schematic model of BBR signaling pathway. BBR
inhibits mitochondria function and decrease intracellular ATP. This leads
to a reduction in gluconeogenic and lipogenic transcription factors
(FoxO1, SREBP1, and ChREBP). As a result, expression of gluconeogenic
genes (PEPCK and G6Pase) and lipogenic gene (FAS) are decreased.
These molecular changes represent a signaling pathway for improve-
ment of fasting glucose and liver steatosis in the BBR-treated diabetic
Berberine Improves Glucose Metabolism
PLoS ONE | www.plosone.org9 February 2011 | Volume 6 | Issue 2 | e16556
decreasing hepatic steatosis in vitro by modulating elements of the insulin
signaling pathway. Hepatology 51: 1584–92.
20. Kimball SR, Siegfried BA, Jefferson LS (2004) Glucagon represses signaling
through the mammalian target of rapamycin in rat liver by activating AMP-
activated protein kinase. J Biol Chem 279: 54103–9.
21. Ko BS, Choi SB, Park SK, Jang JS, Kim YE, et al. (2005) Insulin sensitizing and
insulinotropic action of berberine from Cortidis rhizoma. Biol Pharm Bull 28:
22. Wang H, Liu D, Cao P, Lecker S, Hu Z (2010) Atrogin-1 affects muscle protein
synthesis and degradation when energy metabolism is impaired by the
antidiabetes drug berberine. Diabetes 59: 1879–1889.
23. Stitt TN, Drujan D, Clarke BA, Panaro F, Timofeyva Y, et al. (2004) The IGF-
1/PI3K/Akt pathway prevents expression of muscle atrophy-induced ubiquitin
ligases by inhibiting FOXO transcription factors. Mol Cell 14: 395–403.
24. Sandri M, Sandri C, Gilbert A, Skurk C, Calabria E, et al. (2004) Foxo
transcription factors induce the atrophy-related ubiquitin ligase atrogin-1 and
cause skeletal muscle atrophy. Cell 117: 399–412.
25. Mikes V, Dadak V (1983) Berberine derivatives as cationic fluorescent probes for
the investigation of the energized state of mitochondria. Biochim Biophys Acta
26. Mikes V, Yaguzhinskij LS (1985) Interaction of fluorescent berberine alkyl
derivatives with respiratory chain of rat liver mitochondria. J Bioenerg
Biomembr 17: 23–32.
27. Barreto MC, Pinto RE, Arrabaca JD, Pavao ML (2003) Inhibition of mouse liver
respiration by Chelidonium majus isoquinoline alkaloids. Toxicol Lett 146:
28. Kong W, Wei J, Abidi P, Lin M, Inaba S, et al. (2004) Berberine is a novel
cholesterol-lowering drug working through a unique mechanism distinct from
statins. Nat Med 10: 1344–1351.
29. Zhang Y, Li X, Zou D, Liu W, Yang J, et al. (2008) Treatment of type 2 diabetes
and dyslipidemia with the natural plant alkaloid berberine. J Clin Endocrinol
Metab 93: 2559–2565.
30. Wang L, Ye X, Li X, Chen Z, Chen X, et al. (2010) Metabolism, transformation
and distribution of Coptis chinensis total alkaloids in rat. Zhongguo Zhong Yao
Za Zhi 35: 2017–20.
31. Wang X, Wang R, Xing D, Su H, Ma C, et al. (2005) Kinetic difference of
berberine between hippocampus and plasma in rat after intravenous
administration of Coptidis rhizoma extract. Life Sci 77: 3058–67.
32. Liu YT, Hao HP, Xie HG, Lai L, Wang Q, et al. (2010) Extensive intestinal
first-pass elimination and predominant hepatic distribution of berberine explain
its low plasma levels in rats. Drug Metab Dispos 38: 1779–84.
33. Yu Y, Liu L, Wang X, Liu X, Liu X, et al. (2010) Modulation of glucagon-like
peptide-1 release by berberine: in vivo and in vitro studies. Biochem Pharmacol
34. Lu SS, Yu YL, Zhu HJ, Liu XD, Liu L, et al. (2009) Berberine promotes
glucagon-like peptide-1 (7-36) amide secretion in streptozotocin-induced diabetic
rats. J Endocrinol 200: 159–65.
Berberine Improves Glucose Metabolism
PLoS ONE | www.plosone.org 10 February 2011 | Volume 6 | Issue 2 | e16556