8,8-Dimethyldihydroberberine with improved bioavailability and oral efficacy on obese and diabetic mouse models

Article abstract of DOI:10.1016/j.bmc.2010.06.085

The clinical use of the natural alkaloid berberine (BBR) as an antidiabetic reagent is limited by its poor bioavailability. Our previous work demonstrated that dihydroberberine (dhBBR) has enhanced bioavailability and in vivo efficacy compared with berberine. Here we synthesized the 8,8-dimethyldihydroberberine (Di-Me) with improved stability, and bioavailability over dhBBR. Similar to BBR and dhBBR, Di-Me inhibited mitochondria respiration, increased AMP:ATP ratio, activated AMPK and stimulated glucose uptake in L6 myotubes. In diet-induced obese (DIO) mice, Di-Me counteracted the increased adiposity, tissue triglyceride accumulation and insulin resistance, and improved glucose tolerance at a dosage of 15 mg/kg. Administered to db/db mice with a dosage of 50 mg/kg, Di-Me effectively reduced random fed and fasting blood glucose, improved glucose tolerance, alleviated insulin resistance and reduced plasma triglycerides, with better efficacy than dhBBR at the same dosage. Our work highlights the importance of dihydroberberine analogs as potential therapeutic reagents for type 2 diabetes treatment.

Full text of DOI:10.1016/j.bmc.2010.06.085

Bioorganic & Medicinal Chemistry 18 (2010) 5915–5924
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
8,8-Dimethyldihydroberberine with improved bioavailability and oral
efficacy on obese and diabetic mouse models
Zhe Cheng a, , An-Feng Chen b, , Fang Wu a, Li Sheng a, Han-Kun Zhang b, Min Gu a, Yuan-Yuan Li a,
a,
a,
*
Li-Na Zhang a, Li-Hong Hu b, , Jing-Ya Li , Jia Li
*
*
a National Center for Drug Screening, Shanghai Institute of Materia Medica, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, 189 Guo Shou Jing Road,
Shanghai 201203, PR China
b Shanghai Research Center for Modernization of Traditional Chinese Medicine, Shanghai Institute of Materia Medica, Shanghai Institutes for Biological Science,
Chinese Academy of Sciences, 199 Guo Shou Jing Road, Shanghai 201203, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
The clinical use of the natural alkaloid berberine (BBR) as an antidiabetic reagent is limited by its poor
bioavailability. Our previous work demonstrated that dihydroberberine (dhBBR) has enhanced bioavail-
ability and in vivo efficacy compared with berberine. Here we synthesized the 8,8-dimethyldihydroberb-
erine (Di-Me) with improved stability, and bioavailability over dhBBR. Similar to BBR and dhBBR, Di-Me
inhibited mitochondria respiration, increased AMP:ATP ratio, activated AMPK and stimulated glucose
uptake in L6 myotubes. In diet-induced obese (DIO) mice, Di-Me counteracted the increased adiposity,
tissue triglyceride accumulation and insulin resistance, and improved glucose tolerance at a dosage of
15 mg/kg. Administered to db/db mice with a dosage of 50 mg/kg, Di-Me effectively reduced random
fed and fasting blood glucose, improved glucose tolerance, alleviated insulin resistance and reduced
plasma triglycerides, with better efficacy than dhBBR at the same dosage. Our work highlights the impor-
tance of dihydroberberine analogs as potential therapeutic reagents for type 2 diabetes treatment.
Ó 2010 Elsevier Ltd. All rights reserved.
Received 12 May 2010
Revised 24 June 2010
Accepted 25 June 2010
Available online 1 July 2010
Keywords:
Berberine
Dihydroberberine
Di-Me
AMPK
Diet-induced obese mice
db/db mice
1. Introduction
synthesis18 and enhances glucose consumption in an insulin-inde-
pendent manner.19 In 2006, we reported that in L6 myotubes, ber-
The isoquinoline alkaloid berberine (Fig. 1A) is the main active
constituent of Rhizoma coptidis, a traditional Chinese herb for the
medication of diabetes with a history that can be traced to as early
as A.D. 500.1 The beneficial metabolic effects of berberine have
been reported by several groups in clinical studies with type 2 dia-
betes mellitus patients since 1988.2–5 Improved glucose and lipid
utility and increased insulin sensitivity were observed on several
diabetic rodent models treated with berberine, for example, ob/
ob mice,6 db/db mice,6,7 diet-induced obese rats,7–9 and alloxan-8,10
or streptozotocin-11–13 induced diabetic rats.
berine stimulates glucose uptake by activating AMP-activated
protein kinase (AMPK), a fuel gauge protein sensitive to cellular en-
ergy status, through raising cellular AMP:ATP ratio.20 Similar re-
sults were later demonstrated by several other groups on C2C12
myotubes9 and 3T3-L1 adipocytes.21,22 In the mechanism investiga-
tions, we proved that the inhibition of complex I in the mitochon-
drial respiration chain23 is a key process for the activation of
AMPK by berberine. AMPK activation results in increased glucose
absorption by peripheral tissues via promoting GLUT4 transloca-
tion,24 repressed transcription of gluconeogenesis enzyme genes
such as phosphoenolpyruvate carboxykinase and glucose-6-phos-
phatase,25 and reduced fatty acid and cholesterol synthesis by
inhibiting ACC26 and HMG-CoA reductase.27 Among these in vitro
schemes for berberine’s antidiabetic mechanism, AMPK activation
is proved to contribute at least partially to the anti-hyperglycemic
and anti-lipidemic properties of berberine in vivo.6,7
Several mechanisms were proposed to explain the anti-hyper-
glycemic effects of berberine. Berberine inhibits a-glucosidase, an
enzyme that digests polysaccharides and disaccharides into
monomers, leading to reduced intestinal absorption of carbohy-
drates.14,15 Besides, berberine suppresses adipocyte differentiation
probably through PPAR
c and C/EBPa
inhibition,16 and reduces
cholesterol through elevating the mRNA level of low-density
Nevertheless, the disadvantage of poor bioavailability28 made
high dosage (100–560 mg/kg day) necessary for berberine to yield
its efficacy in vivo.7,8,11,13 In DIO rats, 125 mg/kg twice a day was
necessary for berberine to reduce random fed and fasting blood glu-
cose,9 and 380 mg/kg/day was required for berberine to reduce
plasma triglyceride and relieve insulin resistance.7 For db/db mice,
560 mg/kg berberine was administered to reduce blood glucose
and triglyceride and relieve glucose tolerance.7 P-glycoprotein
lipoprotein receptor.17 In HepG2 cells, berberine inhibits lipid
* Corresponding authors. Tel./fax: +86 21 50802221 (L.-H.H); tel.: +86 21
50801313; fax: +86 21 50800721 (J.-Y.L); tel.: +86 21 50801552; fax: +86 21
50800721 (J.L.).
E-mail addresses: simmhulh@mail.shcnc.ac.cn (L.-H. Hu), jyli@mail.shcnc.ac.cn
(J.-Y. Li), jli@mail.shcnc.ac.cn (J. Li).
These authors contributed equally to this work.
0968-0896/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2010.06.085

5916
Z. Cheng et al. / Bioorg. Med. Chem. 18 (2010) 5915–5924
O
O
O
O
O
A
B
N
Cl
N
N
O
O
O
O
O
O
O
dhBBR
BBR
Di-Me
O
O
O
O
H+
N
N
O
O
O
O
Figure 1. (A) Structures of BBR, dhBBR, and Di-Me. (B) dhBBR converts into BBR via acid-catalyzed aromatization.
inhibitors29 and sodium caprate30 were administered together with
berberine on rodent models to improve its intestinal absorption.
Although berberine analogs were synthesized and evaluated by
several groups23,31–35 in order to find more potent derivatives, ef-
forts aiming to improve its oral absorption and pharmacokinetic
profile are relatively few. In our previous work, we designed a num-
ber of berberine derivatives and found that dhBBR (Fig. 1A) had
markedly improved in vivo efficacy in the treatment of insulin-
resistant rodents.23 In this regard, we previously developed a series
of berberine derivatives, and found that dhBBR preserved the ability
to activate AMPK and had a higher bioavailability of 2.65% com-
pared to berberine which was not detectable in the plasma.23 In
DIO mouse and rat models, dhBBR effectively counteracted the in-
creased insulin resistance and triglyceride accumulation in adipose
tissue at a much lower dosage (100 mg/kg/day). Interestingly, a
majority (90%) of absorbed drug form in rat plasma was berber-
ine,23 suggesting that dhBBR may undergo acid-catalyzed aromati-
zation after absorption. Actually, dhBBR easily converted to BBR in
acid condition in the in vitro stability study (Data not shown). This
may be also the reason why the bioavailability of dhBBR is limited
since certain part of dhBBR may convert back to BBR in the stomach
(Fig. 1B) and which hinders its the in vivo absorption.
of KOH. Treatment of 1 in the presence of POCl3 at 110 °C for 2 h
afforded 8-chloroberberine 2, which was reacted with different
Grignard reagents to yield the 8,8-dialkyldihydroberberines 3a–c.
Finally, salification of the 3a–c with the solution of hydrochloric
acid in EtOAc provided the products 4a–c.
2.2. Active dhBBR derivatives identified
Our previous study23 suggests that dhBBR is readily trans-
formed into BBR after oral administration via acid-catalyzed aro-
matization (Fig. 1B) under the acid conditions in the stomach. To
prove this hypothesis, we designed and synthesized three 8,8-dial-
kyldihydroberberine derivatives (Scheme 1) in order to improve
the chemical stability of dhBBR, and tested their cellular efficacy
on L6 myotubes. As shown in Figure 2A, from 2.5 to 10 lM range,
Di-Me induced a stronger dose-dependent AMPK/ACC phosphory-
lation than Di-Pr, while the effect of Di-Et was the strongest.
Di-Me dose-dependently stimulated glucose uptake (37–85%)
similarly with Di-Pr (63–62%), while the promotion effect of
Di-Et declined (Fig. 2B). While Di-Me and Di-Pr had almost no cell
toxicity at up to 40
toxic at as low as 10
l
l
M after 24 h treatment, Di-Et was obviously
M (Fig. 2C). So Di-Me was selected for further
Here we proposed to block aromatization of dhBBR by 8,8-disub-
stitution with alkyl groups to further optimize dhBBR into the more
aqueous soluble and acid stable 8,8-dialkyldihydroberberine
hydrochlorides.36 8,8-Dimethyl-13,13a-dihydroberberine (Di-Me)
(Fig. 1A) was identified as a promising compound. With improved
aqueous solubility and acid stability, it has significantly higher bio-
availability and is not converted back to berberine in vivo. Similar to
BBR and dhBBR, Di-Me potently activates AMPK pathway in a mito-
chondria-oriented manner, with beneficial metabolic effects in L6
myotubes. Furthermore, Di-Me improves glucose and lipid metab-
olism in diet-induced obese (DIO) mice at a relatively low dosage,
and behaves more effectively than dhBBR in db/db mice. Our work
highlights the potential importance of dihydroberberine analogs
as new candidates for type 2 diabetes therapy.
pharmacological analysis.
2.3. Di-Me activates AMPK and promotes glucose uptake via
inhibition of complex I of mitochondria respiration chain
Thus, the effects of Di-Me on L6 myotubes were investigated in
details. In 1 h, Di-Me dose-dependently increased AMPK/ACC phos-
phorylation from 1.25 to 10 lM (Fig. 3A). This effect reached
climax in as short as 10 min, and sustained for at least 3 h
(Fig. 3B). In 1 h, Di-Me dose-dependently enhanced glucose uptake
from 2.5 (48%) to 20 lM (67%, Fig. 3C), and the stimulation was
observed from 15 min (28%) to 2 h (84%) (Fig. 3D). The enhanced
glucose transport was effectively abolished by a selective AMPK
inhibitor compound C37 but not PI-3 kinase inhibitor wortmannin
(Fig. 3E). In previous studies, BBR and dhBBR activated AMPK by
increasing cellular AMP:ATP ratio through blocking complex I of
the mitochondria respiration chain.20,23 Acute dose-dependent in-
2. Results and discussion
2.1. Chemistry
crease of intracellular AMP:ATP ratio by Di-Me (2.5–10 lM) was
observed after 1 h incubation in L6 myotubes (Fig. 3F). In isolated
The synthesis of the products 4a–c is depicted in Scheme 1. The
lactam 1 was prepared by refluxing of berberine with 20% solution
rat liver mitochondria, Di-Me dose-dependently inhibited respira-
tion (2.5–40 lM) by up to 40% with pyruvate and malate as

Z. Cheng et al. / Bioorg. Med. Chem. 18 (2010) 5915–5924
5917
O
O
O
O
O
Cl
Cl
POCl3
Reflux
20% KOH
N
O
N
N
Cl
O
O
O
O
O
O
O
2
1
berberine
RMgCl
N
O
O
O
O
Cl
HCl/EtOAc
R
R
N
R
R
O
O
O
O
4a R = CH3
3a R = CH3
3b R = CH2CH3
3c R = CH2CH2CH3
4b R = CH2CH3
4c R = CH2CH2CH3
Scheme 1. Synthesis of 8,8-dialkyldihydroberberine hydrochloride.
substrates (Fig. 3G). These results indicate that Di-Me activates
AMPK through a similar mitochondria-oriented mechanism with
BBR9,20,23,38–40 and dhBBR,23 suggesting Di-Me a good candidate
for oral administration with diabetic animal models.
fat proportion reduced for 36% (Fig. 5G) and the plasma insulin de-
creased for 45% (Fig. 5H, 43% for metformin). These results demon-
strate that Di-Me exert beneficial effects on DIO mice at a dosage
much lower than dhBBR.23 Notably, these effects were accompa-
nied with significantly decreased plasma insulin level, indicating
the alleviation of insulin-resistant state. The reduced plasma glu-
cose levels may also be attributed to the possible augmentation
of glucose absorption in peripheral tissues by Di-Me.
2.4. Di-Me displays higher aqueous solubility, stability, and
bioavailability compared to dhBBR
Next, we examined the stability of Di-Me and dhBBR under acid
condition (pH 2). In 6 h, the content of dhBBR significantly declined
by 17%, while Di-Me stably kept its structure (Fig. 4A). Moreover,
Di-Me had a better aqueous solubility (81 mg/mL) over dhBBR
(<1 mg/mL). Orally administered at 10 mg/kg to SD-rats, Di-Me
was rapidly detected in the plasma at 15 min (Fig. 4B), displaying
2.6. Di-Me displays improved oral efficacy on db/db-mouse
model
The diabetic symptom of db/db mice is characterized by in-
creased body weight, blood glucose, plasma insulin, plasma TG,
etc.42 We evaluated both dhBBR and Di-Me on db/db mice at the
same dosage of 50 mg/kg. During the whole process, body weight
and food intake change of both dhBBR and Di-Me groups were
observed with no significant difference with the vehicle group. At
3 weeks, dhBBR significantly reduced random fed blood glucose
(by 25%) and Di-Me achieved a similar result (by 24%). At 4 weeks,
the decrease scope in random fed blood glucose shrunk for dhBBR
(by 18%), but the effects of Di-Me was much more potent (by 40%)
(Fig. 6A). dhBBR decreased fasting blood glucose (FBG) by 17% at
3 weeks, but fluctuated at 4 weeks. In contrast, Di-Me more po-
tently reduced FBG at 3 weeks (24%) and the reduction sustained
at 4 weeks (24%). The positive control metformin decreased FBG
at 5 weeks (Fig. 6B). While dhBBR failed to alleviate glucose toler-
ance capacity, Di-Me significantly improved glucose tolerance at 0,
15, 90, and 120 min, decreasing the area under curve by 16%
(Fig. 6C and D). dhBBR, Di-Me and metformin alleviated insulin
resistance in ITT at multiple time points (Fig. 6E), while Di-Me be-
haved more effectively, and potently reduced the area under curve
(28% for Di-Me, 12% for dhBBR, and 14% for metformin, Fig. 6F).
Furthermore, Di-Me greatly reduced plasma TG level by 38% while
dhBBR was ineffective (Fig. 6G, 23% for metformin). After 6 weeks,
we sacrificed the mice and found that dhBBR reduced liver weight
by 11% and increased subcutaneous fat by 14%, while Di-Me did
not alter the ratio of liver or fat to body weight. These results
indicate that Di-Me is more potent than dhBBR on db/db mice. This
dosage is obviously lower than the effective oral dosage of
BBR reported by previous study with db/db mice (560 mg/kg).
a half-life (t1/2) of 2.9 0.2 h, and a maximum concentration (Cmax
)
of 17.8 4.4 ng/mL. The bioavailability of Di-Me was 10.03%, much
higher than dhBBR (2.65%, administered at 20 mg/kg23). The plas-
ma form of absorbed Di-Me was mostly its original structure. No
detectable BBR or dhBBR was found in plasma. These results
proved that 8,8-substituted methyl groups protect dhBBR from
aromatization and ameliorate its in vivo absorption. Moreover,
the in vivo stability of Di-Me increases the reliability of in vitro cel-
lular assays to understand its pharmacological mechanism.
2.5. Di-Me displays improved oral efficacy on DIO-mouse model
In obesity, adipose tissue releases increased amounts of hor-
mones, inflammatory cytokines and non-esterified fatty acid that
may lead to the development of insulin resistance.41 With main-
tained cellular activity and more efficient oral absorption, Di-Me
was evaluated on DIO mice with a dosage of 15 mg/kg by daily ga-
vage for 6 weeks. During the whole process, body weight and food
intake change of the Di-Me group were observed with no signifi-
cant difference with the vehicle group. In ipGTT test, metformin
decreased blood glucose at 60 and 120 min time points, and Di-
Me significantly improved glucose tolerance at 30 and 120 min
time points (Fig. 5A), decreasing the area under curve by 15%
(Fig. 5B). In ITT test, Di-Me alleviated insulin resistance at 60 and
90 min time points (Fig. 5C). After 6-week treatment, the plasma
triglyceride level declined for 26% (Fig. 5E, 33% for metformin),
the liver triglyceride reduced for 36% (Fig. 5F), the subcutaneous

5918
Z. Cheng et al. / Bioorg. Med. Chem. 18 (2010) 5915–5924
Figure 2. Comparison of Di-Me, Di-Et, and Di-Pr. Fully differentiated L6 myotubes were serum-starved for 2 h before various treatments. (A) Effects of BBR, dhBBR, Di-Me, and
Di-Et on AMPK and ACC phosphorylation in L6 myotubes after 1 h incubation. (B) Effects of BBR, dhBBR, Di-Me, and Di-Et on glucose uptake in L6 myotubes after 1 h
incubation. (C) Cell toxicity of dhBBR, Di-Me, and Di-Et. Data are the mean results of three independent experiments SE (n = 6). *P < 0.05, **P < 0.01 compared with DMSO
group.
Interestingly, unlike BBR, Di-Me did not alter the adipose tissue
mass or body weight of db/db mice, suggesting that the peripheral
glucose absorption might be crucial for the anti-hyperglycemic ef-
fects of Di-Me.
least partially, enhanced oral efficacy of Di-Me in vivo. Our work
suggests that dihydroberberine analogs are an attractive direction
for development of potential therapeutic reagents for the treat-
ment of type 2 diabetes and metabolic syndrome.
3. Conclusion
4. Experimental
By further structural modification, we found Di-Me, a more
aqueous soluble and chemical stable dhBBR derivative. The benefi-
cial effects on both obese and diabetic mouse models indicate that
the improved aqueous solubility, stability and bioavailability, at
4.1. General experimental procedure
1H (300 MHz) and 13C (75 MHz) NMR spectra were recorded on
a Varian Mercury-VX300 Fourier transform spectrometer. The

Z. Cheng et al. / Bioorg. Med. Chem. 18 (2010) 5915–5924
5919
Figure 3. Effects of Di-Me on AMPK activation and glucose uptake in L6 myotubes. Fully differentiated L6 myotubes were serum-starved for 2 h before various treatments
with Di-Me. (A) Di-Me dose-dependently raised AMPK and ACC phosphorylation in L6 myotubes after 1 h incubation. (B) Di-Me (10 M) time-dependently stimulates AMPK
and ACC phosphorylation in L6 myotubes during 3 h incubation. (C) Di-Me dose-dependently promotes glucose uptake in L6 myotubes after 1 h treatment. (D) Di-Me time-
dependently promotes glucose uptake in L6 myotubes within 2 h. (E) Effects of wortmannin (100 nM) and compound C (10 M) on glucose uptake induced by Di-Me (10 M)
and insulin (100 nM) in L6 myotubes. (F) After 1 h incubation with increasing doses of Di-Me, intracellular AMP, ADP, and ATP levels were measured and AMP:ATP ratio was
calculated. Berberine (5 M) was employed as a positive control. (G) Effect of Di-Me on oxygen consumption of mitochondria isolated from rat liver. Data are the mean results
of three independent experiments SE (n = 5–7). P < 0.01 between groups; P < 0.05, P < 0.01 compared with DMSO group.
l
l
l
l
##
*
**
chemical shifts were reported in d (ppm) using the d 7.26 signal of
CDCl3 (1H NMR), the d 77.23 signal of CDCl3 13C NMR), the d 2.50
13C
were redistilled under positive pressure of dry nitrogen atmo-
(
sphere in the presence of proper desiccant when necessary. The
progress of the reactions was monitored by analytical thin-layer
chromatography (TLC) on HSGF254 precoated silica gel plates.
Western blotting detection kits [enhanced chemiluminescence
(ECL)] were purchased from Amersham Biosciences (Uppsala,
signal of DMSO-d6 (1H NMR), the d 39.51 signal of DMSO-d6
(
NMR), the d 4.80 signal of D2O (1H NMR) as internal standards.
HR-ESI-MS was run on a Bruker Atex III spectrometer in CH3CN,
respectively. All commercially available reagents were used with-
out further purification. The solvents used were all AR grade and
Sweden). Radiochemical 2-deoxy-[3H]-
D-glucose was purchased

5920
Z. Cheng et al. / Bioorg. Med. Chem. 18 (2010) 5915–5924
4.2.2. Preparation of the 8-chloroberberine (2)
The lactam (1) (5.0 g, 14 mmol) was dissolved in POCl3 and re-
fluxed for 2 h. The mixture was then cooled to room temperature
and filtrated. The precipitate was washed with CHCl3 and Et2O suc-
cessively to give crude (2) (4.5 g, 78%), which was used for the next
step without further purification.
4.2.3. Preparation of the hydrochloride of 8,8-dimethyldihyd-
roberberine (4a)
The 8-chloroberberine (2) (1 g, 2.47 mmol) was dissolved in
anhydrous Et2O (15 mL), to the suspension were added slowly
5 mL (15 mmol) of MeMgCl (3.0 M in THF) under Ar. When the
addition was completed, the mixture was refluxed for 1 h, the mix-
ture was then allowed to room temperature. The reaction was
quenched with saturated solution of NH4Cl at 0 °C, extracted with
dichloromethane, the organic layers were basified with aqueous
ammonia, dried over anhydrous Na2SO4, concentrated under re-
duced pressure to give the crude product, which was purified using
silica gel column chromatography (petroleum ether/ethyl ace-
tate = 30: 1) to give the compound 8,8-dimethyldihydroberberine
(3a).
The 8,8-dimethyldihydroberberine (3a), gained from the above
reaction, was dissolved in EtOAc immediately, the solution of
hydrochloric acid in EtOAc was then added dropwise. When the
addition was completed, the resulting mixture was stirred for 1 h
and filtrated, the precipitate was washed with EtOAc and dried
in high vacuum to give the compound (4a), the hydrochloride of
8,8-dimethyldihydroberberine (500 mg, 46%).
1H NMR (DMSO-d6, 300 MHz) d: 7.78 (s, 1H), 7.20 (d, J = 8.7 Hz,
1H), 7.17 (s, 1H), 7.02 (d, J = 8.7 Hz, 1H), 6.25 (s, 2H), 4.69 (s, 2H),
4.15 (t, J = 6.9 Hz, 2H), 3.90 (s, 3H), 3.85 (s, 3H), 3.06 (t, J = 6.9 Hz,
2H), 1.92 (s, 6H); 13C NMR (75 Hz) d: 169.9, 153.5, 151.9, 147.3,
144.6, 136.0, 128.8, 122.4, 120.5, 117.3, 114.0, 108.9, 107.8,
103.0, 66.9, 60.5, 56.1, 45.7, 32.0, 26.2, 25.7; HRESIMS m/z
366.1707 (Calcd for C22H24NO4, 366.1700).
Figure 4. Stability and bioavailability of Di-Me. (A) Comparison of the purity of the
DHBBR and Di-Me under acid conditions (pH 2). (B) Pharmacokinetics of Di-Me.
Sprague–Dawley rats were gavaged with Di-Me, and blood samples were collected
at 0 (pre-dosing), 0.083, 0.25, 0.5, 1, 2, 4, 6, 8, 12, and 24 h post administration. Di-
Me, BBR, and dhBBR concentrations were determined. The absolute bioavailability
reaches 10.03%.
4.2.4. 8,8-Diethyldihydroberberine (4b)
The 8,8-diethyldihydroberberine (4b) and 8,8-dipropyldihyd-
roberberine (4c) were prepared with the similar method.
8,8-Diethyldihydroberberine (4b): 1H NMR (DMSO-d6,
300 MHz) d: 7.78 (s, 1H), 7.23 (d, J = 8.4 Hz, 1H), 7.21 (s, 1H),
7.06 (d, J = 8.4 Hz, 1H), 6.26 (s, 2H), 4.89 (s, 2H), 4.13 (t,
J = 6.9 Hz, 2H), 3.85 (s, 3H), 3.84 (s, 3H), 3.06 (t, J = 6.9 Hz, 2H),
2.54 (m, 2H), 2.30 (m, 2H), 0.57 (t, J = 6.9 Hz, 6H); 13C NMR
(75 Hz) d: 172.9, 154.1, 151.8, 147.4, 143.8, 136.4, 123.8, 122.6,
120.0, 119.7, 114.2, 109.2, 107.9, 103.1, 76.5, 60.2, 56.0, 44.2,
31.9, 28.6, 25.9, 8.6; HRESIMS m/z 394.2020 (Calcd for
from PerkinElmer (Waltham, USA). Compound C were from Calbio-
chem and wortmannin were from Sigma. Anti-AMPKa, anti-acetyl-
coenzyme A carboxylase (ACC), antiphospho-ACC (Ser79), anti-
phospho-AMPK (Thr172) primary antibodies, antimouse, and anti-
rabbit IgG HRP-linked antibodies were from Cell Signal
Technology. Rodent diet with 60 kcal% fat (D12492i) was from Re-
search diets, Inc.
L6 cell culture and differentiation, Western blot, glucose uptake,
AMP:ATP ratio tests and mitochondrial were conducted as previ-
ously described.20
C24H28NO4, 394.2013).
4.2. Preparation of 8,8-dialkyldihydroberberine
4.2.5. 8,8-Dipropyldihydroberberine (4c)
8,8-Dipropyldihydroberberine (4c): 1H NMR (D2O, 300 MHz) d:
7.48 (s, 1H), 7.20 (d, J = 8.4 Hz, 1H), 7.10 (d, J = 8.4 Hz, 1H), 6.97 (s,
1H), 6.15 (s, 2H), 4.14 (t, J = 6.9 Hz 2H), 3.91 (s, 3H), 3.90 (s, 3H),
3.11 (t, J = 6.9 Hz, 2H), 2.56 (m, 2H), 2.22 (m, 2H), 1.03 (m, 2H),
0.84 (m, 2H), 0.79 (t, J = 6.6 Hz, 3H), 0.77 (t, J = 6.6 Hz, 3H); 13C
NMR (CDCl3, 75 Hz) d: 171.9, 155.4, 152.5, 148.6, 144.3, 136.5,
125.0, 123.4, 119.8, 118.1, 113.9, 109.2, 108.5, 103.3, 76.4, 60.6,
56.0, 45.7, 38.6, 32.9, 26.8, 18.1, 13.8; HRESIMS m/z 422.2334
(Calcd for C26H32NO4, 422.2326).
4.2.1. Preparation of lactam (1)
Berberine (370 mg, 1.0 mmol) was dissolved in 10 mL of the solu-
tion of 20% KOH and refluxed for 8 h. The mixture was then cooled to
room temperature and extracted with EtOAc (3Â 10 mL). The com-
bined organic layers were dried over anhydrous Na2SO4 and concen-
trated under reduced pressure to give the crude product, which was
purified using silica gel column chromatography (CHCl3/MeOH =
50:1) to give the lactam (1) (210 mg, 59%).
IR (KBr) Vmax 1641; 1H NMR (CDCl3, 300 MHz) d: 7.30 (d,
J = 8.7 Hz, 1H, H-11), 7.26 (d, J = 8.7 Hz, 1H, H-12), 7.19 (s, 1H, H-1),
6.70 (s, 1H, H-4), 6.68 (s, 1H, H-13), 5.98 (s, 2H, –OCH2O–), 4.27 (t,
J = 6.0 Hz, 2H, H-6), 3.99 (s, 3H, OCH3), 3.93 (s, 3H, OCH3), 2.89 (t,
J = 6.0 Hz, 2H, H-5); 13C NMR (75 Hz) d: 160.1, 151.4, 149.4, 148.4,
147.3, 135.6, 132.3, 130.0, 123.7, 122.4, 119.3, 118.8, 107.9, 104.7,
101.5, 101.3, 61.6, 56.8, 39.3, 28.7; ESIMS m/z 352.2 ([M+H]+).
4.3. Aqueous solubility measurements
The solubility of Di-Me in distilled H2O was determined from
undersaturation by adding excess solid phase in water. The sus-
pension was stirred with magnetic stirrers for 24 h at constant
temperature (25 0.5 °C) maintained with a water bath. The

Z. Cheng et al. / Bioorg. Med. Chem. 18 (2010) 5915–5924
5921
Figure 5. Oral efficacy of Di-Me during 6-week treatment with DIO mice with metformin as positive control. (A) Blood glucose levels in the intraperitoneal glucose tolerance
test (ipGTT) after 4 weeks of treatment. (B) Area under curve for the ipGTT after 4 weeks of treatment. (C) Blood glucose levels in the insulin tolerance test (ITT) after 4 weeks
of treatment. (D) Area under curve for ITT after 4 weeks of treatment. (E) Plasma triglyceride content after 4-week treatment. (F) Liver triglyceride content after 6-week
treatment. (G) Subcutaneous fat mass expressed as a percentage of body mass after 6-week treatment. (H) Plasma insulin content after 4-week treatment. Data are the mean
#
*
**
results SE (n = 5–8). P < 0.1, P < 0.05, P < 0.01 compared with vehicle group.
resulting solution was clarified through filter. Concentration was
measured by HPLC.
using a 5
l
m TEDAChrom. Kromasil C18 250 Â 4.6 mm column,
the chromatogram was recorded at 320 nm. The temperature of
the column oven was set at 25 °C. The elution was 58% acetonitrile
in the solution of 0.05 mol/L aqueous KH2PO4 and 0.005 mol/L SDS.
The solution of the Di-Me (0.5 mg/mL) in acetonitrile, which its pH
value was adjusted to 2 with H3PO4, was analyzed by HPLC, using
4.4. Stability determination
The solution of the dhBBR (0.5 mg/mL) in acetonitrile, which its
pH value was adjusted to 2 with H3PO4, was analyzed by HPLC,
a 5
lm TEDAChrom. Kromasil C18 250 Â 4.6 mm column. The

5922
Z. Cheng et al. / Bioorg. Med. Chem. 18 (2010) 5915–5924
Figure 6. Oral efficacy of Di-Me and dhBBR during 6-week treatment with db/db mice. C57BLKS/J-db/db male mice was subjected to 50 mg/kg dhBBR or Di-Me, and positive
metformin treatment for 6 weeks as described in Section 4. (A) Random fed blood glucose levels after 3 and 4 weeks of dhBBR or Di-Me treatment. (B) Fasting blood glucose
levels after 3 and 4 weeks of treatment. (C) Blood glucose levels in the intraperitoneal glucose tolerance test (ipGTT) after 3 weeks of treatment. (D) Area under curve for the
ipGTT after 3 weeks of treatment. (E) Blood glucose levels in the insulin tolerance test (ITT) after 4 weeks of treatment. (F) Area under curve for ITT after 4 weeks of treatment.
*
**
(G) Plasma triglyceride content after 5-week treatment. Data are presented as mean SE (n = 9–12). P < 0.05, P < 0.01.
elution of the Di-Me was 50% acetonitrile in the solution of
0.05 mol/L aqueous KH2PO4 and 0.005 mol/L SDS. Both of the flow
retro-orbital puncture into tubes with heparin anticoagulant at 0
(pre-dosing), 0.083, 0.25, 0.5, 1, 2, 4, 6, 8, and 24 h post administra-
tion. Plasma concentrations of berberine analogs were determined
by LC–MS/MS.43
rates was 1 mL/min. An auto injector was used to inject 5
test solution into the HPLC system.
lL of the
4.5. Pharmacokinetic analysis
4.6. Cytotoxicity assay
After 12 h fasting, male Sprague–Dawley rats were gavaged
with 10 mg/kg Di-Me, and blood samples were collected from
L6 myoblast cells were cultured into 96-well plates. After differ-
entiation, a range of concentrations of compounds were added, and

Z. Cheng et al. / Bioorg. Med. Chem. 18 (2010) 5915–5924
5923
the plate was incubated at 37 °C for 24 h. Then MTT was added to a
final concentration of 5 mg/mL. After 3 h incubation, the medium
was removed and 100 L DMSO was added. The absorbance at
Acknowledgments
l
This work was supported by the Chinese National Science &
Technology Major Project ‘Key New Drug Creation and Manufactur-
ing Program’ (Grants 2009ZX09301-001 and 2009ZX09102-022),
the National Natural Science Foundation of China (Grants
30925040, 90713046, 30772638, and U0633008), CAS Foundation
(Grants KSCX2-YW-R-168 and KSCX2-YW-R-179), Shanghai Com-
mission of Science and Technology (Grants 08DZ2291300,
09DZ2291200, and 09431902100), and the National Hi-Tech
Research and Development Program Grant of China (Grants
2007AA09Z402, 2007AA02Z147, and 2008AA02Z105).
550 nm was measured with a reference wavelength of 690 nm.
4.7. Mitochondria isolation and respiration measurements
Mitochondria were isolated from rat liver with a protocol re-
ported previously.44 The respiration medium contained: 225 mM
mannitol, 75 mM sucrose, 10 mM Tris–HCl, 10 mM KH2PO4,
10 mM KCl, 0.8 mM MgCl2, 0.1 mM EDTA, 0.3% fatty acid free
BSA, pH 7.0, and the respiratory control ratios (state III/state IV res-
piration) determined with 5 mM pyruvate as substrate, were
approximately 5, indicating well coupled mitochondria. Dose–re-
sponse effects of the different compounds on mitochondrial respi-
ration were determined in the presence of excess ADP (2.4 mM),
using substrate combinations targeting Complex I (5 mM pyruvate
plus 2 mM malate) of the respiratory chain.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmc.2010.06.085. These data
include MOL files and InChiKeys of the most important compounds
described in this article.
4.8. Animal studies
References and notes
All animal experiments were approved by the Animal Ethics
Committee of the Shanghai Institute of Materia Medica.
1. Yin, J.; Zhang, H.; Ye, J. Endocr. Metab. Immune Disord. Drug Targets 2008, 8, 99.
2. Ni, Y. X. Zhong Xi Yi Jie He Za Zhi 1988, 8, 711.
Male C57BL/6J mice at the age of 6 weeks were purchased from
Shanghai SLAC Laboratory Animal Co. Ltd. After 8 weeks of high fat
feeding, mice were randomized into various treatment groups by
body weight and fed glucose levels. Treatment groups were as fol-
lows: DIO vehicle (0.5% microcrystalline cellulose and 1.6% lac-
tose), Di-Me (15 mg/kg), metformin (250 mg/kg) as positive
control and chow fed vehicle. Mice were orally administered, body
weights and food intake were measured once daily. Intraperitoneal
glucose tolerance test (ipGTT) was performed by intraperitoneal
3. Zhang, Y.; Li, X.; Zou, D.; Liu, W.; Yang, J.; Zhu, N.; Huo, L.; Wang, M.; Hong, J.;
Wu, P.; Ren, G.; Ning, G. J. Clin. Endocrinol. Metab. 2008, 93, 2559.
4. Yin, J.; Xing, H.; Ye, J. Metabolism 2008, 57, 712.
5. Zhang, H.; Wei, J.; Xue, R.; Wu, J. D.; Zhao, W.; Wang, Z. Z.; Wang, S. K.; Zhou, Z.
X.; Song, D. Q.; Wang, Y. M.; Pan, H. N.; Kong, W. J.; Jiang, J. D. Metabolism 2010,
59, 285.
6. Kim, W. S.; Lee, Y. S.; Cha, S. H.; Jeong, H. W.; Choe, S. S.; Lee, M. R.; Oh, G. T.;
Park, H. S.; Lee, K. U.; Lane, M. D.; Kim, J. B. Am. J. Physiol. Endocrinol. Metab.
2009, 296, E812.
7. Lee, Y. S.; Kim, W. S.; Kim, K. H.; Yoon, M. J.; Cho, H. J.; Shen, Y.; Ye, J. M.; Lee, C.
H.; Oh, W. K.; Kim, C. T.; Hohnen-Behrens, C.; Gosby, A.; Kraegen, E. W.; James,
D. E.; Kim, J. B. Diabetes 2006, 55, 2256.
8. Tang, L. Q.; Wei, W.; Chen, L. M.; Liu, S. J. Ethnopharmacol. 2006, 108, 109.
9. Yin, J.; Gao, Z.; Liu, D.; Liu, Z.; Ye, J. Am. J. Physiol. Endocrinol. Metab. 2008, 294,
E148.
10. Gulfraz, M.; Mehmood, S.; Ahmad, A.; Fatima, N.; Praveen, Z.; Williamson, E. M.
Phytother. Res. 2008, 22, 1208.
11. Zhou, J.; Zhou, S.; Tang, J.; Zhang, K.; Guang, L.; Huang, Y.; Xu, Y.; Ying, Y.;
Zhang, L.; Li, D. Eur. J. Pharmacol. 2009, 606, 262.
12. Kong, W. J.; Zhang, H.; Song, D. Q.; Xue, R.; Zhao, W.; Wei, J.; Wang, Y. M.; Shan,
N.; Zhou, Z. X.; Yang, P.; You, X. F.; Li, Z. R.; Si, S. Y.; Zhao, L. X.; Pan, H. N.; Jiang,
J. D. Metabolism 2009, 58, 109.
13. Leng, S. H.; Lu, F. E.; Xu, L. J. Acta Pharmacol. Sin. 2004, 25, 496.
14. Pan, G. Y.; Wang, G. J.; Sun, J. G.; Huang, Z. J.; Zhao, X. C.; Gu, Y.; Liu, X. D. Yao
Xue Xue Bao 2003, 38, 911.
15. Pan, G. Y.; Huang, Z. J.; Wang, G. J.; Fawcett, J. P.; Liu, X. D.; Zhao, X. C.; Sun, J. G.;
Xie, Y. Y. Planta Med. 2003, 69, 632.
16. Huang, C.; Zhang, Y.; Gong, Z.; Sheng, X.; Li, Z.; Zhang, W.; Qin, Y. Biochem.
Biophys. Res. Commun. 2006, 348, 571.
17. Kong, W.; Wei, J.; Abidi, P.; Lin, M.; Inaba, S.; Li, C.; Wang, Y.; Wang, Z.; Si, S.;
Pan, H.; Wang, S.; Wu, J.; Wang, Y.; Li, Z.; Liu, J.; Jiang, J. D. Nat. Med. 2004, 10,
1344.
18. Brusq, J. M.; Ancellin, N.; Grondin, P.; Guillard, R.; Martin, S.; Saintillan, Y.;
Issandou, M. J. Lipid Res. 2006, 47, 1281.
19. Yin, J.; Hu, R.; Chen, M.; Tang, J.; Li, F.; Yang, Y.; Chen, J. Metabolism 2002, 51,
1439.
20. Cheng, Z.; Pang, T.; Gu, M.; Gao, A. H.; Xie, C. M.; Li, J. Y.; Nan, F. J.; Li, J. Biochim.
Biophys. Acta 2006, 1760, 1682.
injection of D-glucose (2 g/kg). For the insulin tolerance test (ITT),
mice were fasted for 6 hr, then injected intraperitoneally with
0.75 U insulin/kg body weight. Glucose levels were measured at
0, 15, 30, 60, 90, and 120 min. Both the ipGTT and the ITT tests
were carried out at 4 weeks post administration. After 6 weeks of
treatment, mice were sacrificed after a final dose and tissues were
collected for further analysis.
C57BLKS/J-db/db mice (50% male, 50% female) were introduced
from Jackson Laboratories. At the age of 6 weeks, db/db mice were
randomized into the various treatment groups by body weight and
fed glucose levels. Treatment groups were as follows: db/db vehicle
(0.5% microcrystalline cellulose, 1.6% lactose), dhBBR (50 mg/kg),
Di-Me (50 mg/kg), and metformin (250 mg/kg as positive control).
Mice were orally administered once daily. Body weights and food
intake were measured daily, and random fed blood glucose and
fasting blood glucose (FBG) were determined once a week. ITT
and ipGTT tests were performed as described above at 3 weeks
and 4 weeks, respectively. After 6 weeks of treatment, mice were
sacrificed after a final dose and tissues were collected for further
analysis.
21. Kim, S. H.; Shin, E. J.; Kim, E. D.; Bayaraa, T.; Frost, S. C.; Hyun, C. K. Biol. Pharm.
Bull. 2007, 30, 2120.
22. Zhou, L.; Yang, Y.; Wang, X.; Liu, S.; Shang, W.; Yuan, G.; Li, F.; Tang, J.; Chen,
M.; Chen, J. Metabolism 2007, 56, 405.
4.9. Metabolite analysis
Blood glucose was determined with a free-style blood glucose
monitoring system (Abbott Laboratories). Plasma and liver triglyc-
erides were determined with a triglyceride assay kit from Nanjing
Jiancheng Bioengineering Institute (Nanjing, China). Plasma insulin
was determined by ELISA (Linco Research, St. Louis, MO).
23. Turner, N.; Li, J. Y.; Gosby, A.; To, S. W.; Cheng, Z.; Miyoshi, H.; Taketo, M. M.;
Cooney, G. J.; Kraegen, E. W.; James, D. E.; Hu, L. H.; Li, J.; Ye, J. M. Diabetes 2008,
57, 1414.
24. Hawley, J. A.; Lessard, S. J. Acta Physiol. (Oxf.) 2008, 192, 127.
25. Lochhead, P. A.; Salt, I. P.; Walker, K. S.; Hardie, D. G.; Sutherland, C. Diabetes
2000, 49, 896.
26. Hardie, D. G.; Pan, D. A. Biochem. Soc. Trans. 2002, 30, 1064.
27. Gillespie, J. G.; Hardie, D. G. FEBS Lett. 1992, 306, 59.
28. Qiu, F.; Zhu, Z.; Kang, N.; Piao, S.; Qin, G.; Yao, X. Drug Metab. Dispos. 2008, 36,
2159.
4.10. Statistical analysis
29. Pan, G. Y.; Wang, G. J.; Liu, X. D.; Fawcett, J. P.; Xie, Y. Y. Pharmacol. Toxicol.
2002, 91, 193.
30. Lv, X. Y.; Li, J.; Zhang, M.; Wang, C. M.; Fan, Z.; Wang, C. Y.; Chen, L. AAPS
PharmSciTech. 2010.
One-way ANOVA followed by Student’s t-tests were applied for
multiple comparisons. All data are expressed as means SEM. The
significance threshold employed was at a level of 5% (P < 0.05).

5924
Z. Cheng et al. / Bioorg. Med. Chem. 18 (2010) 5915–5924
31. Wang, Y. X.; Wang, Y. P.; Zhang, H.; Kong, W. J.; Li, Y. H.; Liu, F.; Gao, R. M.; Liu,
T.; Jiang, J. D.; Song, D. Q. Bioorg. Med. Chem. Lett. 2009, 19, 6004.
32. Huang, L.; Shi, A.; He, F.; Li, X. Bioorg. Med. Chem. 2009.
33. Ma, Y.; Ou, T. M.; Hou, J. Q.; Lu, Y. J.; Tan, J. H.; Gu, L. Q.; Huang, Z. S. Bioorg. Med.
Chem. 2008, 16, 7582.
37. Hawley, S. A.; Gadalla, A. E.; Olsen, G. S.; Hardie, D. G. Diabetes 2002, 51, 2420.
38. Mikes, V.; Dadak, V. Biochim. Biophys. Acta 1983, 723, 231.
39. Mikes, V.; Kovar, J. Biochim. Biophys. Acta 1981, 640, 341.
40. Barreto, M. C.; Pinto, R. E.; Arrabaca, J. D.; Pavao, M. L. Toxicol. Lett. 2003, 146,
37.
34. Yang, P.; Song, D. Q.; Li, Y. H.; Kong, W. J.; Wang, Y. X.; Gao, L. M.; Liu, S. Y.; Cao,
R. Q.; Jiang, J. D. Bioorg. Med. Chem. Lett. 2008, 18, 4675.
35. Yang, Y.; Ye, X. L.; Li, X. G.; Zhen, J.; Zhang, B.; Yuan, L. Planta Med. 2007, 73,
602.
36. Moniot, J. L.; Kravetz, T. M.; Abd el Rahman, A. H.; Shamma, M. J. Pharm. Sci.
1979, 68, 705.
41. Kahn, S. E.; Hull, R. L.; Utzschneider, K. M. Nature 2006, 444, 840.
42. Kobayashi, K.; Forte, T. M.; Taniguchi, S.; Ishida, B. Y.; Oka, K.; Chan, L.
Metabolism 2000, 49, 22.
43. Chen, X.; Cui, L.; Duan, X.; Ma, B.; Zhong, D. Drug Metab. Dispos. 2006, 34,
1345.
44. Frezza, C.; Cipolat, S.; Scorrano, L. Nat. Protoc. 2007, 2, 287.