Validation of diacyl glycerolacyltransferase I as a novel target for the treatment of obesity and dyslipidemia using a potent and selective small molecule inhibitor

Article abstract of DOI:10.1021/jm7013887

A highly potent and selective DGAT-1 inhibitor was identified and used in rodent models of obesity and postprandial chylomicron excursion to validate DGAT-1 inhibition as a novel approach for the treatment of metabolic diseases. Specifically, compound 4a conferred weight loss and a reduction in liver triglycerides when dosed chronically in DIO mice and depleted serum triglycerides following a lipid challenge in a dose-dependent manner, thus, reproducing major phenotypical characteristics of DGAT-1^-/- mice.

Full text of DOI:10.1021/jm7013887

380
J. Med. Chem. 2008, 51, 380–383
Validation of Diacyl Glycerolacyltransferase I
as a Novel Target for the Treatment of
Obesity and Dyslipidemia Using a Potent and
Selective Small Molecule Inhibitor
Gang Zhao,† Andrew J. Souers,*,† Martin Voorbach,†
H. Doug Falls,† Brian Droz,† Sevan Brodjian,† Yau Yi Lau,‡
Rajesh R. Iyengar,† Ju Gao,† Andrew S. Judd,†
Seble H. Wagaw,§ Matthew M. Ravn,§
Figure 1. DGAT-1 inhibitor 1.
Kenneth M. Engstrom,§ John K. Lynch,†
Despite a significant decrease in tissue triglycerides, DGAT-1
mice have normal plasma triglyceride concentrations. This
suggests that DGAT-1 is not solely responsible for the secretion
of very low-density lipoproteins (VLDL) from the liver, and it
is possible that other enzymes are upregulated in the genetically
altered animals to compensate for its loss.7 Intestinal DGAT-1
has been shown to play a major role in postprandial chylomicron
secretion, however, as DGAT-1-/- mice show substantially
decreased levels of chylomicron-derived plasma triglycerides
following a lipid challenge.
Mathew M. Mulhern,† Jennifer Freeman,† Brian D. Dayton,†
Xiaojun Wang,† Nelson Grihalde,† Dennis Fry,†
David W. A. Beno, Kennan C. Marsh, Zhi Su,#
Gilbert J. Diaz,# Christine A. Collins,† Hing Sham,
Regina M. Reilly,† Michael E. Brune,† and Philip R. Kym†
Metabolic Disease Research, Abbott Laboratories, 100 Abbott Park
Road, Abbott Park, Illinois 60064
ReceiVed NoVember 5, 2007
These and other data indicate that an orally available small
molecule inhibitor could be capable of reducing body weight
through decreased triglyceride absorption, among other mech-
anisms, while improving insulin resistance.8 To provide further
validation for DGAT-1 inhibition as a target for the treatment
of obesity and dyslipidemia, we sought to evaluate a potent and
selective small molecule DGAT-1 inhibitor in rodent models
of metabolic disease. Specifically, we desired a compound
capable of delivering mechanism-based weight loss in DIO mice
in addition to lowering liver triglycerides, thus reproducing two
major components of the DGAT-1-/- phenotype. Additionally,
because intestinal DGAT-1 plays a major role in chylomicron
secretion,7 we felt that a compound-mediated decrease in
postprandial plasma triglycerides would provide further evidence
of mechanism-based efficacy. To this end, we optimized an
existing small molecule for potency and evaluated the resulting
lead compound in rodent models of DIO and acute postprandial
lipemia.
Bayer reported a class of biaryl keto acids, exemplified by
structure 1, that had potency against DGAT-1 in an enzymatic
assay measuring the output of triolein from diolein and
oleoylCoA (Figure 1).10 In our hands, 1 had an IC50 value of
73 and 207 nM against human and mouse DGAT-1, respec-
tively, while several analogs showed enzymatic activities in the
hundreds of nanomolar. Our initial goal was to enhance the
potency of analog 1 against both isoforms of DGAT-1.
Additionally, the relatively high molecular weight of compound
1 and related analogs prompted us to simplify the heterocyclic
terminus of the pharmacophore. We reasoned that the terminal
benzthiazole heterocycle could be opened, revealing a more
polar urea. To this end, we initiated the synthesis of urea-
terminated analogs of 1.
Abstract: A highly potent and selective DGAT-1 inhibitor was
identified and used in rodent models of obesity and postprandial
chylomicron excursion to validate DGAT-1 inhibition as a novel
approach for the treatment of metabolic diseases. Specifically, com-
pound 4a conferred weight loss and a reduction in liver triglycerides
when dosed chronically in DIO mice and depleted serum triglycerides
following a lipid challenge in a dose-dependent manner, thus, reproduc-
ing major phenotypical characteristics of DGAT-1-/- mice.
Diacyl glycerolacyltransferase 1 (DGAT-1a) is one of two
known DGAT enzymes that catalyze the final and only
committed step in triglyceride synthesis.1,2 Multiple lines of
evidence stemming from DGAT-1 deficient mice implicate this
enzyme in the development of both obesity and insulin
resistance.3–5 DGAT-1 deficient mice are resistant to diet-
induced obesity (DIO) and have decreased adiposity through a
mechanism involving increased energy expenditure. This phe-
notype is at least partially attributable to increased ambulatory
activity on a high-fat diet and an increase in the expression of
uncoupling proteins, which play a major role in nonshivering
thermogenesis in rodents. Interestingly, these animals are
hyperphagic relative to their wild-type counterparts, a possible
compensatory mechanism for the increased energy expenditure.
The DGAT-1-/- mice are also protected against diet-induced
hepatic steatosis and show decreased levels of triglycerides in
lipogenic tissues when fed a high-fat diet. Furthermore, genetic
ablation of DGAT-1 leads to increased sensitivity to insulin and
leptin, and DGAT-1 deficiency protects against insulin resistance
and obesity in agouti yellow mice.6
* To whom correspondence should be addressed. Phone: 847-937-5312.
Fax: 847-935-5165 . E-mail: andrew.souers@abbott.com.
†
Metabolic Disease Research.
Advanced Technology.
Process Research.
The synthesis route to the targeted analogs commenced with
the production of (1R,2R)-2-(4-bromo-benzoyl)-cyclopentane-
carboxylic acid methyl ester (2).9,10 Subsequent Suzuki coupling
to 4-nitrophenyl boronic acid and iron-mediated reduction
afforded the primary aniline 3 (Scheme 1). The urea-based
analogs (4) were then rendered via isocyanate coupling and
saponification.
‡
§
Exploratory Pharmacokinetics.
Integrative Pharmacology.
#
aAbbreviations: DGAT-1, diacylglycerol acyltransferase 1; DIO, diet-
induced obesity; VLDL, very low-density lipoprotein; hu, human; mse,
mouse; DMF, dimethylformamide; EtOH, ethanol; THF, tetrahydrofuran;
ACAT, acylcoenzyme A/cholesterol acyltransferase; D-fen, D-fenfluoramine;
Veh, vehicle.
Several of the urea analogs shown in Table 1 demonstrated
good to excellent enzymatic inhibition against both mammalian
10.1021/jm7013887 CCC: $40.75
2008 American Chemical Society
Published on Web 01/10/2008

Letters
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 3 381
Table 2. SAR of Stereoisomers of 4aa
Scheme 1a
b
cmpd
configuration
DGAT-1 IC50 (µM)
4a
5
6
R,R
S,S
R,S
0.007
0.096
0.098
a See Table 1, footnote a. b See Table 1, footnote b.
Table 3. Selected Pharmacokinetic Properties of 4aa
a Reagents and conditions: (a) 4,4,5,5-tetramethyl-2-(4-nitrophenyl)-1,3,2-
dioxaborolane, tris(triphenlyphosphine)palladium, potassium fluoride, DMF/
toluene/EtOH/H2O, 90 °C, o/n, 97%; (b) Fe, NH4Cl, H2O/EtOH, 65 °C,
98%; (c) isocyanate, THF; (d) LiOH, THF/H2O.
mse (10 mg/kg) rat (5 mg/kg) dog (2.5 mg/kg)
ivb
T1/2 (hr)
Vꢀ (L·kg-1 c
AUC (µg·hr·mL-1 d
Clp (L ·hr-1 ·kg-1 e
NA
NA
NA
NA
3.9
0.21
145.4
0.04
3.3
1.21
11.1
0.23
Table 1. SAR of Analogs of 4a
)
)
)
oralb
T1/2 (hr)
6.5
2.7
8.53
80.2
55.2
3.6
2.47
5.5
Cmax (µg·mL-1
)
3.51
15.2
NA
AUC (µg·hr·mL-1 d
)
F (%)f
49.2
a All values are mean values (n ) 3 unless specified otherwise). b A
total of 1% Tween-80 in water. c Volume of distribution. d Total exposure.
e Total plasma clearance. f Bioavailability.
b
c
cmpd
R
huDGAT-1 IC50 (µM) mseDGAT-1 IC50 (µM)
4a
H
0.007
0.009
0.010
0.011
0.006
0.005
0.994
0.009
0.035
0.009
0.007
0.024
0.016
0.017
0.039
0.019
0.021
NT
0.028
0.603
0.014
0.063
4b 2-CF3
4c 3-CF3
4d 3-F
4e
4f
4g 3-C(O)NHMe
4h 4-MeO
4i
4j
acyltransferases DGAT-211 and ACAT1/2,12 as activity at either
off-target could affect any potential in vivo results. No activity
was seen in either assay up to the measured concentration.
Similarly, in a CEREP profile,13 no inhibition greater than 50%
was observed against any of the receptors, ion channels, or
enzymes.
The oral pharmacokinetic properties of 4a were explored in
DIO mice, the model chosen for the preliminary in vivo
evaluation. Single dose studies with both intravenous and oral
routes were also executed in rat and dog. As demonstrated in
Table 3, the oral exposure of 4a was substantial in all species.
The rat and dog pharmacokinetic profiles were both character-
ized by high oral bioavailability and low clearance, while the
volume of distribution was significantly higher in the latter
species.
3-Cl
2-F, 5-CF3
4-CN
4-CF3
4k 3,4 OCH2O
a All compounds were >95% pure by HPLC and characterized by 1H
NMR and MS or EA. See Supporting Information for measurement number
(n) and IC50 range. b Insect cells (e.g., Sf9 or High Five) were infected for
24 to 72 h with recombinant human DGAT-1 containing an N-terminal
His6-epitope tag produced in baculovirus expression system. c Insect cells
(e.g., Sf9 or High Five) were infected for 24 to 72 h with recombinant
mouse DGAT-1 containing an N-terminal His6-epitope tag produced in
baculovirus expression system.
and mouse DGAT-1 enzymes, although they were generally
between 2- and 4-fold less potent against the latter isoform.
Remarkably, the first analog synthesized (4a, entry 1) showed
a 10-fold boost in activity against both isoforms relative to
compound 1. Hydrophobic substituents were tolerated at all
positions of the terminal ring, as various trifluoromethyl,
halogen, and methoxy-substituted analogs showed good enzy-
matic potency. Substituents that are both polar and hydrophilic
at any position were less-tolerated, as exemplified by the
terminal amide-containing compound 4g. Fortuitously, none of
the compounds synthesized showed any measurable activity
against the hERG channel in a dofetilide binding assay (data
not shown).
Because several of the initial analogs were undifferentiated
with respect to potency, we chose the first analog 4a for further
study. To confirm the optimal stereochemistry, the S,S analog
was prepared, along with the racemic cis-diastereomer. As
demonstrated in Table 2, both isomers were significantly less
potent toward the human isoform of DGAT-1, showing 10-fold
diminution in both cases. After confirming the optimal stereo-
chemical arrangement, compound 4a was screened against the
Satisfied with the oral exposure of 4a, we next explored the
effects of administration of 4a in a study measuring body weight,
food intake, and liver triglycerides in DIO mice. For a four-
week period, DIO mice fed a high-fat diet ad libitum were dosed
orally with 4a at 3 mg/kg bid, D-fenfluoramine (D-fen, 10 mg/
kg, qd), or vehicle. Food intake and body weight were measured
at days 1, 5, 7, 14, and 27 for each group. The vehicle group
showed a nonstatistically significant decrease in body weight
through day 14. After this time, the mean group weight
rebounded, ultimately resulting in a nonstatistically significant
increase by day 28. D-Fen caused a rapid decrease in body
weight until day 14, followed by a characteristic rebound.
Compound 4a administered at 3 mg/kg conferred significant
weight loss by day 7. By the end of the study, the drug-treated
animal group weighed 8.45 ( 0.02% (starting weight, 44.70 (
0.76 g; ending weight 41.44 ( 0.84 g; body weight change )
-3.35 ( 0.46 g, p < 0.01) less than the DIO vehicle (starting
weight, 44.74 ( 0.56 g; ending weight 44.95 ( 0.37 g; body
weight change ) +0.21 ( 0.52 g; Figures 1 and 2). No
significant changes in cumulative food intake were observed
for the drug-treated animals (see Supporting Information),

382 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 3
Letters
Figure 2. Effect of compound 4a (dosed at 3 mg/kg, po, bid in 1%
Tween-80 in water) and D-fen (10 mg/kg, po, qd) on the body weight
of DIO mice. Change is registered as the number of grams of body
weight difference for each measurement time point relative to day zero.
All values are mean values ( SEM for n ) 9 (*p < 0.05; **p < 0.01)
for comparisons against vehicle group.
Figure 4. (A) Dose-dependent reduction of chylomicron derived
triglycerides following drug treatment. (B) Reduction of plasma
triglycerides following drug treatment at 2, 4, and 16 h prior to corn
oil challenge. Plasma measurements were taken 1 h after the corn oil
bolus. All values are mean values ( SEM for n ) 9 (**p < 0.01).
DIO mice (bid dosing for days 1–2, fasting overnight, and qd
dosing on day 3), 4a was assessed at 0.3, 3, and 30 mg/kg.
One hour after the final dose, a corn oil bolus was administered
via gavage, and the plasma triglycerides were then measured
one hour later. As demonstrated in Figure 4A, compound 4a
shows a dose-dependent reduction of plasma triglycerides
starting at 0.3 mg/kg and continuing through the higher doses.
The percent reductions in plasma triglycerides corresponding
with the three doses are 73, 93, and 100%.15 Additionally, the
3 mg/kg dose was shown in a subsequent study to retain full
inhibition of chylomicron secretion as measured by plasma
triglyceride concentration when dosed 2, 4, and 16 h before
the corn oil challenge (Figure 4B). These data indicate that
inhibition of intestinal DGAT-1 by 4a has a prolonged action.
Compound 4a shows potent inhibition of both human and
mouse DGAT-1 isoforms, good selectivity over related acyl-
transferases, hERG, and a panel of antitargets, and good oral
pharmacokinetics in multiple species. Chronic drug treatment
in DIO mice confers a phenotype that reproduces characteristics
of the DGAT-1-/- mice with respect to body fat and liver
triglyceride mass. Furthermore, the ability of this compound to
substantially deplete serum triglycerides following a corn oil
bolus indicates an on-target mechanism of intestinal DGAT-1
inhibition. Together, this combination of data provides compel-
ling validation for the hypothesis that small molecule DGAT-1
inhibition is a viable strategy for the treatment of obesity and
various sequelae of metabolic syndrome. Further in vivo
characterization of this compound will be disclosed in due
course.
Figure 3. Reduction of liver triglycerides following treatment with
compound 4a (Veh, vehicle; D-fen). All values are mean values ( SEM
for n ) 9 (*p < 0.05; **p < 0.01), for comparisons against vehicle
group.
Table 4. Exposure Levels of 4a at 1 h (Cmax) and 17 h (C17 h) Postfinal
Dose on Day 28 (3 mg/kg, bid)
time (hours)
plasmaa (µg/mL)
1b
0.91 ( 0.07
0.21 ( 0.05
17c
a All values are mean values ( SEM. b n ) 3. c n ) 9.
indicating a possible role for increased energy expenditure.
Further studies to decipher this result are ongoing.
Consistent with the DGAT-1-/- phenotype, the liver tri-
glycerides were significantly decreased in animals treated
chronically with 3 mg/kg compound 4a (Figure 3). That the
D-fen treated animals lost weight yet showed increased liver
triglycerides provides evidence for a mechanism-based efficacy
profile of 4a, as body weight loss can occur without concomitant
loss of liver triglycerides.14
Gratifyingly, the favorable effects observed with chronic
dosing of DGAT-1 inhibitor 4a occurred at low therapeutic
levels, with peak (1 h post final dose) drug concentrations being
less than 1 µg/mL (Table 4). No measurable differences in
locomotor activity were observed in the drug-treated mice
throughout the study, indicating that treatment did not cause
any overt behavioral effects. Additionally, there were no
statistically significant changes in ALT and AST levels follow-
ing the study (data not shown).
Acknowledgment. The authors thank Jan Waters for support
with NMR analysis and Peter Dandliker and Michael Wendt
for valuable discussion.
To investigate the in vivo effects of specific inhibition of
intestinal DGAT-1, 4a was evaluated in an acute lipid challenge
model measuring chylomicron-derived plasma triglycerides
following a corn oil bolus.9 In a five-dose, three-day study in
Supporting Information Available: Experimental procedures
and characterization data for all compounds and experimental
procedure for biological assays. This material is available free of
charge via the Internet at http://pubs.acs.org.

Letters
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 3 383
(8) Subauste, A.; Burant, C. F. DGAT: Novel therapeutic target for obesity
and type 2 diabetes mellitus. Curr. Drug Targets: Immune Endocr.
Metabol. Disord. 2003, 3, 263–270.
References
(1) Cases, S.; Smith, S. J.; Zheng, Y.-W.; Myers, H. M.; Lear, S. R.;
Sande, E.; Novak, S.; Collins, C.; Welch, C. B.; Lusis, A. J.; Erikson,
S. K.; Farese, R. V., Jr. Identification of a gene encoding acyl CoA/
diacylglycerol acyltransferase, a key enzyme in triacylglycerol syn-
thesis. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 13018–13023.
(2) Cases, S.; Stone, S. J.; Zhou, P.; Yen, E.; Tow, B.; Lardizabal, K. D.;
Voelker, T.; Farese, R. V., Jr. Cloning of DGAT-2, a second
mammalian diacylglycerol acyltransferase, and related family mem-
bers. J. Biol. Chem. 2001, 276, 38870–38876.
(3) Chen, H. C.; Farese, R. V., Jr. DGAT and triglyceride synthesis: A
new target for obesity treatment? Trends CardioVasc. Med. 2000, 10,
188–192.
(4) Chen, H. C. Enhancing energy and glucose metabolism by disrupting
triglyceride synthesis: Lesson from mice lacking DGAT-1. Nutr.
Metab. 2006, 3, 10.
(9) Wilkening, D.; Mundy, B. P. A cyclopentane, cyclopentene and
cyclopenatnone annulation methodology. Synth. Commun. 1984, 14
(3), 227.
(10) Smith, R.; Campbell, A.-M.; Coish, P.; Dai, M.; Jenkins, S.; Lower,
D.; O’Connor, S.; Wang, G.; Zhang; M.; Zhu, L. Z. U.S. Patent
7091228, 2006.
(11) Yu, X. X.; Murray, S. F.; Pandey, S. K.; Booten, S. L.; Bao, D.; Song,
X.-Z.; Kelly, S.; Chen, S.; Mckay, R.; Monia, B. P.; Bhanot, S.
Antisense oligonucleotide reduction of DGAT2 expression improves
hepatic steatosis and hyperlipidemia in obese mice. Hepatology 2005,
42 (2), 362–371.
(12) Chang, C.; Dong, R.; Miyazaki, A.; Sakashita, N.; Zhang, Y.; Liu, J.;
Guo, M.; Li, B.-L.; Chang, T.-Y. Human Acyl-CoA/cholesterol
acyltransferase (ACAT) and its potential as a target for pharmaceutical
intervention against atherosclerosis. Acta Biochim. Biophys. Sin. 2006,
38 (3), 151–156.
(5) Smith, S. J.; Cases, S.; Jensen, D. R.; Chen, H. C.; Sande, E.; Tow,
B.; Sanan, D. A.; Raber, J.; Eckel, R. J.; Farese, R. V., Jr. Obeisty
resistance and multiple mechanisms of triglyceride synthesis in mice
lacking DGAT. Nat. Genet. 2000, 87–90.
(6) Chen, H. C.; Smith, S. J.; Ladha, Z.; Jensen, D. R.; Ferreira, L. D.;
Pulawa, L. K.; McGuire, J. G.; Pitas, R. E.; Eckel, R. H.; Farese,
R. V., Jr. Increased insulin and leptin sensitivity in mice lacking
acylCoA:diacylglycerol acyltransferase 1. J. Clin. InVest. 2002, 109,
1049–1055.
(7) Buhman, K. K.; Smith, S. J.; Stone, S. J.; Repa, J. J.; Wong, J. S.;
Knapp, F. F., Jr.; Burri, B. J.; Hamilton, R. L.; Abumrad, N. A.; Farese,
R. V., Jr J. Biol. Chem. 2002, 277 (28), 25474–25579.
(13) Cross-reactivity assays were performed by CEREP; www.cerep.com.
(14) Liver triglycerides were determined on day 28 only.
(15) The 30 mg/kg dose decreases the plasma triglyceride levels to those
of the saline-gavaged animals. Thus, the percentage inhibition of the
0.3 and 3.0 mg/kg doses are relative to the 30 mg/kg dose.
JM7013887