Small structural changes of the imidazopyridine diacylglycerol acyltransferase 2 (DGAT2) inhibitors produce an improved safety profile

Article abstract of DOI:10.1039/c6md00564k

Small molecule DGAT2 inhibitors have shown promise for the treatment of metabolic diseases in preclinical models. Herein, we report the first toxicological evaluation of imidazopyridine-based DGAT2 inhibitors and show that the arteriopathy associated with

Full text of DOI:10.1039/c6md00564k

MedChemComm
View Article Online
View Journal
RESEARCH ARTICLE
Small structural changes of the imidazopyridine
diacylglycerol acyltransferase 2 (DGAT2) inhibitors
produce an improved safety profile†‡
Cite this: DOI: 10.1039/c6md00564k
a
a
K. Futatsugi, K. Huard, D. W. Kung, b J. C. Pettersen,c D. A. Flynn,c J. R. Gosset,a
G. E. Aspnes,a R. J. Barnes,c S. Cabral,b M. S. Dowling,b D. P. Fernando,b
T. C. Goosen,b W. P. Gorczyca,c D. Hepworth,a M. Herr,b S. Lavergne,b Q. Li,b
M. Niosi,b S. T. M. Orr,b I. D. Pardo,c S. M. Perez,d J. Purkal,d T. J. Schmahai,c
N. Shirai,c A. M. Shoieb,c J. Zhouc and B. Goodwind
*
*
Received 12th October 2016,
Accepted 8th November 2016
Small molecule DGAT2 inhibitors have shown promise for the treatment of metabolic diseases in preclinical
models. Herein, we report the first toxicological evaluation of imidazopyridine-based DGAT2 inhibitors and
show that the arteriopathy associated with imidazopyridine 1 can be mitigated with small structural modifi-
cations, and is thus not mechanism related.
DOI: 10.1039/c6md00564k
www.rsc.org/medchemcomm
DGAT2 inhibition has been derived from antisense oligonu-
cleotides (ASO). Suppression of the DGAT2 isoform with ASO
Introduction
Hypertriglyceridemia is estimated to affect more than one
third of adults in the United States.1 Excessive accumulation
of triglycerides (TG) has been associated with metabolic disor-
ders such as coronary artery disease, type 2 diabetes mellitus
(T2DM) and non-alcoholic steatohepatitis (NASH).2–4 Diac-
ylglycerol acyltransferases 1 and 2 (DGAT1 and DGAT2) cata-
lyze the final step in TG biosynthesis, and their inhibition
has been proposed as a therapeutic strategy for the treatment
of metabolic disease.5 DGAT1 is abundantly expressed in the
intestine while DGAT2 is mainly found in liver and adipose
tissue.6–8 A number of selective small molecule inhibitors of
DGAT1 have been reported to have an attractive preclinical
profile,9–12 however, their clinical development has been hin-
dered by significant gastrointestinal side effects.13,14
improved diet-induced hyperlipidemia, hepatic steatosis and
insulin resistance in rodents.16–18 A number of small mole-
cule DGAT2 inhibitors from distinct chemical series have
been reported in the literature,19–23 and recently our group24
and others25,26 reported in vivo pharmacology results. Treat-
ment of dyslipidemic rodents with the selective DGAT2 inhib-
itor PF-06424439 (compound 1)§ caused a reduction in
plasma cholesterol as well as plasma and hepatic TG levels.24
As for any novel therapeutic mechanism and chemical series,
not only efficacy but also safety must be assessed. Reported
herein are results of safety studies with compound 1 and
close structural analogues.
During optimization of the imidazopyridine series, com-
pound 1 was identified as a potent DGAT2 inhibitor with ex-
cellent selectivity against functionally related mono-
acylglycerol acyltransferases 1, 2 and 3 (MGAT1-3) and
DGAT1,¶ good metabolic stability and high passive perme-
ability in in vitro assays (Fig. 1).‖ As previously reported, ana-
logues from this series were rapidly cleared via N-
glucuronidation of the core, and strategic introduction of a
cyclopropyl ring between the two aromatic systems supressed
clearance by this pathway.24 This modification provided com-
pound 1 with a moderate rate of clearance in rat and dog,
Targeted knock-out (KO) of Dgat2 gene in mice has been
reported to cause early postnatal death and skin barrier dys-
functions.15 Thus, much of the preclinical characterization of
a Pfizer Inc. Medicine Design, 610 Main Street, Cambridge, Massachusetts, 02155
USA. E-mail: Kim.Huard@Pfizer.com
b Pfizer Inc. Medicine Design, Eastern Point Road, Groton, Connecticut, 06340
USA. E-mail: Daniel.W.Kung@pfizer.com
c Pfizer Inc. Drug Safety Research and Development, Eastern Point Road, Groton,
Connecticut, 06340 USA
d Pfizer Inc. Cardiovascular and Metabolic Disease Research Unit, 610 Main
Street, Cambridge, Massachusetts, 02155 USA
† All authors were employed by Pfizer Inc. at the moment the experiments were
conducted.
§ PF-06424439 (compound 1) is available from Aldrich under # PZ0233.
¶ Detailed protocols for human DGAT2, DGAT1, MGAT1, MGAT2 and MGAT3 as
well as dog and rat DGAT2 enzymatic assays were described by Futatsugi.24
‖ Protocols for assessment of metabolic stability in human hepatocytes as well
as passive permeability using RRCK membranes were described previously.36,37
‡ Electronic supplementary information (ESI) available: Protocols and
supporting data for all experiments and compound preparation. See DOI:
10.1039/c6md00564k
This journal is © The Royal Society of Chemistry 2016
Med. Chem. Commun.

View Article Online
Research Article
MedChemComm
nication, we report toxicological findings for compound 1,
summarize the identification and characterization of ana-
logues with a more balanced metabolic profile compared to
compound 1, and report the differentiation of these com-
pounds in in vivo safety studies.
Results
Synthesis of compounds 1–6
Fig.
1 Summary of in vitro pharmacology and ADME data for
Compounds 1–6 were synthesized as shown in Scheme 1;
the initial route to all final compounds involved cyclization
of diamine II with imidate III. However, the chiral imidates
III (R1 ≠ R2) could only be accessed in racemic form by this
route, thus affording diastereomeric mixtures of products at
the methyl-bearing stereocenter (compounds 3–6). The dia-
stereomers were readily separated by chiral chromatography,
and this method provided efficient access to material for un-
derstanding the in vitro and PK properties of these
compounds.
In working toward safety studies, however, further op-
tions for imidazole ring formation were explored. The syn-
thesis of cyclopropyl derivative 1 was previously described.24
An alternative route to 3 and 4 made use of acid V, which
was synthesized in high enantiomeric purity by activation
and SN2-displacement of methyl lactate by 3-chloro-pyrazole.
Hydrogenation of nitro compound I in the presence of Boc2-
O afforded the amino NHBoc intermediate IV. Pyridine IVa
was acylated by acid V using T3P as the coupling reagent to
afford amide VI.33 Boc-deprotection was achieved under
acidic conditions, and subsequent addition of ammonium
acetate to buffer the acidity of the reaction mixture led to
compound 1 (PF-06424439).
and a pharmacokinetic (PK) profile suitable for in vivo experi-
ments in both species.24 Minimizing glucuronidation of the
imidazopyridine core effectively shifted the clearance mecha-
nism towards oxidative metabolism. Metabolism occurred al-
most exclusively via CYP3A4 (Fig. 1) which was identified as a
risk in compound development because drugs that are
cleared by a single enzyme pathway are more sensitive to
drug–drug interactions (DDI).27,28 No additional selectivity
risks were identified in broad panels of in vitro assays (Table
S1‡). On the basis of its overall profile,24 compound 1 was
nominated as a preclinical candidate (PF-06424439) and was
progressed to safety pharmacology and toxicology studies to
enable an investigational new drug (IND) application. In one
month toxicological studies and cardiovascular safety studies
in Beagle dogs, cardiac (coronary artery) and ileum
arteriopathy were observed, along with increases in blood
pressure, heart rate, and prolongation of QTc interval (vide
infra). Arteriopathy is a microscopically observed form of vas-
cular injury and is a significant cause of termination of com-
pounds in nonclinical assessment of safety.29 Arteriopathy
cannot be monitored in human clinical studies by noninva-
sive methods. Multiple mechanisms have been proposed to
explain vascular injury,29 however, the mechanismIJs) driving
these findings with compound 1 is (are) unknown. Due to an
insufficient therapeutic index (TI), the development of com-
pound 1 was halted. More data were required to understand
whether the toxicological findings were linked to the mecha-
nism, or were specific to the chemical series, or compound 1
only. Comparing toxicological findings with those from a dis-
tinct chemotype can help to provide insights on mechanism-
based safety and optimization of structurally differentiated
chemical series was pursued (data not shown). In addition,
within a single chemical series, several examples are reported
in the literature where a small structural change provided an
improved safety profile.30–32 Based on the desirable physico-
chemical properties and attractive in vitro profile of com-
pound 1 (Fig. 1), we continued seeking structurally close-in
analogues to be evaluated in a 1 month toxicology study in
Beagle dogs. Furthermore, while maintaining the positive at-
tributes of compound 1, analogues with a more balanced me-
tabolism profile were sought to avoid the risks associated
with an exclusively CYP3A4-driven clearance. In this commu-
cyclized products
3 or 4, with minimal epimerization
through the amidation and cyclization steps. The synthesis
from single enantiomers of methyl lactate enabled assign-
ment of absolute stereochemistry for compounds 3 and 4.
Notably, however, aminopyrimidine IVb was substantially
less reactive than aminopyridine IVa, with the result that no
amidation conditions for coupling of lactate-derivative V
were identified that would have provided an acceptably
small level of epimerization. The amino NHBoc-pyrimidine
IVb was readily converted to diamine IIb hydrochloride,
which reacted with imidate III to afford a diastereomeric
mixture of 5 and 6, which was separated by chiral chroma-
tography. For simplicity, the drawn structures of 5 and 6 de-
pict single, arbitrarily drawn diastereomers at the methyl
stereocenter, although the absolute configuration at this cen-
ter was not determined.
To ensure consistent PK properties for safety and efficacy
studies, compounds 1, 3, and 5 were isolated as crystalline
solids. Compound
methanesulfonate salt from a solution in acetonitrile.24
Compound crystallized as monohydrate mono-
1 was most readily isolated as its
3
a
hydrochloride salt from a hydrochloric acid and THF solu-
tion. Compound 5 crystallized in a neutral form from a solu-
tion in isopropanol.
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2016

View Article Online
MedChemComm
Research Article
Scheme 1 Synthesis of compounds 1–6. (a) H2, Pd/C, EtOH, (HCl); (b) HOAc, EtOH, (Et3N), 70 °C; (c) Boc2O, H2, Pd/C, EtOH; (d) T3P, EtOAc,
i
(ACH); (e) MsOH, CH2Cl2; then NH4OAc; (f) HCl, dioxane. Note: compounds 5 and 6 are distinct, single diastereomers of undetermined absolute
configuration at the R1/R2 stereocenter; one stereochemistry is arbitrarily drawn for clarity.
Table 1 In vitro human DGAT2 inhibition and clearance rate in HLM and HLM-UGT assays**
A
R1/R2
DGAT2 IC50 (nM)a
HLM (μL min−1 mg−1
)
HLM-UGT (μL min−1 mg−1
)
1
2
3
CH
CH
CH
CH
N
cPr
17
31
20
34 10
26
37
1
2
1
23.7
35.4
27.5
26.0
23.5
21.5
<1.9
ND
10.8
7.6
20.3
15.2
Me/Me
Me/H
H/Me
Me/H
H/Me
4
5b
6b
1
4
N
a
b
DGAT2 potency reported as the geometric mean of at least three replicates standard error of the mean (SEM). Absolute configuration of 5
and 6 was not determined.
Optimization of imidazopyridine analogues for a more
balanced metabolic profile
(3 and 4) maintained good in vitro potency against DGAT2
(Table 1).
While an increased rate of oxidative metabolism was ob-
served in the human liver microsome assay (HLM) for di-
methyl analogue 2, the monomethyl derivatives 3 and 4 were
oxidized at rates comparable to compound 1. In the micro-
somal assay incubated with UDP-glucuronosyltransferase
(UGT) enzyme cofactor UDP-glucuronic acid (HLM-UGT), in-
creased rates of glucuronidation were measured for 3 and 4
(as compared to 1), presumably due to the decreased steric
hindrance at the nucleophilic imidazole ring, thus providing
a more balanced in vitro metabolic profile compared to 1.
Electronic modification of the imidazopyridine core was also
pursued as a strategy to influence rates of glucuronidation.
As previously described, strategic introduction of a sp3
carbon linker between the two aromatic systems shifted the
main clearance pathway from glucuronidation of the
imidazopyridine core to CYP3A4-mediated oxidative metabo-
lism (compound 1, Table 1).24 Presumably, this shift results
from the steric hindrance caused by the cyclopropyl group
adjacent to the site of glucuronidation. To diverge metabo-
lism from CYP3A4-mediated oxidation, modifications of the
sp3 linker were evaluated with the goal of restoring some
level of glucuronidation while maintaining low rates of
oxidative clearance. Replacement of the cyclopropyl ring
with dimethyl (2) or either diastereomer of monomethyl
Incorporation of
a
nitrogen atom at the 6-position
(imidazopyrimidines 5 and 6) provided comparable potency
in the DGAT2 assay with similar low rates of oxidative metab-
olism and slightly increased rates of glucuronidation in vitro
** Protocols and guidance for assessment of metabolic stability in human liver
microsome assays were described previously.38–41
This journal is © The Royal Society of Chemistry 2016
Med. Chem. Commun.

View Article Online
Research Article
MedChemComm
(Table 1). Overall, four analogues were identified that differ
from compound 1 by only small structural changes and ex-
hibit good in vitro profiles as defined by potent DGAT2 inhi-
bition, low rates of metabolic clearance and balanced meta-
bolic profile. For the monomethyl analogues (3–6), in vitro
properties were similar for both methyl diastereomers. The
more potent diastereomer on the imidazopyridine and
imidazopyrimidine cores (compounds 3 and 5, respectively)
were progressed for advanced profiling and are described in
Tables 2 and 3.
In the dog DGAT2 enzymatic assay, compounds 3 and 5
showed IC50 values similar to those measured in the human
assay. Both compounds also inhibited the rat enzyme, al-
though with decreased potency. At a compound concentra-
tion of 10 μM, both compounds were selective for DGAT2
and inactive against broad selectivity panels, including phos-
phodiesterases (PDE), the hERG channel and functionally re-
lated MGAT1-3 and DGAT1 (Tables 2 and S1‡).§ In addition
to the balanced metabolic profile (fraction of metabolism
(fm) of UGT and CYP),34 compounds 3 and 5 showed good
passive permeability and solubility, and reasonable stability
in human hepatocytes (Table 2) which prompted their evalua-
tion in in vivo PK studies.¶
monitoring their effect on circulating TG in rats fed a high-
sucrose diet. Prior experience with prototype DGAT2 inhibi-
tors had demonstrated that the acute effects in circulating
TG were minimal in rats maintained on a standard laboratory
chow and, as a result, the pharmacodynamic effects of com-
pounds 3 and 5 were not assessed in normal rats. The stimu-
latory effect of high-sucrose diets on hepatic TG synthesis
and secretion in rats are well documented and we therefore
selected this model in order to maximize the window to as-
sess the pharmacologic response to compounds 3 and 5. In
our hands, normal rats exhibited circulating TG levels of ∼65
mg dL−1 whilst high-sucrose fed animals had values of ∼150–
330 mg dL−1. As reported in Fig. 2, oral administration of 3
or 5 caused a concentration-dependent reduction in plasma
TG, confirming in vivo engagement of the target by both
compounds.
Evaluation of compounds 1, 3 and 5 in safety pharmacology
and toxicology studies
The nonclinical safety of compound 1 was evaluated in
Wistar Han rats and Beagle dogs. Initially, 14 day exploratory
toxicity studies were conducted to identify the limits of toler-
ation, pharmacokinetics at high doses, and potential target
organs. Follow-up 1 month studies were conducted to sup-
port potential human clinical studies. The 1 month toxicity
studies were designed and conducted in accordance with
Good Laboratory Practice and Animal Welfare guidelines.
Doses of compound 1 in toxicology studies were selected to
explore suitable exposure margins to support early clinical
studies to fully evaluate DGAT2 inhibition. These doses
attained multiples of the dog in vitro DGAT2 IC50 value, with
the lowest dose of 2 mg kg−1 per day achieving 60% average
of dog DGAT2 IC50 over the dose interval. See Tables 4–6 for
more details about doses and compound exposure. At most
doses evaluated in the toxicology studies, decreased plasma
concentrations of TG and cholesterol were observed, which is
an expected result of DGAT2 inhibition. In the 14 day explor-
atory rat study (Table 4), the high dose of compound 1 (500
mg kg−1 per day) exceeded the maximum tolerated dose. Key
In vivo evaluation of pharmacokinetic and pharmacodynamic
profile
When dosed intravenously (iv) in dog PK studies, compounds
3 and 5 exhibited moderate clearance which, combined with
low volume of distribution, resulted in moderate half-life for
both analogues (Table 3). As was the case for oral administra-
tion (po) of compound 1, acceptable bioavailability was
driven by good permeability and adequate solubility for this
series, and the overall PK profiles of 3 and 5 were found to
be similar to compound 1,24 albeit with decreased bioavail-
ability. Comparable PK properties were measured in rats for
both compounds (Table S2‡).
Having demonstrated acceptable in vitro pharmacology
and in vivo PK properties, compounds 3 and 5 were advanced
to a study in rats to confirm in vivo DGAT2 inhibition by
Table 2 Summary of in vitro pharmacology and ADME data for compounds 3 and 5
Compound
3
5
DGAT2 IC50 SEM (nM)a
Human
Rat
Dog
20
155 10
1
26
1
241 18
53 15
>50b
13.1
12.0
41
4
MGAT1-3, DGAT1 IC50 (μM)
Hepatocytes CLint,app
>50b
16.1
Human
Rat
(μL min−1 per million cells)
12.7
Dog
8.4
0.39/0.61
13.1
285
0.14/0.12
<8
0.52/0.48
16.2
53
0.18/0.25
c
fmUGT/fmCYP
Papp (×10−6 cm s−1 d
)
Thermodynamic solubilitye (μg mL−1
fup (rat/dog)f
)
a
Human, dog and rat DGAT2 potency reported as the geometric mean of at least three replicates. SEM = standard error of the mean.
b
e
c
d
Maximum tested concentration was 50 μM. Fraction of metabolism of UGT and CYP using microsomal assays.34 Measured at pH 7.4.
Measured at pH 6.5 for crystalline forms described in the synthesis section. f Unbound fraction was determined using equilibrium dialysis.35
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2016

View Article Online
MedChemComm
Research Article
Table 3 In vivo pharmacokinetic profile of compounds 3 and 5 in dogsa
Dose (mg kg−1
)
Routec
T1/2 (h)
CLp (mL min−1 kg−1
)
Vdss (L kg−1
)
F (%)
3
5
1
5
1
5
ivb
3.2
3.0
1.1
3.8
4.0
0.54
NAd
0.58
NAd
NAd
75
poc
ivb
NAd
7.3
NAd
32
poc
NAd
a
Male Beagle dogs were utilized for PK studies. All data reported here are means of two experiments. The compounds were dosed in the
b
c
crystalline forms described in the synthesis section. Vehicle was 10% PEG 200/90% of 12% SBECD in water. Vehicle was 20% SBECD in
0.5% methylcellulose. d NA = not applicable.
findings included arteriopathy in the kidney and heart, gas-
tric smooth muscle degeneration, gastric ulcers, and foamy
macrophage accumulation in the lungs. Since it was thought
that there was an adequate exposure multiple relative to the
estimated clinically efficacious exposure, the 1 month rat tox-
icity studies evaluated doses of 15, 20, 30, 60 and 300 mg
kg−1 per day. At ≥60 mg kg−1 per day, an increase in inci-
dence and/or severity of microscopic findings in the heart,
mainly myocardial necrosis and fibrosis which was consistent
with rat cardiomyopathy was observed in male rats (Table 4).
In addition, an increase in heart weight was observed in
males and females at the same doses. No adverse findings
were noted at 30 mg kg−1 per day.
In the 14 day exploratory dog study, compound 1 was not
tolerated at 300 mg kg−1 per day (moribund euthanasia on
day 9). In the kidney, mild degeneration/regeneration of tu-
bular epithelium was noted at 300 mg kg−1 per day along
with arteriopathy, in an intramural artery of the heart and in
a submucosal artery of the stomach. The maximum tolerated
dose was 75 mg kg−1 per day (Table 5). In the first 1 month
toxicity dog study, doses of 10, 50 and 150 mg kg−1 per day
were evaluated. The 150 mg kg−1 per day dose was not toler-
ated (adverse clinical observations, decreased body weight
and body weight gain, decreased food consumption) and the
dogs were given a dosing holiday on study day 10. The dose
was decreased to 100 mg kg−1 per day beginning day 11 for
the remainder of the study. Subsequently, improvements in
clinical condition, body weight gain, and food consumption
were observed. Microscopic findings included coronary
arteriopathy at 50 and 150/100 mg kg−1 per day and small ar-
tery arteriopathy of the ileum at 50 mg kg−1 per day only, and
ovary at 10 mg kg−1 per day only. Mild renal tubular degener-
ation was also noted at 150/100 mg kg−1 per day. Increases in
heart rate were also observed at 50 and 150/100 mg kg−1 per
day and prolongation of QTc interval occurred at 150/100 mg
kg−1 per day on study day 25. A no observed adverse effect
level (NOAEL) was not established. Therefore, a second 1
month dog toxicity study was conducted at lower doses (2
and 5 mg kg−1 per day). In the muscularis of the esophagus
and stomach in one male at 5 mg kg−1 per day and renal cor-
tex of one female at 2 mg kg−1 per day, there was arteriopathy
of single small arteries characterized by minimal fibrinoid
necrosis,
hemorrhage
and
inflammation.
Overall,
arteriopathy was observed at all doses and a NOAEL could
not be established for compound 1 within a dose range of 2
to 150 mg kg−1 per day.
To determine if compounds 3 and 5, closely related ana-
logues of compound 1, show differentiated safety profile, 1
month toxicology studies in dogs were designed. The doses
of compounds 3 and 5 were selected to achieve similar
ranges of average DGAT2 inhibition as to that of the lower
doses of compound 1 (∼60–80% average of the dog enzy-
matic DGAT2 IC50) and the exposure levels achieved at these
doses can be found in Table 6. Although compounds 3 and 5
were structurally very similar to compound 1, they were well
tolerated at all doses in the 1 month dog study. Test article-
Fig. 2 Acute relationship between plasma free level of 3 and 5 and
the pharmacodynamic effect (TG lowering). Dose-dependent effects of
3 and 5 on plasma TG levels in sucrose-fed rats. TG levels were deter-
mined in blood drawn from the tail vein 2 h after oral administration of
the indicated dose of 3, 5 or vehicle (n = 7 or 8 animals per group).
Data are expressed as percent change from the vehicle treated group
and were analysed using one-way ANOVA followed by Dunnett's mul-
tiple comparison test. **p < 0.01, ***p < 0.001, ****p < 0.0001 com-
pared to vehicle treated animals. The values shown adjacent to the in-
dividual data points indicate the dose group (mg kg−1).
This journal is © The Royal Society of Chemistry 2016
Med. Chem. Commun.

View Article Online
Research Article
MedChemComm
Table 4 Summary of key findings for rat toxicity studies with compound 1
Study
14 day exploratorya
1 month toxicitya
Dose (po, mg kg−1 per day)
10
390
10
1190
75
None
100
4650
122
25 600
98
None
500
18 800
495
88 300
30
2020
53
4630
92
None
60
5470
144
22 600
300
13 900
366
62 500
Cmax (nM free)b
Fold rat IC50 at Cmax
AUC (ng h mL−1 free)b
Coverage of rat IC50 (%)c
Key findingsd
>99
98
>99
Arteriopathy
(kidney, 1/5 M)
(heart, 1/5 M),
cardiomyocyte
vacuolation (1/5 M)
↑ heart weight,
myocardial
necrosis/fibrosis
(1/10 M)
↑ heart weight,
myocardial
necrosis/fibrosis
(4/10 M, 1/10 F)
a
b
Exploratory groups N = 5/sex/dose, toxicity groups N = 10/sex/dose. Mean values (N = 3–4/sex/dose for toxicokinetics, male and female
c
d
systemic values combined). Average coverage (%) above DGAT2 rat IC50 at steady-state over the dosing interval of 24 hours. Incidence of
finding by sex. M = males, F = females.
Table 5 Summary of key findings for dog toxicity studies with compound 1
Study
14 day exploratorya
1 month toxicitya
Dose po
75
300
2
5
10
50
150/100
(mg kg−1 per day)
Cmax (nM free)b
Fold dog IC50 at Cmax
AUC (ng h mL−1 free)b
6540
409
24 800
99
15 300
956
78 300
>99
Unscheduled death,
arteriopathy (heart, 1/2)
(stomach, 1/2), kidney –
mild renal tubular
deg./reg. (2/2) and
vacuolation (1/2)
305
19
238
60
Arteriopathy
(kidney, 1/3F)
1010
63
1120
87
Arteriopathy
(esophagus
and stomach,
1/3 M)
2030
127
5280
97
Arteriopathy
(ovary, 1/3F)
12 700
794
57 500
>99
Arteriopathy
(heart, 1/3 M)
(ileum, 1/3F)
16 300
1019
92 600
>99
Arteriopathy
(heart, 2/3F),
kidney – mild
renal tubular
deg./reg.
c
Coverage of dog IC50
Key findingsd
None
(1/3 M, 1/3F)
a
b
c
Exploratory groups N = 1–2/sex/dose, toxicity groups N = 3/sex/dose. Mean values (male and female systemic values combined). Average
d
coverage (%) above DGAT2 dog IC50 at steady-state over the dosing interval of 24 hours. Incidence of finding by sex. M = males, F = females.
deg./reg. = degeneration/regeneration.
related findings were limited to mild clinical signs (intermit-
tent episodes of emesis), minimal body weight loss, and de-
creased food consumption, thus establishing NOAEL for both
compounds at comparable exposure levels of compound 1,
where no NOAEL was achieved.
foamy white salivation and multiple bouts of emesis for 2/8
dogs at 5 mg kg−1, 3/8 dogs at 10 mg kg−1, 5/8 dogs at 25 mg
kg−1 and all 8 dogs at 75 mg kg−1. At 5 mg kg−1, no changes
were noted in blood pressure, heart rate, ECG intervals, left
ventricular parameters and activity level. At 10 mg kg−1
,
threshold increases in systolic pressure (3 mmHg) and heart
rate (4–10 bpm) were noted while the CV effects at 25 and 75
mg kg−1 are summarized in Table 7. Oral administration of
compound 1 to dogs was associated with increases in systolic
blood pressure at doses of 10 mg kg−1 and higher that
Evaluation of compounds 1 and 3 in cardiovascular studies
In the single dose cardiovascular (CV) study in Beagle dogs,
clinical signs associated with compound 1 were limited to
Table 6 Summary of 1 month dog toxicity studies with compounds 3 and 5
Compound
3
5
Dose po (mg kg−1 per day)
3
454
11
797
65
None
10
1270
31
2680
86
None
30
3760
92
9530
96
None
3
140
3
149
21
None
10
487
9
540
50
None
30
1760
33
1560
74
None
Cmax (nM free)a
Fold dog IC50 at Cmax
AUC (ng h mL−1 free)a
Coverage of dog IC50 (%)b
Key findingsc
a
b
Mean values (male and female systemic values combined). Average coverage (%) above DGAT2 dog IC50 at steady-state over the dosing inter-
val of 24 hours. c Toxicity groups N = 2/sex/dose.
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2016

View Article Online
MedChemComm
Research Article
Table 7 Summary of in vivo cardiovascular studies with compounds 1
and 3 in dogs
close-in analogues to evaluate in subsequent toxicology stud-
ies in dogs. In this work, we report the identification of ana-
logues 3 and 5 with suitable profiles as defined by good
in vitro and in vivo DGAT2 inhibitory potency, outstanding
in vitro safety profile, balanced metabolic profile and accept-
able dog PK profile. In one month studies in dogs treated
with 3 or 5, test article-related findings were limited to eme-
sis, demonstrating that the previously observed arteriopathy
was specifically driven by compound 1 and not by DGAT2 in-
hibition or by the imidazopyridine chemical series. Con-
versely, both compounds 1 and 3 demonstrated similar CV
profiles in telemetry-instrumented dogs. Increases in blood
pressure that were small but notable and did not appear to
be dose related were observed along with dose- and exposure-
dependent increases in heart rate. Differentiated chemical
matter can be used to demonstrate that the observed in-
creases in blood pressure and heart rate are associated with
the imidazopyridine series but not with the mechanism of ac-
tion. In line with this, further efforts from our group to opti-
mize a different chemical series to fully elucidate the efficacy
and safety of DGAT2 inhibition in human will be published
separately. This report provides an example where a small
structural modification, replacing a cyclopropyl group with
methyl, is sufficient to improve the preclinical safety profile.
Compound
1
3
Dose (po, mg kg−1
Cmax (nM free)
Fold dog DGAT2 IC50
Heart rate (bpm)
Systolic pressure (mmHg)
Diastolic pressure (mmHg)
)
25a
4060
250
75a
25b
2140
52
↑10
↑6
50b
4700
110
↑15
↑5
7510
470
↑8–30
↑5
↑6–20
↑5
NOc
↑6
↑5
↑6
a
b
c
8 animals per dose group. 3 animals per dose group. NO = no
observable effect.
appeared to align with Cmax but the magnitude of increase
was not dose related (5–6 mmHg). Heart rate increased in a
dose dependent manner with maximal increases of 10, 20
and 30 bpm at 10, 25 and 75 mg kg−1, respectively.
Diastolic blood pressure was increased only at 75 mg kg−1
(Table 7). One question posed by the data generated in this
study was whether the clinical signs observed (white foamy
salivation and multiple bouts of emesis within 4 h of dosing)
were contributing partly or wholly to the observed cardiovas-
cular signs and whether this could be mitigated by avoiding
gastrointestinal irritation. A follow up study in telemetered
dogs administering compound 1 by the intravenous route
was conducted to address these questions. A bolus dose of 20
mg kg−1 iv produced a plasma exposure in line with the previ-
ous 75 mg kg−1 po dose and the observed CV profile was also
similar. A 27 bpm increase in heart rate was evident and clin-
ical signs of foamy salivation and emesis, albeit less pro-
nounced than the po route, indicated a systemic mode of ac-
tion rather than local gastrointestinal irritation (data not
shown). Systolic blood pressure was not increased.
Experimental
Synthesis and characterization of compounds 1–6, in vitro se-
lectivity of compounds 1, 3 and 5 as well as detailed method-
ologies and supporting data for pharmacokinetic, toxicology
and cardiovascular experiments are found in the ESI.‡ The
acute effect of compounds 3 and 5 on plasma TG was
assessed according to the protocol described by Futatsugi.24
All experiments and procedures involving animals were
conducted as per the guidelines and protocols reviewed and
approved by Pfizer Institutional Animal Care and Use Com-
mittee. Detailed protocols for human DGAT2, DGAT1,
MGAT1, MGAT2 and MGAT3 as well as dog and rat DGAT2
enzymatic assays were described by Futatsugi.24 Protocols for
assessment of metabolic stability in human hepatocytes as
well as passive permeability using RRCK membranes were de-
scribed previously.36,37 Protocols and guidance for assess-
ment of metabolic stability in human liver microsome assays
were described previously.38–41
Compound 3 was tested in the same model to determine
if the CV signal was evident in a close in analogue. The clini-
cal signs observed during the study were again primarily
emesis occurring within the first 4 h after dosing. Oral ad-
ministration of compound 3 caused increases in systolic and
diastolic blood pressure at all dose levels (5, 25 and 50 mg
kg−1) during the first 4 hours post dose. These effects were
relatively small in magnitude (+4 to +6 mmHg), did not in-
crease with dose level and also occurred during periods of
emesis (Table 7). No other effects were observed at 5 mg
kg−1. However, at 25 mg kg−1 compound 3 caused an increase
in heart rate (+10 bpm for 4 hours post dose) while at 50 mg
kg−1, heart rate was further increased (+15 bpm) and
persisted for up to 9 hours post dose (Table 7).
Acknowledgements
The authors would like to thank Robert Dullea and David A.
Price for help with the manuscript, Jady Rugg, Amanda King-
Ahmad and Thomas S. McDonald for bioanalytical analysis,
Paul Nkansah, Stephanie Sobotka and Samuel Bell for formu-
lation work, Brandon Pabst, Melissa Martin and Paula Loria
for the in vitro assays, Alfin Vaz for biotransformation support
and, Andrew J. Jensen and Gregory S. Goeken for their contri-
bution to solid form identification.
Conclusions
While encouraging data emerged from in vivo pharmacology
studies with compound 1, dogs treated for one month were
observed with multi-organ arteriopathy, increases in blood
pressure and heart rate and prolongation of QTc intervals.
Based on the desirable physicochemical properties and at-
tractive in vitro profile of compound 1, we sought structurally
This journal is © The Royal Society of Chemistry 2016
Med. Chem. Commun.

View Article Online
Research Article
MedChemComm
Notes and references
23 H. R. Wurie, L. Buckett and V. A. Zammit, FEBS J.,
2012, 279, 3033.
1 E. S. Ford, C. Li, G. Zhao, W. S. Pearson and A. H. Mokdad,
Arch. Intern. Med., 2009, 169, 572.
2 R. H. Unger, Annu. Rev. Med., 2002, 53, 319.
24 K. Futatsugi, D. W. Kung, S. T. M. Orr, S. Cabral, D.
Hepworth, G. Aspnes, S. Bader, J. Bian, M. Boehm, P. A.
Carpino, S. B. Coffey, M. S. Dowling, M. Herr, W. Jiao, S. Y.
Lavergne, Q. Li, R. W. Clark, D. M. Erion, K. Kou, K. Lee,
B. A. Pabst, S. M. Perez, J. Purkal, C. C. Jorgensen, T. C.
Goosen, J. R. Gosset, M. Niosi, J. C. Pettersen, J. A.
Pfefferkorn, K. Ahn and B. Goodwin, J. Med. Chem.,
2015, 58, 7173.
3 J. Friedman, Nature, 2002, 415, 268.
4 J. E. Schaffer, Curr. Opin. Lipidol., 2003, 14, 281.
5 C. L. Yen, S. J. Stone, S. Koliwad, C. Harris and R. V. Farese
Jr., J. Lipid Res., 2008, 49, 2283.
6 V. A. Zammit, Biochem. J., 2013, 451, 1.
7 S. Cases, S. J. Stone, P. Zhou, E. Yen, B. Tow, K. D.
Lardizabal, T. Voelker and R. V. Farese Jr., J. Biol. Chem.,
2001, 276, 38870.
8 C. L. Yen, M. Monetti, B. J. Burri and R. V. Farese Jr., J. Lipid
Res., 2005, 46, 1502.
9 R. J. DeVita and S. Pinto, J. Med. Chem., 2013, 56, 9820.
10 R. Naik, B. W. Obiang-Obounou, M. Kim, Y. Choi, H. S. Lee
and K. Lee, ChemMedChem, 2014, 9, 2410.
11 R. L. Dow, RSC Drug Discovery Ser., 2012, 27, 215.
12 A. M. Birch, L. K. Buckett and A. V. Turnbull, Curr. Opin.
Drug Discovery Dev., 2010, 13, 489.
13 H. Denison, C. Nilsson, L. Lofgren, A. Himmelmann, G.
Martensson, M. Knutsson, A. Al-Shurbaji, H. Tornqvist and
J. W. Eriksson, Diabetes, Obes. Metab., 2014, 16, 334.
14 B. S. Maciejewski, J. L. LaPerle, D. Chen, A. Ghosh, W. J.
Zavadoski, T. S. McDonald, T. B. Manion, D. Mather, T. A.
Patterson, M. Hanna, S. Watkins, E. M. Gibbs, R. A. Calle
and C. M. Steppan, Am. J. Physiol., 2013, 304, G958.
15 S. J. Stone, H. M. Myers, S. M. Watkins, B. E. Brown, K. R.
Feingold, P. M. Elias and R. V. Farese, J. Biol. Chem.,
2004, 279, 11767.
16 C. S. Choi, D. B. Savage, A. Kulkarni, X. X. Yu, Z. X. Liu, K.
Morino, S. Kim, A. Distefano, V. T. Samuel, S. Neschen, D.
Zhang, A. Wang, X. M. Zhang, M. Kahn, G. W. Cline, S. K.
Pandey, J. G. Geisler, S. Bhanot, B. P. Monia and G. I.
Shulman, J. Biol. Chem., 2007, 282, 22678.
17 Y. Liu, J. S. Millar, D. A. Cromley, M. Graham, R. Crooke,
J. T. Billheimer and D. J. Rader, Biochim. Biophys. Acta,
2008, 1781, 97.
18 X. X. Yu, S. F. Murray, S. K. Pandey, S. L. Booten, D. Bao,
X. Z. Song, S. Kelly, S. Chen, R. McKay, B. P. Monia and S.
Bhanot, Hepatology, 2005, 42, 362.
19 M. O. Kim, S. U. Lee, H.-J. Lee, K. Choi, H. Kim, S. Lee, S. J.
Oh, S. Kim, J. S. Kang, H. S. Lee, Y.-S. Kwak and S. Cho, Biol.
Pharm. Bull., 2013, 36, 1167.
20 M. O. Kim, S. Lee, K. Choi, S. Lee, H. Kim, H. Kang,
M. Choi, E. B. Kwon, M. J. Kang, S. Kim, H.-J. Lee,
H. S. Lee, Y.-S. Kwak and S. Cho, Biol. Pharm. Bull.,
2014, 37, 1655.
25 J. E. Imbriglio, D.-M. Shen, R. Liang, K. Marby, M. You,
H. W. Youm, Z. Feng, C. London, Y. Xiong, J. Tata, A. Verras,
M. Garcia-Calvo, X. Song, G. H. Addona, D. G. McLaren, T.
He, B. Murphy, D. E. Metzger, G. Salituro, D. Deckman, Q.
Chen, X. Jin, S. J. Stout, S.-P. Wang, L. Wilsie, O. Palyha, S.
Han, B. K. Hubbard, S. F. Previs, S. Pinto and A. Taggart,
J. Med. Chem., 2015, 58, 9345.
26 N. P. Camp and M. Naik, WO2015077299A1, 2015.
27 H. S. Brown, K. Ito, A. Galetin and J. B. Houston, Br. J. Clin.
Pharmacol., 2005, 60, 508.
28 J. H. Lin and A. Y. Lu, Clin. Pharmacokinet., 1998, 35,
361.
29 I. Mikaelian, M. Cameron, D. A. Dalmas, B. E. Enerson,
R. J. Gonzalez, S. Guionaud, P. K. Hoffmann, N. M. King,
M. P. Lawton, M. S. Scicchitano, H. W. Smith, R. A.
Thomas, J. L. Weaver and T. S. Zabka, Toxicol. Pathol.,
2014, 42, 635.
30 P. M. Scola, L. Q. Sun, A. X. Wang, J. Chen, N. Sin, B. L.
Venables, S. Y. Sit, Y. Chen, A. Cocuzza, D. M. Bilder, S. V.
D'Andrea, B. Zheng, P. Hewawasam, Y. Tu, J. Friborg, P.
Falk, D. Hernandez, S. Levine, C. Chen, F. Yu, A. K. Sheaffer,
G. Zhai, D. Barry, J. O. Knipe, Y. H. Han, R. Schartman, M.
Donoso, K. Mosure, M. W. Sinz, T. Zvyaga, A. C. Good, R.
Rajamani, K. Kish, J. Tredup, H. E. Klei, Q. Gao, L. Mueller,
R. J. Colonno, D. M. Grasela, S. P. Adams, J. Loy, P. C.
Levesque, H. Sun, H. Shi, L. Sun, W. Warner, D. Li, J. Zhu,
N. A. Meanwell and F. McPhee, J. Med. Chem., 2014, 57,
1730.
31 L. Olbe, E. Carlsson and P. Lindberg, Nat. Rev. Drug
Discovery, 2003, 2, 132.
32 M. Labelle, M. Belley, Y. Gareau, J. Y. Gauthier, D. Guay, R.
Gordon, S. G. Grossman, T. R. Jones, Y. Leblanc, M.
McAuliffe, C. McFarlane, P. Masson, K. M. Metters, N.
Ouimet, D. H. Patrick, H. Piechuta, C. Rochette, N. Sawyer,
Y. B. Xiang, C. B. Pickett, A. W. Ford-Hutchinson, R. J.
Zamboni and R. N. Young, Bioorg. Med. Chem. Lett., 1995, 5,
283.
33 J. R. Dunetz, Y. Xiang, A. Baldwin and J. Ringling, Org. Lett.,
2011, 13, 5048.
21 K. Lee, M. Kim, B. Lee, J. Goo, J. Kim, R. Naik, J. H. Seo,
M. O. Kim, Y. Byun, G. Y. Song, H. S. Lee and Y. Choi, Org.
Biomol. Chem., 2013, 11, 849.
22 X. S. Song, J. Zhang, X. Chen, O. Palyha, C. Chung, L. M.
Sonatore, L. Wilsie, S. Stout, D. G. McLaren, A. Taggart, J. E.
Imbriglio, S. Pinto, M. Garcia-Calvo and G. H. Addona,
J. Biomol. Screening, 2016, 21, 117.
34 P. J. Kilford, R. Stringer, B. Sohal, J. B. Houston and A.
Galetin, Drug Metab. Dispos., 2009, 37, 82.
35 M. J. Banker, T. H. Clark and J. A. Williams, J. Pharm. Sci.,
2003, 92, 967.
36 L. Di, C. Keefer, D. O. Scott, T. J. Strelevitz, G. Chang, Y. A.
Bi, Y. Lai, J. Duckworth, K. Fenner, M. D. Troutman and
R. S. Obach, Eur. J. Med. Chem., 2012, 57, 441.
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2016

View Article Online
MedChemComm
Research Article
37 L. Di, C. Whitney-Pickett, J. P. Umland, H. Zhang, X. Zhang,
D. F. Gebhard, Y. Lai, J. J. Federico 3rd, R. E. Davidson, R.
Smith, E. L. Reyner, C. Lee, B. Feng, C. Rotter, M. V. Varma,
S. Kempshall, K. Fenner, A. F. El-Kattan, T. E. Liston and
M. D. Troutman, J. Pharm. Sci., 2011, 100, 4974.
38 R. S. Obach, Curr. Opin. Drug Discovery Dev., 2001, 4,
36.
39 R. L. Walsky, J. N. Bauman, K. Bourcier, G. Giddens, K.
Lapham, A. Negahban, T. F. Ryder, R. S. Obach, R. Hyland
and T. C. Goosen, Drug Metab. Dispos., 2012, 40, 1051.
40 K. L. Gill, J. B. Houston and A. Galetin, Drug Metab. Dispos.,
2012, 40, 825.
41 A. Rowland, K. M. Knights, P. I. Mackenzie and J. O. Miners,
Drug Metab. Dispos., 2008, 36, 1056.
This journal is © The Royal Society of Chemistry 2016
Med. Chem. Commun.