Investigation of cardiovascular effects of tetrahydro-β-carboline sstr3 antagonists

Article abstract of DOI:10.1021/ml500028c

Antagonism of somatostatin subtype receptor 3 (sstr3) has emerged as a potential treatment of Type 2 diabetes. Unfortunately, the development of our first preclinical candidate, MK-4256, was discontinued due to a dose-dependent QTc (QT interval corrected for heart rate) prolongation observed in a conscious cardiovascular (CV) dog model. As the fate of the entire program rested on resolving this issue, it was imperative to determine whether the observed QTc prolongation was associated with hERG channel (the protein encoded by the human Ether-à-go-go-Related Gene) binding or was mechanism-based as a result of antagonizing sstr3. We investigated a structural series containing carboxylic acids to reduce the putative hERG off-target activity. A key tool compound, 3A, was identified from this SAR effort. As a potent sstr3 antagonist, 3A was shown to reduce glucose excursion in a mouse oGTT assay. Consistent with its minimal hERG activity from in vitro assays, 3A elicited little to no effect in an anesthetized, vagus-intact CV dog model at high plasma drug levels. These results afforded the critical conclusion that sstr3 antagonism is not responsible for the QTc effects and therefore cleared a path for the program to progress.

Full text of DOI:10.1021/ml500028c

Letter
pubs.acs.org/acsmedchemlett
Investigation of Cardiovascular Effects of Tetrahydro-β-carboline
sstr3 antagonists
Shuwen He,*,† Zhong Lai,† Zhixiong Ye,† Peter H. Dobbelaar,† Shrenik K. Shah,† Quang Truong,†
Wu Du,† Liangqin Guo,† Jian Liu,† Tianying Jian,† Hongbo Qi,† Raman K. Bakshi,† Qingmei Hong,†
James Dellureficio,† Mikhail Reibarkh,† Koppara Samuel,‡ Vijay B. Reddy,‡ Stan Mitelman,‡
Sharon X. Tong,‡ Gary G. Chicchi,§ Kwei-Lan Tsao,§ Dorina Trusca,§ Margaret Wu,§ Qing Shao,§
Maria E. Trujillo,§ Guillermo Fernandez,◇ Donald Nelson,◇ Patricia Bunting,◇ Janet Kerr,◇
Patrick Fitzgerald,◇ Pierre Morissette,◇ Sylvia Volksdorf,◇ George J. Eiermann,§ Cai Li,§ Bei Zhang,§
Andrew D. Howard,§ Yun-Ping Zhou,§ Ravi P. Nargund,† and William K. Hagmann†
†
‡
§
Merck Research Laboratories, Departments of Medicinal Chemistry, Drug Metabolism and Pharmacokinetics, and Diabetes
Research, 2000 Galloping Hill Road, Kenilworth, New Jersey 07033, United States
◇Department of Safety Assessment, 770 Sumneytown Pike, West Point, Pennsylvania 19486, United States
S
* Supporting Information
ABSTRACT: Antagonism of somatostatin subtype receptor 3 (sstr3) has emerged as
a potential treatment of Type 2 diabetes. Unfortunately, the development of our first
preclinical candidate, MK-4256, was discontinued due to a dose-dependent QTc (QT
interval corrected for heart rate) prolongation observed in a conscious cardiovascular
(CV) dog model. As the fate of the entire program rested on resolving this issue, it was
imperative to determine whether the observed QTc prolongation was associated with
hERG channel (the protein encoded by the human Ether-a-go-go-Related Gene)
̀
binding or was mechanism-based as a result of antagonizing sstr3. We investigated a
structural series containing carboxylic acids to reduce the putative hERG off-target
activity. A key tool compound, 3A, was identified from this SAR effort. As a potent sstr3 antagonist, 3A was shown to reduce
glucose excursion in a mouse oGTT assay. Consistent with its minimal hERG activity from in vitro assays, 3A elicited little to no
effect in an anesthetized, vagus-intact CV dog model at high plasma drug levels. These results afforded the critical conclusion that
sstr3 antagonism is not responsible for the QTc effects and therefore cleared a path for the program to progress.
KEYWORDS: sstr3, antagonist, Type-2 diabetes, β-tetrahydrocarboline, carboxylic acid, hERG channel, QTc prolongation,
cardiovascular dog models
omatostatin receptor 3 (sstr3) is a member of a group of
S
five G-protein coupled somatostatin receptors (sstr1−
sstr5).1 Two different structural classes of selective small
molecule sstr3 antagonists have been reported: imidazolyl-
tetrahydro-β-carbolines derived from D-tryptophan (D-Trp) and
substituted decahydroisoquinolines.2−6 Recently, we disclosed
that antagonism of sstr3 represents a potential novel
mechanism for the treatment of Type-2 diabetes mellitus
(T2DM) through the evaluations of compound 1 in both in
vitro assays and animal efficacy models (Figure 1).7
Subsequently, we reported that the optimization of this
tetrahydro-β-carboline series led to the discovery of MK-4256
(Figure 1).8 MK-4256 possesses excellent sstr3 potency and
subtype selectivity, a good pharmacokinetic (PK) profile in
preclinical species, and superior efficacy in a mouse oGTT
assay. Although, MK-4256 was shown to have high protein
plasma binding with ∼1% free fractions across several species
(Table 1), only a minimal shift (∼2×) was observed from in
vitro assays with 20% human serum added (Figure 1).9 This
lack of serum shift on sstr3 was attributed to a slow dissociation
Figure 1. Tetrahydro-β-carboline sstr3 antagonists.
rate of MK-4256 from the receptor. More significantly, MK-
4256 reduced glucose excursion by 86% in a mouse oGTT
assay at a dose as low as 0.1 mg/kg with the maximal plasma
concentration of 88 nM.8 On the other hand, MK-4256 had
Received: January 22, 2014
Accepted: April 21, 2014
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ml500028c | ACS Med. Chem. Lett. XXXX, XXX, XXX−XXX

ACS Medicinal Chemistry Letters
Letter
a
Indeed, in anesthetized, vagus-intact dogs, iv fusion of MK-
4256 did result in a dose-dependent increased QTc interval
(Table 2). This result suggested that anesthesia was not
involved but that MK-4256 exerted these unwanted effects
through an unknown mechanism involving vagal signaling.
Nevertheless, since QTc prolongation was observed with MK-
4256 in the anesthetized, vagus-intact dog model, this model
was deemed to be more appropriate for assessing the CV effects
of this series of compounds than the anesthetized, vagotomized
dog model. Therefore, the anesthetized, vagus-intact dog model
was subsequently adopted to screen later lead compounds prior
to a more definitive evaluation in the more resource-intensive
conscious CV dog telemetry model.
The observed QTc prolongation for MK-4256 cast a serious
doubt on the entire sstr3 antagonist program. It was crucial to
elucidate whether the QTc prolongation was caused by
inhibition of the hERG channels (i.e., MK-4256 is deficient)
or the adverse effect was due to the antagonism of sstr3. The
latter scenario would lead to an immediate termination of the
sstr3 program. It is well-known that somatostatin receptors are
coupled to several types of potassium channels, including the
delayed rectifier, inward rectifier and ATP-sensitive potassium
channels.1 It was reported that octreotide, a synthetic analogue
of somatostatin, is able to improve and even normalize the QT
prolongation frequently observed in acromegalic patients.19
Given that sstr agonism with octreotide was able to reduce QTc
prolongation, an antagonist of sstr might have the potential to
produce QTc prolongation. In order to define the path forward
for sstr3 antagonists as a potential treatment for T2DM, it was
critical to confirm that antagonism of sstr3 alone does not cause
QTc prolongation. Therefore, further SAR exploration focused
on identifying compounds with little or no inhibition of the
hERG channel but still engaged sstr3.
Table 1. In Vitro Plasma Protein Binding of MK-4256
MK-4256 plasma protein binding (%)
human
99.2
monkey
99.4
dog
rat
mouse
98.2
99.2
97.4
a
Determined by equilibirum dialysis.
modest activity on the hERG channel (the protein encoded by
the human Ether-a-go-go-Related Gene), which raised some
̀
concerns from safety perspectives.10,11 MK-4256 inhibits
radiolabeled MK-499 binding of the hERG channel with an
IC50 = 1.74 μM.12,13 These results represent >2000-fold
selectivity for sstr3 over hERG binding. In a functional patch
clamp assay (PatchXpress), MK-4256 exhibited 50% blockade
of hERG at 3.4 μM concentration.14 When MK-4256 was
dosed in an anesthetized, vagotomized cardiovascular (CV) dog
model, there were no cardiovacular findings at plasma levels up
to 13.7 μM.15 There seemed to be a shift in hERG activity of
MK-4256 from in vitro assays to the CV dog models, probably
due to its high plasma protein binding. Based on its highly
attractive activity on sstr3 and sufficient margins expected
against the hERG channel, MK-4256 was selected as a
preclinical development candidate.
During the course of toxicological evaluation, MK-4256
exhibited a dose-dependent prolongation of the QTc interval
(QT interval corrected for heart rate) in a conscious
cardiovascular (CV) dog telemetry model, which is widely
considered to be the “gold standard” for assessing the CV
effects including QTc prolongation (Table 2).16,17 Increases in
Table 2. Effect of MK-4256 in CV Dog Models
MK-4256 Caused QTc Prolongation in a Conscious CV Dog Model
oral dose of MK-4256 (mg/kg)
plasma Cmax
QTc
increase
It has been well documented in the literature that
introduction of a carboxylic acid moiety into a structure can
mitigate binding to the hERG channels.20−23 In these reported
examples, some carboxylic acid compounds were able to
maintain the activity on the biological targets while minimizing
the interaction with the hERG channel.
a
(μM)
20
4.6
32
5%
100
500
10%
14%
b
ND
Effect of MK-4256 in Anesthetized, Vagus-Intact CV Dog Model
Our initial introduction of a carboxylic acid group into the
indole ring or the fluorophenyl imidazole moiety of the
structure abolished the activity on sstr3. Our earlier experience
indicated that the C-1 position of the β-carboline is flexible
enough to accommodate different functional groups. Indeed,
this area of the molecule tolerates an appended carboxylic acid
as well as a variety of aryls and 5 or 6- membered heteroaryls.
We decided that it was synthetically more expedient to focus on
the compounds with a carboxylic acid group attached to a
pyridyl ring at the C-1 position rather than oxadiazole-
carboxylic acid analogues while keeping the N-methyl pyrazole
moiety existing in the structure of MK-4256. Earlier SAR results
indicated that phenyl derivatives at C-1 had significant
problems with hERG activities. As the result, we prepared the
pyridyl carboxylic acid compounds (2A-5A and 2B-5B) from
their corresponding ethyl ester precursors (6A-9A and (6B-9B)
(Figure 2). The A and B series differ by their configuration at
the C-1 position.
dosage of MK-4256 by iv infusion
(mg/kg/30 min)
plasma Cmax
QTc
a
(μM)
increase
10
20
30
18
45
92
6%
8%
10%
a
Plasma drug levels were determined from separate PK studies in
b
dogs. Not determined.
the QTc interval of 5, 10, and 14% were noted after a single
oral dosing of 20, 100, and 500 mg/kg (at plasma concentration
of 4.6 and 32 μM for 20 and 100 mg/kg, respectively). Based
on these observed CV adverse effects, further development of
MK-4256 was discontinued.
The adverse finding of MK-4256 in a conscious CV dog
telemetry model was an unexpected observation as a re-
evaluation in anesthetized, vagotemized dogs at higher doses
afforded no QTc interval increases after intravenous (iv) dosing
at plasma concentrations up to 100 μM. These results indicated
that an anesthetized, vagotemized CV dog model may not be
predicative. Subsequently, MK-4256 was tested in an
anesthetized, vagus-intact CV dog model, since our studies
with a few compounds well-documented for their QTc
prolongation effects indicated there was a trend of higher
sensitivity in the vagus-intact dogs to QTc prolongation.18
The synthesis of the compounds followed the procedure
reported previously (Scheme 1). The Pictect-Spengler reaction
of fluorophenyl imidazole intermediate 10 with the correspond-
ing ketones gave the product as a mixture of diastereomers,
which were separated by chiral HPLC or SFC to give the ethyl
esters as single compounds (6A-9A and 6B-9B). Each ethyl
B
dx.doi.org/10.1021/ml500028c | ACS Med. Chem. Lett. XXXX, XXX, XXX−XXX

ACS Medicinal Chemistry Letters
Letter
Figure 2. Imidazole tetrahydro-β-carboline incorporating a pyridyl carboxylic acid moiety.
a
The in vitro profiles of ethyl esters 6A-9A and 6B-9B are
summarized in Table 3. In general, the compounds in the A
series (6A-9A) are slightly more potent on sstr3 than their
corresponding diastereomers in the B series (6B-9B) as
measured in the competitive binding assay and the functional
antagonism assay using recombinant human or mouse sstr3.
Some compounds even achieve subnanomolar potencies (e.g.,
7A). However, all these ester compounds suffer from strong
binding to the hERG channel as observed in the MK-499
binding assay, with the binding affinities comparable to that of
MK-4256.
The profiles of the carboxylic acid compounds are
summarized in Table 4. Following the same SAR trend as
observed in the ethyl ester series, the carboxylic acid
compounds in the A series (2A-5A) are more potent than
their corresponding diastereomers 2B-5B. However, the
carboxylic acid compounds are all less potent than their
corresponding ethyl ester precursors. It appears that sstr3 is less
tolerant of a carboxylic acid group compared with neutral
moieties such as an ethyl ester group. Fortunately, compounds
2A, 3A and 4A still maintain single digit nanomolar potencies
in the human sstr3 binding assay. Compound 5A is much less
potent than 2A, 3A and 4A, probably due to the unfavorable
positioning of the carboxylic acid group para to the β-carboline
core in contrast to 2A-4A, all of which have a carboxylic group
meta to the β-carboline core. Compared with ethyl ester
Scheme 1. Synthesis of Acid and Ethyl Ester Compounds
a
Experimental details are included in the Supporting Information: a.
Pictet-Spengler reaction with corresponding ketone; b. Separation of
diastereomers; c. Hydrolysis of the ester group to a carboxylic acid
group.
ester was then hydrolyzed to give the corresponding carboxylic
acid compound (2A-5A and 2B-5B). The configuration at the
C-3 position of β-carboline was set as R from the starting
material 10, which was derived from D-tryptophan. The
configuration at the C-1 position relative to that of the C-3
positon was established by NOE experiments. The compounds
in the A series have the imidazole ring and the N-methyl
pyrazole ring on the same face of the β-carboline core. The
compounds in the B series have the opposite configuration at
C-1 position of the β-carboline core.
a,b
Table 3. Profiles of Ethyl Ester Compounds
compd human sstr3 binding
human sstr3 functional cAMP
antagonism IC50, nM (% inh.)
mouse sstr3 binding
IC50, nM
mouse sstr3 functional cAMP
antagonism IC50, nM (% inh.)
MK-499 binding
assay, IC50, nM
no.
IC50, nM
6A
6B
7A
7B
8A
8B
9A
9B
0.83 0.19 (n = 6)
1.52 0.74 (n = 6)
0.64 0.30 (n = 4)
4.87 0.36 (n = 2)
0.78 0.00 (n = 2)
3.36 0.18 (n = 4)
1.72 0.19 (n = 2)
2.91 0.40 (n = 2)
1.97 0.94 (n = 3) (81%)
5.36 3.05 (n = 4) (81%)
0.78 0.11 (n = 3) (101%)
47.1 11.1(n = 2) (98%)
0.15 0.01 (n = 2) (85%)
24.0 16.0 (n = 4) (134%)
5.25 2.57 (n = 4) (91%)
31.6 18.1 (n = 4) (65%)
0.41 0.00 (n = 2)
0.70 0.06 (n = 2)
0.41 0.08 (n = 4)
0.58 0.23 (n = 2) (80%)
3.48 0.57 (n = 2) (87%)
0.36 0.25(n = 4) (91%)
14.7 (n = 1) (85%)
707 27 (n = 2)
625 (n = 1)
1420 (n = 1)
1205 (n = 1)
2257 (n = 1)
2520 843 (n = 2)
1611 (n = 1)
2845 (n = 1)
c
ND
c
ND
0.35 0.09 (n = 2) (99%)
8.37 3.74 (n = 4) (107%)
1.53 0.22 (n = 3) (96%)
14.2 2.7 (n = 3) (66%)
c
ND
c
ND
c
ND
a
b
c
Assay protocols are provided in the Supporting Information. The standard deviations are included as well as the numbers of repeats. Not
determined.
C
dx.doi.org/10.1021/ml500028c | ACS Med. Chem. Lett. XXXX, XXX, XXX−XXX

ACS Medicinal Chemistry Letters
Letter
compound 3A has little interaction with the hERG channel. It
has a weak binding to the hERG channel according to MK-499
binding assay. In the functional PatchXpress assay, it has
minimal inhibition (26%) of hERG at the maximal dose tested
(30 μM).
Table 4. sstr3 in Vitro Activity of Carboxylic Acid
Compounds
a,b
human sstr3
functional cAMP
mouse sstr3
functional cAMP
human sstr3
mouse sstr3
compd binding IC50, antagonism IC50, binding IC50, antagonism IC50,
no.
nM
nM (% inh.)
nM
nM (% inh.)
Due to its favorable overall profile, compound 3A was
evaluated in mouse oGTT assay to help confirm target
engagement with sstr3. When acid 3A was dosed orally at 3
and 10 mg/kg in mice, the reduction of glucose excursion was
minimal (24% and 35%) compared to MK-4256, the positive
control, which reduced glucose excursion by 63%. The poor
efficacy was likely due to the poor oral bioavailability for 3A in
mice since the plasma drug levels of 3A at 2.5 h post dosing
were only 2 nM and 3 nM for 3 mg/kg and 10 mg/kg,
respectively.24 To circumvent the poor oral bioavailability, 3A
was dosed intraperitoneally (i.p.) to improve its drug exposure.
When compound 3A was dosed i.p. at 1 mg/kg, the reduction
of glucose excursion was 64%, comparable to the positive
control MK-4256, which reduced the glucose excrusion by 74%.
The plasma drug level of 3A at 2.5 h post i.p. dosing was 74
nM, which is much higher than the drug exposures achieved
after oral dosings. Given the good oGTT efficacy established
together with excellent in vitro potencies in binding and
functional assays, we were confident that with sufficient plasma
concentrations,3A is capable of engaging the sstr3 receptor.
After confirming the in vivo target engagement of 3A on
sstr3, we evaluated 3A in the anesthetized, vagus-intact CV dog
model (Figure 3). Compound 3A was intravenously infused
2A
4.41 1.04
(n = 8)
67.0 69.0
(n = 8) (129%)
3.93 0.75
(n = 2)
34.0 11.2
(n = 11)
(101%)
c
2B
3A
3B
4A
4B
5A
5B
74.7 13.1
(n = 4)
105 106 (n = 3)
(83%)
ND
111 25 (n = 3)
(69%)
2.13 0.37
(n = 8)
1.71 0.73
(n = 4) (101%)
1.72 0.34
0.73 0.34
(n = 3) (81%)
(n = 4)
c
72.7 1.5
(n = 2)
52.3 9.5 (n = 2)
(98%)
ND
17.1 (n = 1)
(70%)
c
1.22 0.17
(n = 4)
0.52 0.28
(n = 4) (88%)
ND
0.16 0.03
(n = 3) (67%)
c
10.8 1.1
(n = 2)
6.12 2.17
(n = 2) (117%)
ND
2.90 1.71
(n = 2) (80%)
c
23.7 6.8
(n = 2)
87.2 16.5
(n = 2) (88%)
ND
57.1 34.2
(n = 4) (87%)
c
409 21
(n = 2)
>2000 (n = 2)
(27%)
ND
1863 58 (n = 2)
(51%)
a
b
Assay protocols are provided in the Supporting Information. The
standard deviations are included as well as the numbers of repeats.
Not determined.
c
analogues as well as MK-4256, carboxylic acid compounds 2A,
3A and 4A all have a much reduced binding to hERG as
observed in the MK-499 binding assay (Table 5).
Table 5. Activities of Selected Acid Compounds on the
a,b
hERG Channel
compd
no.
MK-499 binding assay, IC50,
nM
PatchXpress, hERG
2A
3A
4A
8703 504 (n = 2)
19560 2913 (n = 2)
11990 3877 (n = 2)
39% inh. at 10 μM
26% inh. at 30 μM
70% inh. at 30 μM
(IC50 = 12 μM)
a
The standard deviations are included as well as the numbers of
b
repeats. PatchXpress assay (n = 1) is sufficient to differentiate 3A
from 2A and 4A.
Compound 2A, the first carboxylic acid compound prepared
in this series, had good potency in human sstr3 binding assay
and some potency erosion in the human sstr3 functional assay.
With MK-499 binding Ki of 8526 nM and 39% blockade of
hERG at 10 μM (the highest concentration tested) in the
PatchXpress assay, compound 2A clearly demonstrated an
improvement over MK-4256 (Table 5). Compound 2A was
evaluated for PK in mice. It demonstrated moderate clearance
(15.6 mL min−1kg−1) and a reasonable half-life (2.62 h).
However, it had very low oral bioavailability (1.5%).
Compound 4A exhibited the best potency on sstr3 among all
of the acid compounds prepared according to in vitro sstr3
binding and functional assays (Table 4). However, it showed
∼12 μM binding in the MK-499 binding assay (Table 5). The
functional blockade of the hERG channel was confirmed in the
PatchXpress assay (IC50 12 μM). Compound 4A represented
some improvement over MK-4256 in term of the hERG profile.
Compound 3A had the best overall profile. It is potent in
binding and functional assays for human as well as mouse sstr3.
Furthermore, 3A is highly selective for sstr3 against other sstr
subtypes with human binding IC50 higher than 10 μM on sstr1,
sstr2, sstr4, and sstr5, respectively (>4000x). More importantly,
Figure 3. Effect of 3A on QTc in an anesthetized, vagus-intact CV dog
model.
continuously in dogs at increasing doses of 10, 20, and 50 mg/
kg over three 30 min intervals. High plasma concentrations of
3A were detected (38 uM, 199 uM and 909 uM, respectively).
At Cmax =38 uM, compound 3A did not cause QTc
prolongation. At Cmax of 199 uM, the QTc prolongation did
not exceed 5% cutoff. At plasma concentration as high as 909
uM, QTc prolongation barely exceeded the 5% cutoff. In
contrast to the results for MK-4256, which showed dose-
dependent increases in QTc, we concluded compound 3A
caused little or no QTc prolongation at high plasma drug levels
in this CV dog model.
Since MK-4256 is highly plasma protein bound (>99% in
dog and human plasma), a significant serum shift on the hERG
channel might be expected. However, QTc prolongation was
observed with plasma concentrations in the CV dog model near
D
dx.doi.org/10.1021/ml500028c | ACS Med. Chem. Lett. XXXX, XXX, XXX−XXX

ACS Medicinal Chemistry Letters
Letter
the IC50 obtained from in vitro assays on hERG channel. There
was no observed serum shift. A thallium flux assay was used to
assess the effect of added serum to the response of the hERG
channel.25 Table 6 shows the effect of added dog serum to the
ACKNOWLEDGMENTS
■
We thank Dr. Li-Kang Zhang for measuring the high resolution
mass spectra of 3A, Dr. Mary Senior, Dr. Alexei Buevich, Dr.
Mengfen (Melissa) Lin, and Mr. Mark A. Holmes for NMR
support, and Dr. Gregory J. Kaczorowski and his team (Kevin
S. Ratliff, Birgit T. Priest, and James Herrington) for evaluating
MK-4256 in the thallium flux assay. All of them are current or
former employees at Merck Research Laboratories.
Table 6. Effect of Added Dog Serum to the Blockade of
hERG Function by MK-4256
% blockade of hERG
conc of MK-4256 (μM)
0% added dog serum
37 10
50% added dog serum
REFERENCES
■
3
32
62
8
8
(1) For an excellent review on somatostatin and somatostatin
receptors, see: Kumar, U.; Grant, M. Somatostatin and Somatostatin
Receptors. Results and Probl. Cell Differ 2010, 50, 137−184.
(2) Poitout, L.; Roubert, P.; Contour-Galcera, M.; Moinet, C.;
Lannoy, J.; Pommier, J.; Plas, P.; Bigg, D.; Thurieau, C. Identification
of Potent Non-Peptide Somatostatin Antagonists with sst3 Selectivity.
J. Med. Chem. 2001, 44, 2990.
(3) Thurieau, C.; Poitout, L.; Galcera, M.; Moinet, C.; Gordon, T.
D.; Morgan, B. A.; Bigg, D.; Pommier, J. β-Carboline Compounds.
WO 1999/64420, December 16, 1999.
10
61
9
ability of MK-4256 to block the response of the hERG channel.
Despite being highly protein bound (dog Fu = 0.8%), MK-4256
exhibited no serum shift. This is consistent with the observed
QTc increase in dogs which occurred at a plasma concentration
(4.6 uM) very near the IC50 in the MK-499 inhibition assay
(1.7 uM) and the patch clamp assay (3.4 uM). Despite the lack
of an expected serum shift, these results would still seem to
intimate a role for the hERG channel in the observed QTc
prolongation. The underlying cause(s) for a lack of a serum
shift on the hERG channel remain unresolved.
(4) Troxler, T.; Hurth, K.; Hoyer, D. Beta-Carboline Derivatives and
Its Pharmaceutical Use Against Depression and Anxiety. WO 2002/
081471, April 2, 2002.
(5) Troxler, T.; Hurth, K.; Schuh, K.; Schoeffter, P.; Langenegger, D.;
Enz, A.; Hoyer, D. Decahydroisoquinoline derivatives as novel non-
peptidic, potent and subtype-selective somatostatin sst3 receptor
antagonists. Bioorg. Med. Chem. Lett. 2010, 20, 1728−1734.
(6) Baenziger, M.; Cercus, J.; Hirt, H.; Laumen, K.; Malan, C.;
Spindler, F.; Struber, F.; Troxler, T. The development of a practical
synthesis of the potent and selective somatostatin sst3 receptor
antagonist [4-(3,4-difluoro-phenyl)-piperazine-1-yl]-{(4S,4aS,8aR)-
2[(S)-3-(6-methoxy-pyridin-3-yl)-2-methyl-propyl]-decahydroisoqui-
noline-4-yl}-methanone (NVP-ACQ090). Tetrahedron: Asymmetry
2003, 14, 3469−3477.
(7) Pasternak, A.; Feng, Z.; deJesus, R.; Ye, Z.; He, S.; Dobbelaar, P.
H.; Bradley, S. A.; Chicchi, G. G.; Tsao, K.; Trusca, D.; Eiermann, G.
J.; Feng, Y.; Wu, M.; Shao, Q.; Zhang, B.; Nargund, R. P.; Mills, S. G.;
Howard, A. D.; Yang, L.; Zhou, Y. Stimulation of glucose-dependent
insulin secretion by a potent, selective sstr3 antagonist. ACS Med.
Chem. Lett. 2012, 3, 289−293.
(8) He, S.; Ye, Z.; Truong, Q.; Shah, S.; Du, W.; Guo, L.; Dobbelaar,
P. H.; Lai, Z.; Liu, J.; Jian, T.; Qi, H.; Bakshi, R. K.; Hong, Q.;
Dellureficio, J.; Pasternak, A.; Feng, Z.; deJesus, R.; Yang, L.; Reibarkh,
M.; Bradley, S. A.; Holmes, M. A.; Ball, R. G.; Ruck, R. T.; Huffman,
M. A.; Wong, F.; Samuel, K.; Reddy, V. B.; Mitelman, S.; Tong, S. X.;
Chicchi, G. G.; Tsao, K.-L.; Trusca, D.; Wu, M.; Shao, Q.; Trujillo, M.
E.; Eiermann, G. J.; Li, C.; Zhang, B. B.; Howard, A. D.; Zhou, Y.-P.;
Nargund, R. P.; Hagmann, W. K. The Discovery of MK-4256, a Potent
SSTR3 Antagonist as a Potential Treatment of Type 2 Diabetes. ACS
Med. Chem. Lett. 2012, 3, 484−489.
(9) For a reference describing the measurment of plasma protein
binding by equilibrium dialysis, see: Waters, N. J.; Jones, R.; Williams,
G.; Sohal, B. Validation of a rapid equilibrium dialysis approach for the
measurement of plasma protein binding. J. Pharm. Sci. 2008, 97,
4586−4595.
(10) Rampe, D.; Brown, A. M. A history of the role of the hERG
channel in cardiac risk assessment. J. Pharmacol. Toxicol. Methods
2013, 68, 13−22.
(11) Jamieson, C.; Moir, E. M.; Rankovic, Z.; Wishart, G. Medicinal
chemistry of hERG optimizations: Highlights and hang-ups. J. Med.
Chem. 2006, 49, 5029−5046.
In summary, we have investigated the cause of an adverse CV
effect observed for MK-4256 in the CV dog model. The
observed QTc prolongation with MK-4256 was dependent
upon the “vagal” status of the animals: no QTc interval increase
in vagotemized dogs but an observed increase in vagus-intact
animals. A carboxylic acid moiety was introduced into the
structure to mitigate the off-target interaction with the hERG
channel. From this effort, we identified a tool compound, 3A,
which has excellent potency in sstr3 binding and functional
assays. Compound 3A demonstrated effective target engage-
ment by suppressing glucose excursion in oGTT in mice. More
significantly, compound 3A demonstrated little to no effect in
anesthetized, vagus-intact CV dog models. In contrast, MK-
4256 caused a dose-dependent increase in QTc in the same
model. This investigation demonstrated the QTc prolongation
of MK-4256 was due to the blockade of the hERG channel and
was not due to antagonism of sstr3. During this investigation,
we also found that, for this class of compounds, the
anesthetized, vagus-intact CV dog models appeared to be
more appropriate than the anesthetized, vagotemized CV dog
model for the evaluation of QTc prolongation. Compound 3A
could not be progressed due to its suboptimal PK profile in
preclinical species. However, this tool compound cleared a path
forward for the discovery of sstr3 antagonist devoid of the CV
liability of MK-4256. Additional efforts in discovering non-
carboxylic acid compounds with minimal hERG activity will be
disclosed separately.26
ASSOCIATED CONTENT
* Supporting Information
Syntheses and characterization data for compounds 2A −9A
and 2B −9B and biological assay protocols. This material is
available free of charge via the Internet at http://pubs.acs.org.
■
S
(12) Radiolabeled MK-499 has been used in a hERG (potassium ion
channel Kv11.1) binding assay to provide an initial screen of a
compound’s potential to interact with hERG. Raab, C. E.; Butcher, J.
W.; Connolly, T. M.; Karczewski, J.; Yu, N. X.; Staskiewicz, S. J.;
Liverton, N.; Dean, D. C.; Melillo, D. G. Synthesis of the first sulfur-
35-labeled hERG radioligand. Bioorg. Med. Chem. Lett. 2006, 16,
1692−1695.
AUTHOR INFORMATION
Corresponding Author
*Phone: 908-740-0881. E-mail: shuwen_he@merck.com.
Notes
■
The authors declare no competing financial interest.
E
dx.doi.org/10.1021/ml500028c | ACS Med. Chem. Lett. XXXX, XXX, XXX−XXX

ACS Medicinal Chemistry Letters
Letter
(13) For a description of the protocol of MK-499 binding assay, see:
Wang, J.; Della Penna, K.; Wang, H.; Karczewski, J.; Connolly, T. M.;
Koblan, K. S.; Bennett, P. B.; Salata, J. J. Functional and
pharmacological properties of canine ERG potassium channels. Am.
J. Physiol. Heart Circ. Physiol. 2003, 284, H256−H267.
(14) For a reference of an automated patch-clamp hERG screening
methold (PatchXpress assay), see: Zeng, H.; Penniman, J. R.; Kinose,
F.; Kim, D.; Trepakova, E. S.; Malik, M. G.; Dech, S. J.;
Balasubramanian, B.; Salata, J. J. Improved Throughput of PatchXpress
hERG Assay Using Intracellular Potassium Fluoride. Assay Drug
Development Technol. 2008, 6, 235−241.
(15) For a description of the anesthetized CV dog model, see:
Tashibu, H.; Miyazaki, H.; Aoki, K.; Akie, Y.; Yamamoto, K. QT
PRODACT: in vivo QT assay in anesthetized dog for detecting the
potential for QT interval prolongation by human pharmaceuticals. J.
Pharmacol. Sci. (Tokyo, Japan) 2005, 99, 473−486.
(26) Shah S. K.; He, S.; Guo, L.; Truong Q.; Qi H.; Wu, D.; Lai, Z.;
Liu, J.; Jian, T.; Hong, Q.; Samuel, K.; Madiera, M.; Reddy, V. B.;
Mitelman, S.; Tong, S. X.; Chicchi, G. G.; Tsao, K.-L.; Trusca, D.; Wu,
M.; Shao, Q.; Trujillo, M. E.; Eiermann, G. J.; Li, C.; Pachanski, M.;
Fernandez, G.; Nelson, D.; Bunting, P.; Morisette, P.; Volksdorf, S.;
Kerr, K.; Zhang, B.; Howard, A. D.; Zhou, Y.-P.; Nargund, R. P.;
Hagmann, W. K. The Discovery of MK-1421 as a Potential Treatment
for Type 2 Diabetes: sstr3 Antagonists with Reduced QTc
Prolongation. Manucript in preparation.
NOTE ADDED AFTER ASAP PUBLICATION
■
This paper was published ASAP on April 28, 2014. Due to a
production error, Table 2 was corrected and the revised version
was reposted on May 9, 2014.
(16) For a reference on the conscious CV dog telemetry model, see:
Chaves, A. A.; Zingaro, G. J.; Yordy, M. A.; Bustard, K. A.; O’Sullivan,
S.; Galijatovic-Idrizbegovic, A.; Schuck, H.; Christian, D. B.; Hoe, C.
M.; Briscoe, R. J. A highly sensitive canine telemetry model for
detection of QT interval prolongation: Studies with moxifloxacin,
haloperidol and MK-499. J. Pharmacol. Toxicol. Methods 2007, 56,
103−114.
(17) For a reference on the conscious CV dog telemetry model, see:
Toyoshima, S.; Kanno, A.; Kitayama, T.; Sekiya, K.; Nakai, K.; Haruna,
M.; Mino, T.; Miyazaki, H.; Yano, K.; Yamamoto, K. QT PRODACT:
in vivo QT assay in the conscious dog for assessing the potential for
QT interval prolongation by human pharmaceuticals. J. Pharmacol. Sci.
(Tokyo, Japan) 2005, 99, 459−471.
(18) Woltmann, R.; Morissette, P. Effect of vagotomy on the
cardiovascular function in pentobarbital anesthetized dogs. J.
Pharmacol. Toxicol. Methods 2009, 60, 210−258, DOI: 10.1016/
j.vascn.2009.04.026.
(19) Fatti, L. M.; Scacchi, M.; Lavezzi, E.; Giraldi, F. P.; De Martin,
M.; Toja, P.; Michailidis, G.; Stramba-Badiale, M.; Cavagnini, F. Effects
of treatment with somatostatin analogues on QT interval duration in
acromegalic patients. Clinical Endocrinol. 2006, 65, 626−630.
(20) Zhu, B.-Y.; Jia, Z. J.; Zhang, P.; Su, T.; Huang, W.; Goldman, E.;
Tumas, D.; Kadambi, V.; Eddy, P.; Sinha, U.; Scarborough, R. M.;
Song, Y. Inhibitory effect of carboxylic acid group on hERG binding.
Bioorg. Med. Chem. Lett. 2006, 16, 5507−5512.
(21) Aslanian, R.; Piwinski, J. J.; Zhu, X.; Priestley, T.; Sorota, S.; Du,
X.-Y.; Zhang, X.-S.; McLeod, R. L.; West, R. E.; Williams, S. M.; Hey, J.
A. Structural determinants for histamine H1 affinity, hERG affinity and
QTc prolongation in a series of terfenadine analogs. Bioorg. Med.
Chem. Lett. 2009, 19, 5043−5047.
(22) Meyers, M. J.; Arhancet, G. B.; Hockerman, S. L.; Chen, X.;
Long, S. A.; Mahoney, M. W.; Rico, J. R.; Garland, D. J.; Blinn, J.. R.;
Collins, J. T.; Yang, S.; Huang, H.-C.; McGee, K. F.; Wendling, J. M.;
Dietz, J. D.; Payne, M. A.; Homer, B. L.; Heron, M. I.; Reitz, D. B.; Hu,
X. Discovery of (3S,3aR)-2-(3-Chloro-4-cyanophenyl)-3-cyclopentyl-
3,3a,4,5-tetrahydro 2H-benzo[g]indazole-7-carboxylic Acid (PF-
3882845), an Orally Efficacious Mineralocorticoid Receptor (MR)
Antagonist for Hypertension and Nephropathy. J. Med. Chem. 2010,
53, 5979−6002.
(23) Mihalic, J. T.; Fan, P.; Chen, X.; Chen, X.; Fu, Y.; Motani, A.;
Liang, L.; Lindstrom, M.; Tang, L.; Chen, J.-L.; Jaen, J.; Dai, K.; Li, L.
Discovery of a novel melanin concentrating hormone receptor 1
(MCHR1) antagonist with reduced hERG inhibition. Bioorg. Med.
Chem. Lett. 2012, 22, 3781−3785.
(24) The PK profile of compound 3A in rats was also not optimal
with a high clearance (119 mL min−1 kg−1).
(25) For a description of thallium flux assay, see: Schmalhofer, W. A.;
Swensen, A. M.; Thomas, B. S.; Felix, J. P.; Haedo, R. J.; Solly, K.; Kiss,
L.; Kaczorowski, G. J.; Garcia, M. L. A Pharmacologically Validated,
High-Capacity, Functional Thallium Flux Assay for the Human Ether-
a-
̀
go-go Related Gene Potassium Channel. Assay Drug Development
Technol. 2010, 8, 714−726.
F
dx.doi.org/10.1021/ml500028c | ACS Med. Chem. Lett. XXXX, XXX, XXX−XXX