Direct observation (NMR) of the efficacy of glucagon receptor antagonists in murine liver expressing the human glucagon receptor

Article abstract of DOI:10.1016/j.bmc.2005.10.008

The demonstration of pharmacodynamic efficacy of novel chemical entities represents a formidable challenge in the early exploration of synthetic lead classes. Here, we demonstrate a technique to validate the biological efficacy of novel antagonists of the

Full text of DOI:10.1016/j.bmc.2005.10.008

Bioorganic & Medicinal Chemistry 14 (2006) 1506–1517
Direct observation (NMR) of the efficacy of glucagon receptor
antagonists in murine liver expressing the human glucagon receptor
Sheila M. Cohen,a Joseph L. Duffy,b,* Corin Miller,a Brian A. Kirk,b
Mari Rios Candelore,c Victor D. H. Ding,c Gregory Kaczorowski,d Laurie M. Tota,c
Jeffrey G. Werrmann,a Michael Wright,c Emma R. Parmee,b
James R. Tatab and Bei B. Zhangc
aDepartment of Research Imaging, Merck Research Laboratories, Rahway, NJ 07065, USA
bDepartment of Medicinal Chemistry, Merck Research Laboratories, Rahway, NJ 07065, USA
cDepartment of Molecular Endocrinology, Merck Research Laboratories, Rahway, NJ 07065, USA
dDepartment of Biochemistry and Physiology, Merck Research Laboratories, Rahway, NJ 07065, USA
Received 5 July 2005; revised 3 October 2005; accepted 3 October 2005
Available online 25 October 2005
Abstract—The demonstration of pharmacodynamic efficacy of novel chemical entities represents a formidable challenge in the early
exploration of synthetic lead classes. Here, we demonstrate a technique to validate the biological efficacy of novel antagonists of the
human glucagon receptor (hGCGR) in the surgically removed perfused liver prior to the optimization of the pharmacokinetic prop-
erties of the compounds. The technique involves the direct observation by 13C NMR of the biosynthesis of [13C]glycogen from
[13C]pyruvate via the gluconeogenic pathway. The rapid breakdown of [13C]glycogen (glycogenolysis) following the addition of
50 pM exogenous glucagon is then monitored in real time in the perfused liver by 13C NMR. The concentration-dependent inhibi-
tion of glucagon-mediated glycogenolysis is demonstrated for both the peptidyl glucagon receptor antagonist 1 and structurally
diverse synthetic antagonists 2–7. Perfused livers were obtained from a transgenic mouse strain that exclusively expresses the func-
tional human glucagon receptor, conferring human relevance to the activity observed with glucagon receptor antagonists. This tech-
nique does not provide adequate quantitative precision for the comparative ranking of active compounds, but does afford
physiological evidence of efficacy in the early development of a chemical series of antagonists.
Ó 2005 Elsevier Ltd. All rights reserved.
1. Introduction
The peptide activates the hepatic glucagon receptor
(GCGR) and triggers the synthesis of glucose (gluconeo-
genesis) and processing and release of hepatic glycogen
stores (glycogenolysis) to restore plasma glucose and
maintain homeostasis. In diabetic individuals, inappro-
priately high hepatic glucose production arising from
elevated glucagon exacerbates inappropriately low glu-
cose utilization. The combination of these two deficien-
cies results in chronic hyperglycemia. Many therapeutic
approaches to the treatment of diabetes have focused on
the normalization and utilization of plasma insulin.1c,2
An alternative or additional therapy may be realized
by blocking hepatic glucose production using glucagon
receptor antagonists.3 Peptidyl antagonists are known,4
and several small molecule synthetic antagonists have
been reported as well.5
Diabetes mellitus is a condition characterized by chron-
ically elevated levels of blood glucose. This condition
arises from inappropriate secretion and activity of the
two major pancreatic hormones that control glucose
homeostasis, insulin and glucagon. In the diabetic state,
insulin resistance and insulin deficiency impair the
bodyꢀs ability to absorb and utilize glucose from the
bloodstream. The condition is commonly aggravated
by chronically elevated levels of the peptide glucagon
or hyperglucagonemia.1 Glucagon is a 29 amino acid
peptide that is released from pancreatic a cells. In
healthy individuals, the hormone is synthesized and
secreted in response to insufficient plasma glucose levels.
Keywords: Glucagon; Antagonist; Glycogenolysis; Perfused liver;
Diabetes; NMR; Transgenic.
The activity of peptide or small molecule glucagon
receptor antagonists may be readily measured at the cel-
lular level with in vitro receptor binding and functional
*
Corresponding author. Tel.: +1 732 594 1062; fax: +1 732 594
5350; e-mail: joseph_duffy@merck.com
0968-0896/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2005.10.008

S. M. Cohen et al. / Bioorg. Med. Chem. 14 (2006) 1506–1517
1507
assays. The antagonists may be further investigated in vi-
vo using a variety of animal models of diabetes mellitus.
However, the demonstration of pharmacodynamic effi-
cacy requires that the compound be optimized beyond
intrinsic potency to overcome several additional obsta-
cles to an interpretable experiment. For example, the
investigational compound must be sufficiently potent
on the glucagon receptor that is endogenous to the ani-
mal model. In the event that the antagonist is highly po-
tent against the human receptor, but lacks potency
against other mammalian receptor variants, the proba-
bility of demonstrating a pharmacological effect in other
species is low. Indeed, in early reports from these labo-
ratories we observed a variable but substantial decrease
in the binding affinity of glucagon receptor antagonists
in the murine receptor.9b Additionally, the pharmacoki-
netic parameters of the compound must be suitable for
the route of administration, such that the compound
can reach sufficient concentration in hepatic tissue to ex-
ert a measurable physiological response. The optimiza-
tion of these parameters may add substantial
additional barriers to the early demonstration of efficacy
of a given lead class of glucagon receptor antagonists.
was compared with the intrinsic potency observed in
the in vitro functional (cAMP) assay to examine the
pharmacodynamic relevance of gains in intrinsic
potency.
2. Results
2.1. Chemistry
The glucagon receptor antagonists investigated in the
present study are illustrated in Table 1. The peptidyl
antagonist 1,4a the triarylpyrrole antagonist 2,9 and the
4-phenylpyridine antagonists 6 and 7 have been reported
previously.10 Recently, we reported the discovery of the
thiophene-derived glucagon receptor antagonists 3 and
4, and an inactive variant of these compounds 5, in brief
communications.11 The synthesis of the cyclohexyl
substituted thiophenes 3 and 5 is illustrated in Scheme
1. The cyclohexanone derivative I was combined with
malononitrile and elemental sulfur under basic condi-
tions to afford the aminothiophene II directly by the Ge-
wald cyclization.12 Acylation with the appropriate acid
chloride under basic conditions afforded 3 or 5. In a sim-
ilar fashion, the isoxadiazole substituted aminothioph-
ene antagonist 4 was prepared, as illustrated in
Scheme 2. A Gewald cyclization on the the b-ketoester
III with malononitrile afforded the aminothiophene
IV. Acylation of this material with the appropriate acid
chloride afforded the thiophene amide V. This interme-
diate was hydrolyzed to the corresponding carboxylic
acid, followed by condensation with the hydroxamidine
amide VI under standard coupling conditions to afford
intermediate VII in good yield. Dehydration of this
intermediate was accomplished by heating in diglyme,
affording the isoxadiazole derivative 4.
In previous reports from these laboratories, we addressed
the species-dependent receptor variation with the devel-
opment and characterization of a strain of transgenic
mice that express the human glucagon receptor.6 The
homozygous receptor replacement mice (hGCGR mice)
were healthy and fertile, and employ a human hepatic glu-
cagon receptor that was demonstrated to be functionally
intact both in vitro and in vivo. The utilization of hGCGR
mice allows for a meaningful correlation of murine hepat-
ic gluconeogenic activity with in vitro potency of antago-
nists optimized against the human glucagon receptor.
The challenge of attaining sufficient in vivo compound
exposure for validation of efficacy may be overcome
by examining the activity of compounds in the entire
functioning organ ex vivo. The mouse liver is an attrac-
tive candidate organ for ex vivo study by NMR spec-
troscopy owing to its surgical accessibility and small
size.7 With the infusion of 13C-labeled pyruvate into a
surgically removed and perfused murine liver, we have
monitored the biosynthesis of glycogen in real time by
perfused-liver 13C NMR and observed the rapid break-
down of glycogen in response to exogenously adminis-
tered glucagon. An NMR spectrometer equipped with
a standard vertical-bore NMR magnet was used for
these studies.8
2.2. In vitro potency of glucagon receptor antagonists
The compounds were assayed by measurement of the
inhibition of binding of 125I-glucagon to the hGCGR ex-
pressed in CHO cell membrane, affording the binding
IC50 values in Table 1. The compounds were further
examined for the functional antagonist potency, as mea-
sured by the inhibition of glucagon-induced cAMP
accumulation in hGCGR-transfected CHO cells (cAMP
IC50, Table 1). The peptide 1 is a potent glucagon recep-
tor antagonist, with a binding IC50 of 18 4 nM, and an
equal potency in the functional assay with a cAMP IC50
of 26 10 nM. The triaryl pyrrole antagonist 2 was
reported previously from these laboratories.9b Although
the compound retains similar inhibitory potency to 1 in
the membrane-based binding assay (IC50 = 14 5 nM),
the compound is substantially less active as a functional
antagonist of the glucagon receptor in the cell-based
cAMP assay (IC50 = 120 70 nM).13 The thiophene-de-
rived compounds 3, 4, and 5 have also been reported
previously from these laboratories, and their potency
in the binding and functional assays is presented in
Table 1 for comparison.11 The inactive thiophene
derivative 5 was included in these NMR investigations
as a structurally similar negative control. Finally, an
additional class of potent antagonists are represented
Herein we report that the time-resolved perfused-liver
NMR technique may be used for the direct observation
of the inhibition of glucagon-induced glycogenolysis by
glucagon receptor antagonists. We have employed per-
fused liver from the murine hGCGR replacement strain
to minimize species-dependent differences in the efficacy
of the antagonists and to confer human relevance to the
experiments. We examined a structurally diverse set of
small molecule glucagon receptor antagonists at multi-
ple concentrations for the ability to inhibit glucagon-in-
duced glycogenolysis in the perfused hGCGR liver. The
efficacy of these compounds in this whole organ assay

1508
S. M. Cohen et al. / Bioorg. Med. Chem. 14 (2006) 1506–1517
Table 1. Binding and functional activity of antagonists of the human glucagon receptor
Compound
Glucagon in vitro assays (nM)a
Binding IC50 cAMP IC50
1
2
des-His1[Glu9]glucagon-NH2
18
4
26 10
14
5
120 70
3b
181 10
129 70
4
89 17
34 10
5b
>10,000
>10,000
6
1.6 0.2
0.7 0.3
7
1270 720
147 50
a The IC50 values are given as means of three to four measurements SEM.
b Compounds were prepared and tested as racemic mixtures.
Scheme 1. Reagents and conditions: (a) malononitrile, S8, morpholine, EtOH, 70 °C. (b) 2-Ethylbutyryl chloride, DIEA, CH2Cl2.
(c) 4-Bromobenzoyl chloride, DIEA, CH2Cl2.

S. M. Cohen et al. / Bioorg. Med. Chem. 14 (2006) 1506–1517
1509
Scheme 2. Reagents and conditions: (a) malononitrile, S8, morpholine, EtOH, 70 °C. (b) 2-Ethylbutyryl chloride, DIEA, CH2Cl2. (c) TFA, CH2Cl2.
(d) EDC, HOBT, DIEA, DMF. (e) Diglyme, 135 °C.
by the recently reported phenylpyridine 6.10 The less ac-
tive enantiomeric phenylpyridine 7 was also investigated
in the liver NMR assay. The enantiomers 6 and 7 retain
identical physical properties, but differ by over three
orders of magnitude in their binding and functional
potency in the hGCGR in vitro assays.14 Thus, any dif-
ference in the efficacy of these compounds in the liver
NMR assay may be ascribed directly to the difference
in intrinsic potency of these compounds.
marized here for clarity. As illustrated in Figure 1, per-
fused livers from hGCGR transgenic mice were
surgically excised and positioned in a 20 mm NMR tube
with the portal vein cannula suspended through the center
of a modified Teflon plug. During the experiment, the
Krebs buffer perfusion fluid containing dialyzed BSA
was continuously oxygenated and recirculated through
the liver. An initial 25 min period of flow-through, sub-
strate-free perfusion washed out endogenous hormones
and substrates. For absolute quantification of 13C and
31P metabolites, a reference compound reservoir was
incorporated in the liver NMR perfusion cell; this reser-
voir contained methylenediphosphonic acid and
[13C]urea at effective amounts of 12 and 35 lmol,
respectively.
2.3. Perfused liver preparations and NMR measurements
The isolated perfused liver technique and the apparatus
developed for liver perfusion in the NMR magnet have
been described in detail previously,7a,7c and they are sum-
Figure 1. Essential components of perfusion of an isolated liver in the NMR magnet. Inflow of perfusion fluid is via the portal vein cannula. An
expanded view of NMR perfusion cell (20-mm diameter) is shown on the right. The sealed referenced substance reservoir shown is described in the
text.

1510
S. M. Cohen et al. / Bioorg. Med. Chem. 14 (2006) 1506–1517
>4.5 h observation period were used in this study. One
assessment of hepatic viability is stable and acceptable
ATP levels over the course of the experiment; this was
measured experimentally by 31P NMR. A representative
pair of 31P NMR spectra from before and after an infu-
sion experiment is illustrated in Figure 2. The bottom
spectrum (a) shows a typical baseline profile of phosp-
hometabolites in hGCGR mouse liver for which initial
ATP = 3.3 lmol/g liver wet weight (glww). After 4 h per-
fusion (c) and after the addition of 50 pmol of glucagon
under our NMR conditions, liver ATP levels remained
constant (3.7 lmol/glww). Direct treatment with any
one of the seven glucagon receptor antagonists studied
did not affect hepatic ATP levels. Across the several
treatments reported here, representative hepatic ATP
levels were 3.6 lmol/glww 0.05 (N = 20, mean SEM),
and the ratio of the ATP level measured after accumula-
tion of all 13C spectra to the ATP level measured at
baseline was 0.99 0.06 (mean SD).
2.3.2. Time-resolved 13C NMR observations of hGCGR
mouse liver. The biosynthesis of glycogen in the perfused
liver was monitored by infusing the liver with the isoto-
pically labeled (>99%) substrate [2-13C]pyruvate using
the apparatus illustrated in Figure 1. The 13C isotopic
label was incorporated into specific carbons of glu-
cose-6-phosphate and eventually glycogen by the glu-
coneogenic pathway, illustrated in Scheme 3.8a
Synthesis of glycogen and glucose at an acceptable, lin-
ear rate from substrates that enter the pathway at the le-
vel of pyruvate is a rigorous test of liver function. The
kinetics of net glycogen synthesis were monitored in real
time by measurement and integration of the discrete
peak resulting from 13C enrichment at C-1 of the gluco-
syl units in glycogen by 13C NMR. As can be seen in
Figure 3, glycogen C-1 (labeled peak A) gives rise to
13C NMR signals free of overlapping peaks. Thus,
hepatic metabolism was monitored under approxi-
mately metabolic and isotopic steady-state conditions
Figure 2. 31P NMR spectra of isolated perfused hGCGR mouse liver
at baseline (a) and 4 h later after accumulation of 18 13C spectra (c).
The difference spectrum is shown in (b). This liver was exposed to
50 pM glucagon during the 13C NMR phase.
2.3.1. Control studies of perfused liver NMR. 31P NMR
measurements. Only perfused liver preparations that
were metabolically stable in the NMR magnet over the
Scheme 3. Simplified gluconeogenic pathway monitored in perfused liver from exogenously administered [2-13C]pyruvate to observed [13C]glycogen
and [13C]glucose, shown in red. The key enzymes are shown in green.

S. M. Cohen et al. / Bioorg. Med. Chem. 14 (2006) 1506–1517
1511
The accumulation of [13C]glycogen over time was mea-
sured by integration of the C-1 signal (Peak A in
Fig. 3), and integration area versus time is plotted in
Figure 4. Both groups of livers were treated with
DMSO vehicle rather than a synthetic antagonist at t
ꢀ25 min. One control group was treated with vehicle
only, denoted as no glucagon controls. The rate of net
accumulation of newly synthesized [13C]glycogen in
perfused hGCGR transgenic mouse livers was linear
over 3 h in this group. In contrast, when the perfused
livers were treated with vehicle plus 50 pM bolus of glu-
cagon, a significant reduction of [13C]glycogen was ob-
served over the following 40 min. This control group is
denoted as DMSO + glucagon controls in Figure 4.
The rapid depletion of hepatic glycogen coupled with
the appearance of 13C-labeled glucose (Fig. 3) in this
control group is attributed to glucagon-mediated
glycogenolysis. The subsequent resynthesis of glycogen
continues to be inhibited throughout the remainder of
the experiment, from 40 to 80 min after glucagon
addition.
2.3.3. Direct 13C NMR observations of efficacy of
glucagon receptor antagonists in hGCGR mouse liver.
Addition of a glucagon receptor antagonist to the per-
fused liver prior to administration of glucagon was
expected to diminish the subsequent reduction of
[13C]glycogen. As a validation of this NMR approach,
perfused livers from our transgenic mice were treated
with varying concentrations of the glucagon receptor
antagonist peptide 1 25 min prior to administration of
glucagon. As illustrated in Figure 5, the presence of
1 lM of antagonist 1 almost completely blocked the
induction of glycogenolysis by glucagon (Fig. 5b) and
allowed the continued accumulation of newly synthe-
sized [13C]glycogen without appreciable buildup of
13C-labeled glucose (Fig. 5c). The accumulation of
[13C]glycogen over time was again monitored by integra-
tion of the C-1 signal. The concentration-dependent
inhibition of glucagon-induced glycogenolysis by 1, as
measured by these spectra, is illustrated in Figure 6.
Figure 3. 13C NMR spectra of a representative perfused liver from
hGCGR mouse; (a and b) at the indicated times after the addition of
[2-13C]pyruvate and (c) after the subsequent addition of 50 pM
glucagon. The spectral region 105–60 ppm contains signals of newly
synthesized 13C-labeled glycogen and glucose. The 13C natural
abundance background spectrum of the liver, measured under identical
conditions before addition of 13C-labeled substrate, was subtracted
from each of the spectra shown. Abbreviation: (A) C-1 of glucosyl units
of glycogen. (B and C) C-1 of b and a anomers of glucose. (D) C-2 to
C-5 of glucosyl units in glycogen. (E) Lactate C-2 and C-2 of the
glycerol backbone of triacylglycerols (TG). (F) C-1 and C-3 of TG
glycerol. (G) C-6 of glucose and glucosyl units of glycogen. (H)
Glutamate C-2. (I) Glutamine C-2. (J) Alanine C-2. (a–e) Correspond
to signals arising from C-2 to C-5 of b and a anomers of D-glucose.7
by infusion of [2-13C]pyruvate (6.7 mM) + NH4Cl
(2.7 mM). The biosynthesis of 13C-labeled glycogen
can be detected by 13C NMR within 24 min of the addi-
tion of labeled substrate (Fig. 3a). [13C]Glycogen is seen
to accumulate monotonically in each subsequent spec-
trum for 105 min (Fig. 3b) as monitored by the glycogen
C-1 signal (peak A). In control perfused livers main-
tained under these same steady-state conditions, newly
synthesized [13C]glycogen continues to accumulate for
at least 3 h. However, administration of 50 pM glucagon
to the perfused liver initiates an immediate cessation of
glycogen accumulation, followed by the breakdown of
the [13C]glycogen synthesized to 13C-labeled glucose
(Fig. 3c). By subtracting the initial 13C spectrum of the
13C natural abundance of each liver from each spectra
accumulated after the addition of 13C-labeled substrate,
only those signals that could be traced back directly to
[2-13C]pyruvate, such as [13C1]glycogen (Peak A), were
selectively quantitated by integration.
Figure 4. Quantitation of the accumulation of
[
13C]glycogen in
DMSO vehicle treated perfused liver from hGCGR mouse liver over
3 h (-h-). The effect of treatment with DMSO vehicle plus 50 pM
glucagon (-n-) is also illustrated. Data represents the mean values
and standard errors from the indicated number of hGCGR mouse
perfused livers assayed.

1512
S. M. Cohen et al. / Bioorg. Med. Chem. 14 (2006) 1506–1517
The 13C NMR integration plots from similar adminis-
tration of the synthetic glucagon receptor antagonists
2 and 3 to perfused hGCGR mouse livers are illustrated
in Figures 7a and b, respectively. Each compound miti-
gates the glycogenolysis due to glucagon in a concentra-
tion-dependent fashion, albeit at substantially higher
concentrations than are required for a similar mitigation
from the more potent peptide antagonist 1 (Fig. 6). The
triarylpyrrole antagonist 2 afforded a near-complete
impediment to glucagon-induced glycogenolysis at a
concentration of 25 lM, whereas the thiophene-derived
antagonist 3 only partially inhibited glycogenolysis at
concentrations as high as 50 lM.
The activity of the more potent thiophene-derived gluca-
gon receptor antagonist 4 is illustrated in Figure 8, along
with the inactive thiophene derivative 5. Again in this in-
stance, the antagonist 4 effectively blocks glucagon-in-
duced glycogenolysis in the transgenic hGCGR mouse
livers in a concentration-dependent fashion. The struc-
turally related thiophene 5 was inactive even at 50 lM,
commensurate with the lack of activity observed with
this compound in the in vitro glucagon assays (Table 1).
Finally, the activities of the potent phenylpyridine
antagonist 6 and the less active enantiomer 7 are dis-
played in Figure 9. The glycogenolysis inhibition assay
for 6 was measured twice at each concentration, and
Figure 5. 13C NMR spectra illustrating the action of the peptide
glucagon receptor antagonist 1 in a representative perfused liver from
and hGCGR mouse. Peptide 1 was administered at the indicated time
prior to treatment with 50 pM glucagon under the conditions described
in Section 5 of the text. NMR conditions and abbreviations are as
denoted in Figure 3.
Figure 6. Quantitation of the inhibition of glucagon induced glyco-
genolysis by the peptide glucagon receptor antagonist 1 in perfused
hGCGR mouse liver. Time of addition of glucagon, 50 pM, is
designated as t = 0 min.
The control plots from Figure 4 are also presented for
comparison. The peptide partially inhibits glucagon-
mediated glycogenolysis at concentrations as low as
0.1 lM. At peptide concentrations above 1.1 lM glyco-
genolysis due to added glucagon is entirely eliminated.
Figure 7. Concentration-dependent inhibition of glucagon induced
glycogenolysis by synthetic glucagon antagonists 2 (a) and 3 (b) in
perfused hGCGR mouse liver.

S. M. Cohen et al. / Bioorg. Med. Chem. 14 (2006) 1506–1517
1513
class may hinder the interpretation of in vivo results.
Thus, additional tools are needed that may demonstrate
a pharmacological effect (or lack of effect) of explorato-
ry compounds as early as possible in a medicinal chem-
istry lead optimization program.
In order to provide pharmacological validation for lead
classes in our glucagon receptor antagonist program, we
sought to remove as many barriers as possible to the
observation of a direct pharmacological result of recep-
tor antagonism. Species differences in receptor interac-
tions with molecular antagonists were minimized by
employing transgenic mice that express a functional hu-
man glucagon receptor. Furthermore, the variability of
in vivo molecular exposure to the target of interest
was minimized by examining the effect of the antago-
nists in the functioning organ ex vivo. The livers were
exposed to 50 pM glucagon, which is approximately
the physiological level of this hormone in vivo in wild-
type mice.16 Finally, the effect of the antagonist on the
specific biochemical pathway of interest was examined
by spectroscopically isolating this pathway for observa-
tion using 13C-labeled substrate. In this way, the effect of
novel classes of glucagon receptor antagonists could be
demonstrated prior to optimization of other pharmaco-
logical properties.
Figure 8. Concentration-dependent inhibition of glucagon induced
glycogenolysis by synthetic glucagon antagonist 4, and corresponding
lack of inhibition by the inactive analogue 5, in perfused hGCGR
mouse liver.
The compounds studied in this investigation were cho-
sen because they represent a variety of structural classes
and a range of intrinsic potencies as glucagon receptor
antagonists. Four separate structural classes of antago-
nists were examined in the perfused-liver NMR assay.
The peptidyl antagonist 1, the triarylpyrrole antagonist
2, the aminothiophene antagonists 3 and 4, and the phe-
nylpyridine antagonist 6 all mitigated glucagon-induced
glycogenolysis in a concentration-dependent fashion.
The aminothiphene analogue 5 was inactive in the
NMR assay, commensurate with the lack of activity of
this compound in the binding and functional assays
(Table 1). Similarly, while the potent phenylpyridine
enantiomer 6 afforded near-complete inhibition of
glycogenolysis at a concentration of 1 lM (Fig. 9), the
less active phenylpyridine 7 afforded a substantially
reduced inhibition of glycogenolysis at this concentra-
tion, commensurate with the diminished activity of 7
in the binding and functional assays (Table 1).
Figure 9. Concentration-dependent inhibition of glucagon induced
glycogenolysis by synthetic glucagon antagonist 6, and the substan-
tially weaker inhibition by the less active enantiomer 7, in perfused
hGCGR mouse liver.
the data from the individual assays are plotted in Figure
9 as a measure of the reproducibility of the experiment
between individual perfused livers. The potent phenyl-
pyridine antagonist 6 afforded near-complete inhibition
of glucagon-induced glycogenolysis at a concentration
of 1 lM, whereas the less active enantiomer 7 afforded
little inhibition at this concentration.
The ex vivo NMR assay may afford validation of a given
lead class for pharmacological efficacy. However, the
correlation of ex vivo glycogenolysis inhibition with
in vitro potency is limited when comparing between
two different lead classes. For example, the phenylpyri-
dine antagonist 6 displayed a substantially greater
intrinsic potency in the binding and cAMP assays than
the peptide 1 (Table 1), but each afforded a similar inhi-
bition of hepatic glycogenolysis at similar concentra-
tions in perfused livers (Figs. 6 and 9).
3. Discussion
The validation of activity of a chemical lead class
toward a biological target represents one of the most
formidable challenges in early lead development.
Factors such as poor solubility, molecular aggregation,
or chemical reactivity may conspire to produce artifac-
tual activity in a cellular or membrane-based assay.15
The confirmation of efficacy of chemical leads may be
obtained by in vivo experiments, where the measurable
pharmacological effect of a compound interaction with
the biological target may be observed. However, the
possibility of poor bioavailability or interfering biologi-
cal interactions in the early examples of chemical lead
The antagonists 2 and 3 exhibit similar activity in the liv-
er NMR assay at concentrations of 10 lM (Fig. 7), but
higher concentrations of 3 do not afford a concomitant
increase in the inhibition of glycogenolysis. This dispar-
ity in activity between 2 and 3 at higher concentrations is

1514
S. M. Cohen et al. / Bioorg. Med. Chem. 14 (2006) 1506–1517
in contrast to the similar intrinsic potency exhibited by
these antagonists in the cAMP assay (Table 1). The
source of the discrepancies between the ex vivo liver as-
say and the in vitro functional assay is not entirely clear
at present. During the course of the experiment, the per-
fusate solution containing substrate, glucagon, and
antagonist was recirculated through the functioning
hGCGR liver. It is possible that even under these con-
trolled conditions the antagonists afford differing levels
of accumulation in liver tissue, and this could be inves-
tigated using a modification of the tissue lysis methods
reported by Dallas-Yang et al.6b Alternatively, differen-
tial rates of hepatic metabolism of the antagonists stud-
ied may be a factor affecting the direct comparison of
compounds. The recirculating solution could be sam-
pled and analyzed during the course of the experiment
to account for this possibility, but such an analysis
was not performed in the present study.
compound. Whereas excellent reproducibility was
observed in multiple control experiments (Fig. 4), only
limited duplicate concentration assays have been
run with synthetic antagonists. In those instances where
duplicate antagonist experiments were performed
(Fig. 9), the NMR assay exhibits excellent reproducibility.
4. Conclusions
The hGCGR perfused-liver NMR experiment has been
developed as an assay that affords the direct observation
of the effect of novel glucagon receptor antagonists spe-
cifically on the intended biochemical pathway. The tech-
nique allows for the pharmacological validation of lead
compounds in the entire organ of interest prior to the
optimization of pharmacokinetic properties. By appro-
priate choice of the 13C-labeled substrate, the biochem-
ical pathway of interest was spectroscopically isolated
for observation. The assay described here utilized per-
fused livers from transgenic mice expressing the human
glucagon receptor, thus minimizing potential species
variations in compound efficacy. This technique does
not provide adequate quantitative data for an accurate
comparison of compounds with the precision of in vitro
assays. Rather, this assay system represents a new devel-
opment for validating the activities of glucagon receptor
antagonists under more physiological conditions.
Notwithstanding the possibility of accumulation or
hepatic metabolism of the compounds during the exper-
iment, there are multiple additional variables that are
absent in the ex vivo liver preparation, but that would
be introduced during in vivo pharmacodynamic experi-
ments. For example, in vivo experiments are subjected
to differences in molecular absorption and tissue distri-
bution, and binding of compounds to components of
plasma, and any of these may hinder the early validation
of exploratory lead classes of glucagon receptor antago-
nists. Indeed, we have reported previously the capricious
in vivo efficacy of 2, 3, and 4 in blocking glucagon-in-
duced glucose elevation in mice.9a,11 The perfused-liver
NMR assay presented herein affords validation of these
compounds as effective antagonists of the human gluca-
gon receptor, in that they block glucagon-mediated gly-
cogenolysis in a concentration-dependent fashion in the
entire organ expressing the human variant of the biolog-
ical target. This ex vivo assay is not intended to supplant
the demonstration of pharmacodyamic efficacy in the
drug discovery process. It serves to bridge the gap be-
tween the observation of in vitro activity of a lead class
and the demonstration of efficacy in vivo.
5. Experimental
5.1. Chemistry
All reagents and chemicals used were of the highest purity
commercially available and were used without further
1
purification. H NMR spectra and 13C NMR spectra of
synthetic compounds were recorded on a Varian InNova
500 MHz instrument in CDCl3 or CD3OD as specified.
Low-resolution mass spectra (MS) were determined on
a Micromass Platform liquid chromatography–mass
spectrometer (LC–MS). High-resolution mass spectra
were acquired from a Micromass Q-TOF quadropole-
time-of-flight mass spectrometer. All MS experiments
were performed using electrospray ionization in positive
ion mode. Leucine enkephalin was applied as a lock-mass
reference for accurate mass analysis. All synthetic inter-
mediates and final products were purified to >96% chem-
ical purity as indicated by LC–MS, HPLC–UV, and
NMR. Silica gel chromatography was performed with a
Biotage Horizon(R) Flash Chromatography system from
Dyax, Inc. using the gradient solvent systems indicated
as per the manufacturerꢀs instructions. The peptidyl
antagonist 1 is commercially available from Bachem
Bioscience, King of Prussia, PA, USA. The triarylpyrrole
antagonist 2 was synthesized as described previously.9
One shortcoming of the NMR glycogenolysis assay is the
relatively low throughput for optimization of com-
pounds, and in this respect the assay is similar to in vivo
pharmacodynamic assays. Animal-to-animal variability
may also be a factor in the observed results, although
within a given mouse strain the activities of the hepatic en-
zymes controlling flux through the gluconeogenic path-
way tend to be closely similar for similarly treated
animals.7a–c,8a
The plot from the control experiments in Figure 4 in-
cludes standard deviation error bars and number of
experiments for each group. It is evident from the small
size of the error bars in Figure 4 that the reproducibility
of glycogen synthesis and glucagon-induced glycogenol-
ysis in control livers was excellent over multiple experi-
ments under these meticulously controlled conditions.
The NMR integration plots involving glucagon receptor
antagonists illustrated in Figures 6–8 represent a single
hGCGR mouse liver for each concentration of each
5.1.1. N-(3-Cyano-6-tert-pentyl-4,5,6,7-tetrahydro-1-benz-
othien-2-yl)cyclopentanecarboxamide (3)
5.1.1.1. Step A. 2-Amino-6-(1,1-dimethylpropyl)-4,5,6,7-
tetrahydro-1-benzothiophene-3-carbonitrile (II). To a solu-
tion of 5.00 g (29.8 mmol) of 4-tert-pentylcyclohexanone
in 20 mL ethanol at 0 °C was added 1.97 g (29.8 mmol)

S. M. Cohen et al. / Bioorg. Med. Chem. 14 (2006) 1506–1517
1515
ofmalononitrile, followed by3.89 mL(44.6 mmol)of mor-
pholine, then 1.90 g (59.5 mmol) of elemental sulfur. The
mixture was warmed to ambient temperature and stirred
for 16 h. The reaction was then concentrated and purified
by flash chromatography (5–15% ethyl acetate in hexane
gradient) affording 7.32 g (29.5 mmol) of the title com-
pound as a beige solid (99%). 1H NMR (500 MHz, CDCl3)
d 4.57(s, 2H), 2.66(m, 1H), 2.49(m, 1H), 2.44(m, 1H), 2.32
(m, 1H), 1.97 (m, 1H), 1.63 (m, 1H), 1.35 (m, 2H), 0.89 (s,
3H), 0.87 (s, 3H), 8.85 (t, J = 7.5 Hz, 3H); 13C NMR
(125 MHz, CDCl3) d 160.3, 132.6, 121.7, 115.7, 88.7,
43.1, 35.1, 32.8, 25.8, 25.6, 24.3, 24.2, 23.6, 8.4; HRMS:
calcd for C14H20ON2SH: 249.142. Found: 249.140.
ature,the mixture was poured onto 300 mL of saturated
aqueous NaHCO3 and extracted with two 200 mL por-
tions of CH2Cl2. The combined organic layers were
dried (Na2SO4) and concentrated. Purification by flash
chromatography (2–7% ethyl acetate in hexanes) affor-
ded 3.39 g (10.1 mmol) the title compound as a yellow
solid (50%). 1H NMR (500 MHz, CDCl3) d 9.04 (s,
1H), 2.61 (s, 3H), 2.38 (m, 1H), 1.78 (m, 2H), 1.68 (m,
2H), 1.58 (m, 9H), 0.98 (t, J = 7.5 Hz, 6H); 13C NMR
(125 MHz, CDCl3) d 174.4, 161.6, 151.5, 142.8, 120.4,
114.2, 82.5, 50.6, 28.5, 25.7, 14.6, 12.1; HRMS: calcd
for C17H24N2O3SH: 337.159. Found: 337.157.
5.1.2.3. Step C. (1Z)-2-(2,4-Dichlorophenyl)-N0-hydroxy-
ethanimidamide (VI). To a solution of 2.00 g (10.7 mmol)
(2,4-dichlorophenyl)acetonitrile in 50 mL ethanol was add-
ed 3.75 mL (26.9 mmol) of triethylamine, followed by
1.64 g (23.7 mmol) of hydroxylamine hydrochloride. The
mixture was heated to 100 °C for 2 h, then cooled to ambi-
ent temperature and concentrated. The residue was purified
by flash chromatography (50–100% ethyl acetate in hex-
anes) affording 1.23 g (5.64 mmol) of the title compound
5.1.1.2. Step B. N-(3-Cyano-6-tert-pentyl-4,5,6,7-tet-
rahydro-1-benzothien-2-yl)cyclopentanecarboxamide (3).
To a solution of 3.97 g (16.0 mmol) of 2-amino-6-(1,1-
dimethylpropyl)-4,5,6,7-tetrahydro-1-benzothiophene-
3- carbonitrile (II) in 100 mL CH2Cl2 was added 8.36 mL
(48.1 mmol) of di-iso-propylethylamine, followed by
2.75 mL (20.0 mmol) of 2-ethylbutanoyl chloride. After
16 h at ambient temperature, the mixture was poured
onto 300 mL of saturated aqueous NaHCO3 and extract-
ed with two 100 mL portions of CH2Cl2. The combined
organic layers were dried (Na2SO4) and concentrated.
Purification by flash chromatography (2–7% ethyl acetate
in hexane gradient) afforded 4.65 g (13.4 mmol) of the title
compound as an off-white solid (84%). 1H NMR
(500 MHz, CDCl3) d 8.45 (s, 1H), 2.76 (dd, J = 5.0 Hz,
J = 16.5 Hz, 2H), 2.65 (dd, J = 5.0 Hz, J = 16.0 Hz,
2H), 2.50 (m, 2H), 2.41 (m, 2H), 2.24 (m, 2H), 2.04 (m,
2H), 1.73 (m, 3H), 1.64 (m, 4H), 1.35 (m, 4H), 0.95 (t,
J = 7.5 Hz, 6H), 0.90 (d, J = 8.0 Hz, 4H), 0.852 (t,
J = 7.5 Hz, 3H); 13C NMR (125 MHz, CDCl3) d 174.7,
159.6, 156.6, 151.8, 145.1, 134.8, 134.5, 131.9, 129.7,
128.1, 116.4, 114.0, 97.1, 50.4, 34.3, 25.6, 14.9, 12.1;
HRMS: calcd for C20H30N2OSH: 347.216. Found:
347.217.
1
as a white solid (53%). H NMR (500 MHz, CDCl3) d
7.43 (d, J = 2.5 Hz, 1H), 7.32 (d, J = 8.0 Hz, 1H), 7.25
(dd, J = 2.0 Hz, J = 8.5 Hz, 1H), 4.61 (s, 3H), 3.60 (s,
2H); 13C NMR (125 MHz, CD3OD) d 153.1, 134.8,
133.8, 133.1, 131.3, 128.8, 127.2, 33.9; HRMS: calcd for
C8H8Cl2N2OH: 219.010. Found: 219.008.
5.1.2.4. Step E. N-{5-[({[(1Z)-1-Amino-2-(2,4-dichlor-
ophenyl)-ethylidene]amino}-oxy)carbonyl]-3-cyano-4-meth-
yl-2-thienyl}-2-ethylbutanamide (VII). To a solution of
3.35 g (1.00 mmol) of tert-butyl 4-cyano-5-[(2-ethylbuta-
noyl)amino]-3-methylthiophene-2-carboxylate (V) in
50 mL CH2Cl2 was added 20 mL trifluoroacetic acid.
After 2 h at ambient temperature, the mixture was con-
centrated, affording 2.80 g (1.00 mmol) of the corre-
1
sponding carboxylic acid as a white solid (100%). H
NMR (500 MHz, CD3OD) d 2.62 (m, 1H), 2.57 (s, 3H),
1.70 (m, 2H), 1.59 (m, 2H), 0.92 (t, J = 7.5 Hz); mass
spectrum (ES) m/e 281 (M+1). To a solution of 1.58 g
(0.56 mmol) of this intermediate in 100 mL dimethyl-
formamide was added 1.23 g (0.56 mmol) of (1Z)-2-
5.1.2. N-{3-Cyano-5-[3-(2,4-dichlorobenzyl)-1,2,4-oxadiazol-
5-yl]-4-methyl-2-thienyl}-2-ethylbutanamide (4)
5.1.2.1. Step A. tert-Butyl 5-amino-4-cyano-3-methylthi-
ophene-2-carboxylate (IV). To a solution of 4.98 mL
(30.0 mmol) of tert-butyl acetoacetate in 30 mL ethanol
was added 1.98 g (30.0 mmol) of malononitrile, followed
by 3.92 mL (45.0 mmol) of morpholine, then 1.92 g
(60 mmol) of elemental sulfur. The mixture was heated to
70 °C for 2.5 h, then cooled to ambient temperature and con-
centrated. Purification by flash chromatography (5–20%
ethyl acetate in hexanes) afforded 4.81 g (20.2 mmol) of the
title compound as an off-white solid (67%). 1H NMR
(500 MHz, CDCl3) d 5.20 (s, 2H), 2.49 (s, 3H), 1.55 (s,
9H); 13C NMR (125 MHz, CDCl3) d 164.2, 161.6, 145.3,
114.8, 113.4, 92.9, 82.2, 28.6, 15.3; LC–MS (ES) m/e 183
(M+1 tert-butyl).
(2,4-dichlorophenyl)-N0-hydroxyethanimidamide
(VI),
followed by 2.90 mL (1.69 mmol) of di-iso-propylethyl-
amine, 0.762 g (0.56 mmol) of 1-hydroxybenzotriazole,
then 1.08 g (0.56 mmol) of N-(3-dimethylaminopropyl)-
N0-ethylcarbodiimide hydrochloride. After 16 h at ambi-
ent temperature, the mixture was poured onto 300 mL
of 0.5 M aqueous NaHCO3 and extracted with 300 mL
ethyl acetate. The organic layer was washed with brine,
dried (MgSO4), and concentrated. Purification by flash
chromatography (15–35% ethyl acetate in hexanes) affor-
ded 1.97 g (41.0 mmol) of the title compound as a yellow
solid (73%). 1H NMR (500 MHz, CDCl3) d 9.38 (s, 1H),
7.44 (d, J = 2.0 Hz, 1H), 7.42 (d, J = 18.5 Hz, 1H), 7.26
(dd, J = 2.0 Hz, J = 8.5 Hz, 1H), 4.91 (s, 2H), 3.75 (s,
2H), 2.67 (s, 3H), 2.43 (m, 1H), 1.76 (m, 2H), 1.67 (m,
2H), 0.97 (t, J = 7.5 Hz, 6H); 13C NMR (125 MHz,
CDCl3) d 163.1, 147.0, 132.6, 131.4, 131.0, 130.2, 129.3,
128.2, 114.8, 94.1, 43.0, 35.1, 32.9, 32.8, 25.4, 25.3, 24.6,
24.5, 24.3, 24.2, 23.6, 8.4; HRMS: calcd for
C21H22Cl2N4O3SH: 481.087. Found 481.084.
5.1.2.2. Step B. tert-Butyl 4-cyano-5-[(2-ethylbuta-
noyl)amino]-3-methylthiophene-2-carboxylate (V). To a
solution of 4.81 g (20.2 mmol) of tert-butyl 5-amino-4-
cyano-3-methylthiophene-2-carboxylate (IV) in 100 mL
CH2Cl2 was added 10.5 mL (60.6 mmol) of di-iso-pro-
pylethylamine, followed by 3.45 mL (25.3 mmol) of
2-ethylbutanoyl chloride. After 16 h at ambient temper-

1516
S. M. Cohen et al. / Bioorg. Med. Chem. 14 (2006) 1506–1517
5.1.2.5. Step F. N-{3-Cyano-5-[3-(2,4-dichlorobenzyl)-
bation at room temperature, the radioactivity bound to
the cell membranes was determined in a scintillation
counter (MicrobetaÒ—Perkin-Elmer Life Sciences, Bos-
ton, MA). The IC50 values were calculated using nonlin-
ear regression (curve fit) analysis assuming single site
competition.
1,2,4-oxadiazol-5-yl]-4-methyl-2-thienyl}-2-ethylbutanamide
(4). A solution of 1.80 g (3.74 mmol) of N-{5-[({[(1Z)-1-
amino-2-(2,4-dichlorophenyl)ethylidene]amino}oxy)carbon-
yl]-3-cyano-4-methyl-2-thienyl}-2-ethylbutanamide (VII) in
80 mL diglyme was heated to 135 °C for 16 h. The reaction
was cooled to ambient temperature and concentrated. Puri-
fication of the residue by flash chromatography (5–15%
ethyl acetate in hexanes) afforded 1.23 g (26.6 mmol) of
the title compound as a white solid (71%). 1H NMR
(500 MHz, CDCl3) d 9.12 (s, 1H), 7.45 (d, J = 2.5 Hz,
1H), 7.33 (d, J = 8.0 Hz, 1H), 7.26 (dd, J = 2.5 Hz,
J = 6.0 Hz, 1H), 4.24 (s, 2H), 2.69 (s, 3H), 2.40 (m, 1H),
1.79 (m, 2H), 1.69 (m, 2H), 0.98 (t, J = 7.5 Hz, 6H); 13C
NMR (125 MHz, CDCl3) d 174.3, 171.4, 168.6, 152.6,
141.1, 135.3, 134.1, 132.1, 129.8, 127.6, 115.3, 113.9,
112.7, 96.9, 50.6, 29.8, 25.6, 15.1, 12.1; HRMS: calcd for
C21H20Cl2N4O2SH: 463.076. Found: 463.076.
5.2.2. Adenylyl cyclase assay. cAMP levels in CHO-
hGCGR were determined with the aid of an adenylyl
cyclase assay kit (SMP-004B, Perkin-Elmer Life Scienc-
es, Boston, MA) as per manufacturerꢀs instructions. For
assessment of the antagonistic activity, cells were
incubated with compound or 5% DMSO (as a control)
in Flash plates(R) coated with anti-cAMP antibodies
for 30 min and then stimulated with glucagon
(250 pM) for an additional 30 min. All incubations were
performed at room temperature in a 1:1 mixture of
stimulation buffer and cell resuspension buffer. The cell
stimulation was stopped by addition of an equal amount
of a detection buffer containing cell lysis agent and
125I-labeled cAMP tracer. After 3–6 h of incubation at
room temperature, the 125I-cAMP bound to the plate
was determined using a liquid scintillation counter
(MicrobetaÒ—Perkin-Elmer Life Sciences, Boston,
MA) and used to quantitate the amount of cAMP
present in each sample. The IC50 values were calculated
using nonlinear regression (curve fit) analysis.
5.1.3. 4-Bromo-N-[3-cyano-6-(1,1-dimethylpropyl)-4,5,6,
7-tetrahydro-1-benzothien-2-yl]benzamide (5). To a solu-
tion of 3.23 g (13.0 mmol) of 2-amino-6-(1,1-dimethyl-
propyl)-4,5,6,7-tetrahydro-1-benzothiophene-3-carbonitrile
(II) in 100 mL CH2Cl2 was added 6.79 mL (39.1 mmol) if
di-iso-propylethylamine, followed by 3.57 g (16.3 mmol)
of 4-bromobenzoyl chloride. After 16 h at ambient tem-
perature, the reaction was poured onto 200 mL of saturat-
ed aqueous NaHCO3. The organic layer was dried
(Na2SO4) and concentrated. Purification by flash chro-
matography afforded 0.340 g (0.787 mmol) of the title
5.3. Animals
1
compound as a yellow solid (6%). H NMR (500 MHz,
The perfused livers from male mice, 28–34 g, from the
hGCGR replacement strain reported previously6 were
used in all NMR experiments. The animals used in these
experiments were bred and housed within our facility,
and received chow (Harlan Teklad 7012) and water ad li-
bitum. Perfusion studies were initiated during the mid-
dle of the dark sequence of a 12-h/12-h lighting cycle.
The use of animals was done under the purview of an
Institutional Animal Care and Use Committee, and all
applicable regulations and laws pertaining to the use
of laboratory animals were followed.
CDCl3) d 9.05 (s, 1H), 7.83 (dd, J = 1.5 Hz, J = 6.5 Hz,
2H), 7.69 (dd, J = 2.0 Hz, J = 7.0 Hz, 2H), 2.77 (dd,
J = 5.0 Hz, J = 16.5 Hz, 1H), 2.70 (dd, J = 4.5 Hz,
J = 15.5 Hz, 1H), 2.52 (m, 2H), 2.37 (m, 1H), 2.04 (m,
1H), 1.65 (m, 1H), 1.40 (m, 2H), 0.92 (s, 3H), 0.91 (s,
3H), 0.87 (t, J = 8.0 Hz, 3H); 13C NMR (125 MHz,
CDCl3) d 162.9, 146.9, 132.6, 131.4, 130.9, 130.2, 129.2,
128.3, 114.7, 94.1, 43.0, 35.1, 32.8, 25.4, 25.3, 24.5, 24.3,
24.2, 23.6; HRMS: calcd for C21H23BrN2OSH: 431.079
and 433.077. Found: 431.078 and 433.078.
5.2. Biology
5.4. Perfused liver preparations and NMR measurements
Chinese hamster ovary (CHO) cells expressing the cloned
human glucagon receptor (CHO-hGCGR) have been de-
scribed.17 [2-13C]Pyruvate (P99%) isotopic purity was
purchased from Cambridge Isotope Labs (Andover,
MA). Glucagon (1–29) and des-His1 [Glu9]glucagon-
NH2 (1) were purchased from Bachem Bioscience (King
of Prussia, PA, USA). Glucagon stock solutions were
assayed using glucagon RIA kits purchased from Linco
Research (St. Charles, MO, USA).
13C and 31P NMR spectra of perfused liver at 37 °C were
measured at either 8.5 or 11.7 T (AM360WB or Avance
500WB NMR spectrometers, Bruker Instruments, Bille-
rica, MA) using probes in which the main coil was dou-
ble-tuned for both the 13C and 31P frequencies. First,
four baseline 31P spectra of the perfused liver, each
200 scans (8 min), were accumulated over the first
32 min. Second, one natural abundance 13C spectrum
of 800 scans was accumulated over 11 min. Third, initial
addition of 13C-labeled substrate was followed by accu-
mulation of 17 additional 13C spectra under isotopic
steady-state conditions. After measurement of the final
13C spectrum, two additional 31P spectra were
accumulated.
5.2.1. Glucagon receptor binding assay. To determine the
binding affinity of selected compounds, 2 lg of cell
membranes from the CHO-hGCGR cells was incubated
with 125I-glucagon (NEN) in a buffer containing 50 mM
TrisÆHCl, pH 7.5, 5 mM MgCl2, 2 mM EDTA, 12%
glycerol, and 200 lg wheat germ agglutinin coated
PVT SPA beads (Amersham), compounds. Non-spe-
cific binding was determined in the same buffer in the
presence of 1 lM unlabeled glucagon. After 4–12 h incu-
5.4.1. NMR observation of glucagon-mediated glycogen
metabolism. In preliminary perfusions of transgenic
GCGR murine livers (N = 20) under the same 13C
NMR conditions described above, glucagon was titrated

S. M. Cohen et al. / Bioorg. Med. Chem. 14 (2006) 1506–1517
1517
In Research in Perfused Liver: Clinical and Basic Applica-
tions; Ballet, F., Thurman, R. G., Eds.; John Libby:
London, 1991, Chapter 4; (d) Bergans, N.; Dresselaers, T.;
Vanhamme, L.; Van Henke, P.; Van Huffel, S.; Vanstapel,
F. NMR Biomed. 2003, 16, 36.
using one single concentration, between 20 and 70 pM,
for each perfused liver to establish the minimum level
of glucagon required to reproducibly induce almost
complete glycogenolysis and inhibition of [13C]glycogen
accumulation (data not shown). Under these conditions,
the optimum concentration was determined to be 50 pM
glucagon; this concentration is approximately equiva-
lent to the physiological level reported for wild-type
mice in vivo.16 In other similar preliminary perfused liv-
ers (N = 5), dimethylsulfoxide (DMSO) was titrated to
establish the level of DMSO that had no effect on hepatic
metabolism under these NMR conditions. The no-effect
level was found to be 0.25% DMSO. Consequently, the
non-peptide hGCGR antagonists 2–7 were prepared in
stock solutions that provided for infusion of exactly
0.25% DMSO in all cases. For the quantitation of net
[13C]glycogen accumulation in perfused hGCGR mouse
livers, peak areas in the 13C NMR spectra were deter-
mined as described previously8a by fitting spectra to
appropriate line shapes with subsequent integration using
the CF module of NMR1 software or similar functions
included in the Matlab software package. Absolute inte-
gral units, normalized as a percent of the maximum pre-
glucagon value for the given liver, were used in comparing
the area of the 13C enrichment at C-1 of the glucosyl units
in glycogen (peak A) in the perfused liver preparations.
8. (a) Cohen, S. M.; Werrmann, J. G.; Tota, M. R. Proc.
Natl. Acad. Sci. U.S.A. 1998, 95, 7385; (b) Cohen, S. M.
Anal. N.Y. Acad. Sci. 1987, 508, 109.
9. (a) Cascieri, M. A.; Koch, G. E.; Ber, E.; Sadowski, S. J.;
Louizides, D.; de Laszlo, S. E.; Hacker, C.; Hagmann, W.
K.; MacCoss, M.; Chicchi, G. G.; Vivario, P. P. J. Biol.
Chem. 1999, 274, 8694; (b) de Laszlo, S. E.; Hacker, C.; Li,
B.; Kim, D.; MacCoss, M.; Mantlo, N.; Pivnichny, J. V.;
Colwell, L.; Koch, G. E.; Cascieri, M. A.; Hagmann, W.
K. Bioorg. Med. Chem. Lett. 1999, 9, 641.
10. Ladouceur, G. H.; Cook, J. H.; Hertzog, D. H.; Jones, H.;
Hunertmark, T.; Korpusik, M.; Lease, T. G.; Livingston,
J. N.; MacDougall, M. L.; Osterhout, M. H.; Phelan, K.;
Romero, R. H.; Schoen, W. R.; Shao, C.; Smith, R.
Bioorg. Med. Chem. Lett. 2002, 12, 3421.
11. (a) Qureshi, S. A.; Candelore, M. R.; Xie, D.; Yang, X.;
Tota, L. M.; Ding, V. D.-H.; Li, Z.; Bansal, A.; Miller, C.;
Cohen, S.; Jiang, G.; Brady, E.; Saperstein, R.; Duffy, J. L.;
Tata, J. R.; Chapman, K. T.; Moller, D. E.; Zhang, B. B.
Diabetes 2004, 53, 3267; (b) Duffy, J. L.; Kirk, B. A.;
Konteatis, Z.; Campbell, E. L.; Liang, R.; Brady, E. J.;
Candelore, M. R.; Ding, V. D.-H.; Jiang, G.; Liu, F.;
Qureshi, S. A.; Saperstein, R.; Szalkowski, D.; Tong, S.;
Tota, L. M.; Xie, D.; Yang, X.; Zafian, P.; Zheng, S.;
Chapman, K. T.; Zhang, B. B.; Tata, J. R. Bioorg. Med.
Chem. Lett. 2005, 15, 1401.
Acknowledgments
12. For a review of this method, see: Sabnis, R. W.;
Rangnekar, D. W.; Sonawane, N. D. J. Heterocycl. Chem.
1999, 36, 3333.
We are very grateful to William Hagmann and Scott
Edmondson for their assistance in the preparation of
the manuscript and for numerous helpful discussions.
We are also grateful to Falguni Patel for assistance in
the characterization of compounds.
13. Previous accounts from these laboratories (Ref. 9) reported
a marked decrease in binding potency of compound 2 in
the presence of 5 mM Mg2+ in a whole cell (filtration)
binding assay, which effected a shift in the binding IC50
from 7 nM (ꢀMg2+) to 170 nM (+Mg2+). This dependence
was not observed in the present membrane-based SPA-
binding assay described in Section 5, which is run in the
presence of 5 mM Mg2+. The source of the discrepancy in
Mg2+ dependence between the two assay formats is not
fully understood at present. The functional antagonist
activity reported here (cAMP IC50 = 120 nM) is in mod-
erate agreement with that reported using a previous cell
line and version of the functional assay (cAMP
IC50 = 41 nM, Ref. 9b). For the sake of consistency, all
activity values reported in Table 1 were derived from
the present assay methods and cell lines as described in
Section 5.
14. The binding IC50 values determined for 6 and 7 are sixfold
more potent than the values reported in Ref. 10 for these
compounds. The cAMP IC50 value determined for 6 is
ninefold more potent that the value reported in Ref. 10,
and a cAMP IC50 for 7 has not been previously reported.
The discrepancies may be due to different cell lines used
for the assays in Table 1 as compared to those used in Ref.
10. For the sake of consistency, all activity values reported
in Table 1 were derived from the present assay methods
and cell lines as described in Section 5.
References and notes
1. (a) Zhang, B. B.; Moller, D. B. Curr. Opin. Chem. Biol.
2000, 4, 461; (b) Jiang, G.; Zhang, B. B. Am. J. Physiol.
Endocrinol. Metab. 2003, 284, E671; (c) Morral, N. Trends
Endocrinol. Metab. 2003, 14, 169.
2. Moller, D. E. Nature 2001, 414, 821.
3. (a) Kurukulasuriya, R.; Link, J. T.; Madar, D. J.; Pei, Z.;
Rhode, J. J.; Richards, S. J.; Souers, A. J.; Szczepankie-
wicz, B. G. Curr. Med. Chem. 2003, 10, 99; (b)
Kurukulasuriya, R.; Link, J. T.; Madar, D. J.; Pei, Z.;
Richards, S. J.; Rhode, J. J.; Souers, A. J.; Szczepankie-
wicz, B. G. Curr. Med. Chem. 2003, 10, 123; (c) Sloop, K.
W.; Michael, M. D. Drugs Future 2004, 29, 835.
4. (a) Unson, C. G.; Andreu, D.; Gurzenda, E. M.; Merri-
field, R. B. Proc. Natl. Acad. Sci. U.S.A. 1987, 84, 4083;
(b) Hruby, V. J.; Ahn, J.-M.; Trivedi, D. Curr. Med.
Chem.—Immunol. Endocr. Metab. Agents 2001, 1, 199.
5. (a) Ling, A. Drugs Future 2002, 27, 987; (b) Ling, A. L.;
Wasserman, J. I. Expert Opin. Ther. Patents 2003, 13, 1.
6. (a) Shiao, L.-L.; Cascieri, M. A.; Trumbauer, M.; Chen,
H.; Sullivan, K. A. Transgenic Res. 1999, 8, 295; (b)
Dallas-Yang, Q.; Shen, X.; Strowski, M.; Brady, E.;
Saperstein, R.; Gibson, R. E.; Szalkowski, D.; Candelore,
M. R.; Fenyk-Melody, J. E.; Parmee, E. R.; Zhang, B. B.;
Jiang, G. Eur. J. Pharmacol. 2004, 501, 225.
15. McGovern, S. L.; Caselli, E.; Grigorieff, N.; Shoichet, B.
K. J. Med. Chem. 2002, 45, 1712.
16. Parker, J. C.; Andrews, K. M.; Allen, M. R.; Stock, J. L.;
McNeish, J. D. Biochem. Biophys. Res. Commun. 2002, 290,
839.
17. Chicchi, G. C.; Graziano, M. P.; Koch, G.; Hey, P.;
Sullivan, K.; Vivario, P. P.; Cascieri, M. A. J. Biol. Chem.
1997, 272, 7765.
7. (a) Cohen, S. M. J. Biol. Chem. 1983, 258, 14294; (b)
Cohen, S. M. Biochemistry 1987, 26, 563; (c) Cohen, S. M.