Integration of optimized substituent patterns to produce highly potent 4-aryl-pyridine glucagon receptor antagonists

Article abstract of DOI:10.1016/S0960-894X(02)00736-9

Optimized substituent patterns in 4-aryl-pyridine glucagon receptor antagonists were merged to produce highly potent derivatives containing both a 3-[(1R)-hydroxyethyl] and a 2′-hydroxy group. Due to restricted rotation of the phenyl-pyridine bond, these analogues exist as four isomers. A diastereoselective methylcopper reaction was developed to facilitate the synthesis, and single isomers were isolated with activities in the range IC₅₀=10-25 nM.

Full text of DOI:10.1016/S0960-894X(02)00736-9

Bioorganic & Medicinal Chemistry Letters 12 (2002) 3421–3424
Integration of Optimized Substituent Patterns to Produce Highly
Potent 4-Aryl-pyridine Glucagon Receptor Antagonists
Gaetan H. Ladouceur,a,* James H. Cook,a Donald L. Hertzog,a J. Howard Jones,a
Thomas Hundertmark,a Mary Korpusik,a Timothy G. Lease,a James N. Livingston,b
Margit L. MacDougall,b Martin H. Osterhout,a Kathleen Phelan,a Romulo H. Romero,a
William R. Schoen,a Chunning Shaoa and Roger A. Smitha
aDepartment of Chemistry Research, Bayer Research Center, 400 Morgan Lane, West Haven, CT 06516, USA
bDepartment of Metabolic Disorders Research, Bayer Research Center, 400 Morgan Lane, West Haven, CT 06516, USA
Received 27 June 2002; accepted 23 August 2002
Abstract—Optimized substituent patterns in 4-aryl-pyridine glucagon receptor antagonists were merged to produce highly potent
derivatives containing both a 3-[(1R)-hydroxyethyl] and a 20-hydroxy group. Due to restricted rotation of the phenyl–pyridine
bond, these analogues exist as four isomers. A diastereoselective methylcopper reaction was developed to facilitate the synthesis,
and single isomers were isolated with activities in the range IC50=10–25 nM.
# 2002 Elsevier Science Ltd. All rights reserved.
The incidence of type 2 diabetes is expected to sig-
nificantly increase by 2010.1 As no current therapies
prevent the disease’s progression, many research groups
are focusing their efforts toward the identification of
novel approaches.2 One such strategy targets modula-
tion of the glucagon receptor. The mechanism of action
of glucagon and its role in the regulation of blood glu-
cose levels have been known for quite some time.3
However, only during the last decade have studies
shown that potent peptide antagonists significantly
decrease blood glucose levels in diabetic animal mod-
els.4 More recently, it was demonstrated that a small
molecule could act as an antagonist, triggering a strong
interest in the research community.5
hybrid structure of type 4, striving for an additive effect
in potency. The two main moieties contributing to high
binding affinities with the glucagon receptor are the
3-[(1R)-hydroxyethyl] group of structure 2 and the
20-hydroxyl group of structure 3. In this article, we pre-
sent the preparation and SAR analysis of such com-
pounds (4), as well as the development of
diastereoselective methylcopper approach to address
a
challenges encountered in the synthesis of 4.
Compounds of type 4 have four possible isomers by
virtue of their two chiral elements; the stereocenter of
the 3-(1-hydroxyethyl) group, and the axis of chirality
resulting from restricted rotation about the phenyl–
pyridine bond.8 We decided to synthesize compound 4
first in an achiral fashion and separate the mixture of
isomers by chiral chromatography for biological eva-
luation.9 In situ formation of the magnesium complex of
diethyl acetone-1,3-dicarboxylate followed by con-
densation with isobutyryl chloride provided the 2,6-di-
isopropyl pyrone 6, by a one-pot process7 (Scheme 2).
Treatment of pyrone 6 with ammonia gave the
4-hydroxypyridine, which was methylated to the 4-meth-
oxypyridine 7. Treatment of compound 7 with Grignard
reagent 8 afforded phenylpyridine 9 in good yield.7
Selective reduction of one of the ester groups of com-
pound 9 with Red-Al1, followed by a PCC oxidation and
From a high throughput screening assay, we discovered
compound 1 (Scheme 1), which exhibited modest activ-
ity against the human glucagon receptor (binding affi-
nity, IC50=7 mM). Efforts to optimize compound 1 led
to the discovery of more potent analogues, such as 26
(R=1-propyl, IC50=0.11 mM) and 37 (R=1-pentyl,
IC50=0.19 mM). Thereafter, our goal was to merge the
key pharmacophoric groups of these two series into a
*Corresponding author. Tel.:+1-203-812-2055; fax:+1-203-812-2650;
e-mail: gaetan.ladouceur.b@bayer.com
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(02)00736-9

3422
G. H. Ladouceur et al. / Bioorg. Med. Chem. Lett. 12 (2002) 3421–3424
Figure 1. Potent glucagon antagonist analogue of type 4.
Compound 4a (X=F and R=C2H5, Fig. 1) was one of
the first examples of type 4 synthesized and was
obtained as a ca. 1:2 diastereomeric mixture. The dia-
stereomers were separated through standard silica gel
chromatography (10% ethyl acetate in hexanes), and
the faster-eluting diastereomer exhibited high affinity to
the human glucagon receptor (IC50=20 nM), while the
slower-eluting diastereomer had marginal activity
(IC50=2000 nM). Using chiral chromatography,10 the
former (more active) diastereomer was resolved, to
provide only one potent enantiomer (IC50=16 nM vs
5000 nM for the two enantiomers). This highly potent
analogue demonstrated that combining the key phar-
macophoric groups of compounds 3 and 4 resulted in an
additive effect in potency. Thus we focused our efforts
toward a detailed SAR evaluation of this new sub-class
of analogues. However, their synthesis via Scheme 3,
which produced the more potent diastereomer as the
minor product, was inefficient. The development of a
diastereoselective and scalable synthesis became a
priority.
Scheme 1. Optimization summary of a class of 4-aryl-pyridine gluca-
gon antagonists (R=ethyl to pentyl, X=H or F).
During our search for a diastereoselective synthesis, we
discovered that aldehyde 11 gave encouraging results
when treated with alkylcopper reagents using proce-
dures developed by Sato and coworkers.11 After a series
of optimization experiments, we found that the best
results were obtained when the methylcopper reagent
was generated by using a 1:1 ratio of methyllithium and
copper iodide in a mixture of toluene and ether (9:1) at
0 ꢀC (Scheme 4).12 Using these conditions, a nearly
quantitative yield of product was obtained, with a dia-
stereomeric ratio of >90:10 in favor of the desired ste-
reoisomer. This represented a substantial synthesis
improvement, as the original method (Scheme 3) pro-
vided predominantly the less potent diastereomer. The
diastereomers were separated by standard chromato-
graphy, and the racemic mixture of the desired isomer
was carried until the end of the synthesis, where they
were easily separated by chiral chromatography.10
Scheme 2. Synthesis of intermediate 11.
Wittig reaction afforded alkene 10. Standard LiAlH4
reduction of the ester group of compound 10, followed
by PCC oxidation gave the desired aldehyde precursor
11. Compound 11 was obtained as a racemic mixture,
since the benzyloxy group at the 20 position introduces
an element of chirality of the biaryl axis (vide supra).
Intermediate 11 was then treated with methyllithium
or methylmagnesium bromide to provide the
hydroxyethyl compound 12 (Scheme 3). Hydrogena-
tion of 12 served to saturate the olefin moiety as well
as to remove the benzyl protecting group, affording
the desired analogue 4.
With an improved synthesis in hand, we then focused our
efforts on a detailed SAR investigation of analogues of
type 4, and a summary of their activities is listed in Table
1. The activities (IC50) of diastereomers D1 (first-eluting
Scheme 3. Synthesis of analogues 4.
Scheme 4. Diastereoselective synthesis of intermediates 12.

G. H. Ladouceur et al. / Bioorg. Med. Chem. Lett. 12 (2002) 3421–3424
3423
Table 1. Glucagon antagonist activities of analogues 4
of compound 4c (Table 1) exhibited the greatest potency
with an IC50 of 10 nM. Also, a diastereoselective alkyl-
ation was developed to improve the synthetic accessi-
bility to this new series.
Acknowledgements
We would like to thank the members of the analytical
chemistry group for their valuable support. In parti-
cular, we are grateful to Rolf Grosser and John Dority
for their contributions in chiral separations, as well as
Axel Goehrt for X-ray crystal structure analysis.
Compd X
R
D1 IC50 D2 IC50 E1 of D1 IC50 E2 of D1 IC50
(nM)
(nM)
(nM)
(nM)
4a
4b
4c
4d
4e
F
H
F
F
Et
Pr
Pr
Bu
20
40
18
28
30
2000
1100
1000
1000
2200
5000
190
8000
—
16
25
10
—
12
F Pent
8000
References and Notes
D1, first diastereomer to elute (as mixture of E1 and E2); D2, second
diastereomer to elute; E1, first enantiomer to elute; E2, Second enan-
tiomer to elute.
1. Amos, A. F.; McCarty, D. J.; Zimmet, P. Diabetic Med.
1997, 14, S7.
2. (a) Zhang, B. B.; Moller, D. E. Curr. Opin. Chem. Biol.
2000, 4, 461. (b) Cobb, J.; Dukes, I. Annu. Rep. Med. Chem.
1998, 33, 213.
3. (a) Dohlman, H. G.; Thorner, J.; Caron, M. G.; Lefkowitz,
R. J. Annu. Rev. Biochem. 1991, 60, 653. (b) Christophe, J.
Biochim. Biophys. Acta 1995, 1241, 45. (c) Unger, R. H.; Orci,
L. N. Engl. J. Med. 1981, 1518. (d) Gerich, J. E.; Lorenzi, M.;
Bier, D.; Schneider, V.; Tsalikian, E.; Karam, J.; Forsham,
P. H. New Engl. J. Med. 1975, 292, 985. (e) Unger, R. H.
Metabolism 1978, 27, 1691.
diastereomer) ranged from 18 to 40 nM and were up to
100-fold more potent than diastereomers D2 (second-
eluting diastereomer).13 The second-eluting enantiomer
(E2) of diastereomer D1 was the more potent enantio-
mer, exhibiting slightly stronger activities as compared
to their parent diastereomers D1, while the first-eluting
enantiomers (E1) were considerably less active. As pre-
viously observed,14 the fluoro substituent at the 40 posi-
tion confers an approximately 2-fold increase in potency
(compare diastereomers D1 and enantiomers E2 of 4b
and 4c). The most potent analogue, E2 of 4c, exhibited
an IC50 value of 10 nM, establishing it as one of the
most active glucagon antagonists reported to date. This
compound was also evaluated for antagonistic activity
according to a published procedure.15 In that assay, the
potent isomer of analogue 4c (E2 of D1) exhibited
functional activity (cAMP production IC50=6.5 nM)
comparable to that measured in the receptor binding
affinity assay.
4. (a) Unson, C. G.; Gurzenda, E. M.; Merrifield, R. B. Pep-
tides 1989, 10, 1171. (b) Salzig, T. M.; Unson, C.; Merrifield,
B. Diabetes 1996, 45, 220.
5. (a) Livingston, J. N.; Schoen, W. R. Annu. Rep. Med.
Chem. 1999, 34, 189. (b) Connell, R. D. Exp. Opin. Ther. Pat
1999, 9, 701. (c) 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.; Vicario, P. P.
J. Biol. Chem. 1999, 274, 8694. (d) Madsen, P.; Knudsen, L. B.;
Wiberg, F. C.; Carr, R. D. J. Med. Chem. 1998, 41, 5150.
6. (a) Ladouceur, G. H.; Cook, J. H.; Doherty, E. M.;
Schoen, W. R.; MacDougall, M. L.; Livingston, J. N. Bioorg.
Med. Chem. Lett. 2002, 12, 461. (b) Schoen, W. R.; Ladou-
ceur, G. H.; Cook, J. H.; Lease, T. G.; Wolanin, D. J.;
Kramss, R. H.; Hertzog, D. L.; Osterhout, M. H. US Patent
6,218,431, 2001. Chem. Abstr. 2001, 134, 311111.
We also sought to determine the absolute configuration
of the most active isomers E2 (Table 1). Assuming that
the hydroxyethyl group retains the R configuration
established for compound 2,16 X-ray crystal structure
analysis17 allowed us to assign the absolute configur-
ation of compound 13 (Fig. 2), in which both hydroxy
groups are syn to each other and appear to form an
internal hydrogen bond.
7. Smith, R. A.; Hertzog, D. L.; Osterhout, M. H.; Ladou-
ceur, G. H.; Korpusik, M.; Bobko, M. A.; Jones, H.; Phelan,
K.; Romero, R. H.; Hundertmark, T.; MacDougall, M. L.;
Livingston, J. N.; Schoen, W. R. Bioorg. Med. Chem. Lett.
2002, 12, 1303.
8. Lloyd-Williams, P.; Giralt, E. Chem. Soc. Rev. 2001, 30,
145.
In summary, we have discovered new highly potent
glucagon antagonists by integrating the key pharmaco-
phoric groups of compounds 2 and 3. Isomer E2 of D1
9. All data reported herein reflect purified and characterized
(1H NMR, MS) samples. The human glucagon receptor bind-
ing assay was carried out as described previously.6a
10. Racemates were resolved by using a Waters Preparative
LC 2000HPLC system with a chiral column (BRB-9668A,
6ꢁ50 cm ID). The mobile phase consisted of 99% hexanes and
1% of a solution of 1:99 acetic acid/ethanol (150 mL/min). As
an example, for the racemate (D1) of compound 4c, E1 was
collected between 17 and 23 min and E2 was collected between
23 and 32 min.
11. Sato, F.; Kobayashi, Y.; Takahashi, O.; Chiba, T.;
Takeda, Y.; Kusakabe, M. J. Chem. Soc. Comm. 1985, 1636.
12. (a) An example of a typical procedure: To CuI (5.69 g,
29.9 mmol) in toluene (85 mL) was slowly added a solution of
1.4 M of methyllithium in diethyl ether (21 mL, 29.4 mmol) at
Figure 2. X-ray crystal structure (XP plot) of the potent glucagon
antagonist 13 (4a, E2 of D1 in Table 1).

3424
G. H. Ladouceur et al. / Bioorg. Med. Chem. Lett. 12 (2002) 3421–3424
0 ꢀC. The mixture was stirred for 15 min before the addition of
aldehyde 11 (5.5 mmol) in toluene (8 mL). During the addition
of aldehyde 11, the temperature was carefully monitored to
13. Standard errors for the binding data were ꢂ35%.
14. Table 2 in ref 6a.
15. Selected compounds were also tested in a functional cell
assay, and all were determined to be competitive antagonists.
IC50 values for cAMP production are as follows: for 4a-E2-
D1, 8 nM, 4b-E2-D1, 12 nM, 4c-E2-D1, 6.5 nM, 4e-E2-D1, 11
nM. The procedure for the functional assay has been described
previously.6a
16. The secondary alcohol of compound 2 was assigned the R
configuration by X-ray crystal structure analysis of its Mosher
ester derivative.6a
17. Computations and molecular graphics were performed
using SHELXTL-5 on a HP-Vectra 5/90 XU PC. The struc-
ture was solved by direct methods. All remaining non-hydro-
gen atoms were found by successive electron density map
calculations. H-atoms could also be detected, but were con-
structed in geometrically meaningful positions. Anisotropic
displacement parameters were applied to all but the H-atoms
for which an isotropic displacement factor (1.5 times Uequiv
of the corresponding atom for CH3 hydrogens, 1.2 times
Uequiv of the corresponding atom for all other H-atoms in the
structure) is calculated.
ꢀ
ensure that it was maintained around 0 C. The reaction mix-
ꢀ
ture was stirred for an additional 30 min at 0 C, then quen-
ched with a saturated solution of NH4OH (250 mL). After an
additional 30 min, the mixture was further diluted with a
saturated solution of NH4Cl (300 mL) and extracted with
ether (3ꢁ150 mL). The combined organic layers were dried
(MgSO4) and concentrated to give a colorless oil. The oil was
filtered through a pad of silica gel (5:1 hexanes/EtOAc) and
the filtrate concentrated to give the product as a colorless oil
(29.6 mmol, 99%).
(b) A mechanistic rationale for the different stereoselectivity
observed for the methylcopper reagent, as compared to that
observed for methyllithium, is complicated by the lack of a
clear understanding of the structures of these reagents in
solution. Indeed, we have found that the stereoselectivity is
influenced by the reagent as well as the solvent. Coordination
of the reagent to the benzyloxy group may play a role in this
process, but we have insufficient data to be able to propose an
interpretation for these findings.