Glucose-lowering in a db/db mouse model by dihydropyridine diacid glycogen phosphorylase inhibitors

Article abstract of DOI:10.1016/S0960-894X(03)00798-4

The synthesis of a series of novel dihdyropyridine diacid glycogen phosphorylase inhibitors is presented. SAR and functional assay data are discussed, along with the effect of a single inhibitor on blood glucose in a diabetic animal model.

Full text of DOI:10.1016/S0960-894X(03)00798-4

Bioorganic & Medicinal Chemistry Letters 13 (2003) 3405–3408
Glucose-Lowering in a db/db Mouse Model by Dihydropyridine
Diacid Glycogen Phosphorylase Inhibitors
Anthony K. Ogawa,a,* Chris A. Willoughby,a,, Raynald Bergeron,b
Kenneth P. Ellsworth,b Wayne M. Geissler,b Robert W. Myers,b Jun Yao,b
Georgianna Harrisb and Kevin T. Chapmana
aDepartment of Basic Chemistry, Merck Research Laboratories, Rahway, NJ 07065, USA
bDepartment of Metabolic Disorders-Diabetes, Merck Research Laboratories, Rahway, NJ 07065, USA
Received 16 April 2003; revised 22 July 2003; accepted 29 July 2003
This paper is dedicated to the memory of Chris A. Willoughby
Abstract—The synthesis of a series of novel dihdyropyridine diacid glycogen phosphorylase inhibitors is presented. SAR and
functional assay data are discussed, along with the effect of a single inhibitor on blood glucose in a diabetic animal model.
# 2003 Elsevier Ltd. All rights reserved.
A recent World Health Organization report estimates
that approximately 130 million people worldwide are
afflicted with type 2 diabetes.1 Patients suffer from
reduced insulin sensitivity and secretion that result in
glucose intolerance and postabsorptive hyperglycemia.
Hepatic glucose output (HGO), a significant con-
tributor to diabetic hyperglycemia,2 derives from both
the hydrolytic breakdown of stored glycogen (glyco-
genolysis), and gluconeogenesis, the latter of which is
elevated in type 2 diabetics relative to normal patients.3
Recent evidence suggests that a portion of the gluco-
neogenic flux cycles through the glycogenolytic path-
way,4 thereby identifying a viable intervention point for
restoring glycemic control. Phosphorylated glycogen
phosphorylase a (GPa) catalyzes the breakdown of gly-
cogen to glucose-1-phosphate, which represents the rate
determining step in glycogenolytic glucose production.
Glucose lowering by inhibition of GPa with small molcule
inhibitors has been reported for several unique structural
classes.5 GPa, therefore, represents an attractive ther-
apeutic target for the treatment of type 2 diabetes.
co-crystal structure (3AMV) indicated a binding site
also utilized by AMP to allosterically activate glycogen
phosphorylase b.7 Although AMP binding does not
impact GPa activity, the spatial proximity of the afore-
mentioned bound inhibitor to Ser14, the residue phos-
phorylated to activate GP, suggests an inhibitory
mechanism for this compound class.
At present, SAR exploration of the dihydropyridine
dicarboxylate scaffold and demonstrated efficacy in a
diabetic model remain unreported. We report herein the
synthesis and evaluation of racemic dihdyropyridine
GPa inhibitors that address the functional role of sev-
eral substituents, as well as an optimized GPa inhibitor
that effected sustained glucose lowering in a db/db
mouse model.
The core 1,4-dihydropyridine scaffold for all reported
inhibitors derived from a Hantzsch synthesis that
Recent literature reported the inhibition of glycogen-
olysis by a dihydropyridine dicarboxylate inhibitor
(U6751) in rat liver.6 (Fig. 1) An enzyme-inhibitor
*Corresponding author. Tel.: +1-732-594-4199; fax: +1-732-594-
9473; e-mail: anthony_ogawa@merck.com
,Deceased.
Figure 1. Previously reported inhibitor U6751.
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00798-4

3406
A. K. Ogawa et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3405–3408
Table 1. In vitro liver enzyme, muscle enzyme and hepatocyte inhi-
bition assay results for compounds 1–310,11
Compd
Structure
HLGPa
HMGPa
Cell
IC50 (nM) IC50 (nM) EC50 (mM)
rac-U6751
39
138
2.23
6.50
1
2
3
4
5
395
1195
15,600
692
43
15,000
172
52
NA
11.0
1.89
1.13
Scheme 1. Reagents and conditions: (i) R2NH2, MeOH, rt (crude
product >95% pure by 1H NMR); (ii) cat pTSA, CH3C(O)CH2-
CO2R4, benzene, reflux; (iii) 140 ꢀC, neat; (iv) NaOH, (1:1) dioxane/
H2O, 50 ꢀC; (v) TFA.
involved regio-selective condensation of enamine sub-
strates with a-carboxy-a,b-unsaturated ketones at ele-
vated temperature8 (Scheme 1). Cyclization was
followed by methyl ester hydrolysis, which proceeded
cleanly at 50 ꢀC overnight with no hydrolysis of the C5-
isopropyl ester observed. Isolation of the C2,C3-dicar-
boxylate as the bis-sodium salt yielded the desired pro-
duct.9 The synthesis of inhibitor 3 required TFA
deprotection 3,4-dimethoxybenzyl group (12) prior to
ester hydrolysis.
11
26
substitution (4) had little effect on potency relative to
rac-U6751, but the N1-benzyl inhibitor (5) was dis-
tinctly more potent against GPa, both in vitro and in
the hepatocyte cell-based assay as well. In order to
evaluate the potential for this inhibitor series, sub-
sequent SAR of the corresponding N1-benzylic moiety
was pursued and will be discussed later.
In vitro analysis of rac-U6751 and inhibitors 1–5 pro-
vided broad SAR for the dihydropyridine scaffold.
Qualitative analysis of the 3AMV structure suggested
that the C4-o-chlorophenyl substituent from U6751
oriented the C4-aryl group perpendicular to the dihy-
dropyridine ring system.7 Assay of the C4-phenyl deri-
vative (1) resulted in a 10-fold increase in IC50, thereby
supporting the aforementioned structure–function
hypothesis (Table 1). Removal of the C2-carboxylate (2)
verifed the necessity for the C2,C3-diacid functionality,
as the potency fell by >200-fold. Clearly, disruption of
salt bridges formed with one or more of the proximal
arginine residues in 3AMV imparted grave
consequences on GP inhibition.
C5-carboxamide inhibitors (16–19) were synthesized to
examine the SAR of the hydrophobic pocket occupied
by the isopropyl ester in 3AMV. The syntheses of 16–19
followed from tert-butyl ester deprotection of 10 to
afford an intermediate C5-acid (Scheme 2). Amide for-
mation under standard conditions, followed by diester
hydrolysis yielded the corresponding inhibitors.
Unfortunately, carboxamide replacement of the C5-
ester yielded very poor GPa inhibitors (Table 2). Some-
what surprising was the potency loss for amide 16 com-
pared to isosteric rac-U6751 (>200-fold), thereby
suggesting that the isopropyl amide unable to adopt a
productive conformation.
Regarding the N1-substituent, the N1-desalkyl inhibitor
(3) was significantly less active, and although the physical
underpinnings for this effect were unclear, productive van
der Waals interactions between the ethyl group and prox-
imal GPa residues must clearly exist. N1-trifluoroethyl
In order to expand on the SAR at N1, several analogues
of N1-benzyl inhibitor 5 were made utilizing the

A. K. Ogawa et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3405–3408
3407
Table 3. In vitro liver enzyme, muscle enzyme and hepatocyte inhib-
iton assay data for compounds 20–29.10,11
Compd
R
HLGPa
IC50 (nM)
HMGPa
IC50 (nM)
Cell
EC50 (mM)
5
H
2-Cl
3-Cl
4-Cl
11
79
8
26
178
29
1.13
3.71
4.01
0.48
6.19
7.13
6.71
1.06
2.22
0.44
0.27
20
21
22
23
24
25
26
27
28
29
2
6
2,3-Cl2
2,4-Cl2
2,5-Cl2
3,4-Cl2
3,5-Cl2
3-NO2
3,4-(OMe)2
109
104
65
13
15
5
233
185
91
41
37
Scheme 2. Reagents and conditions: (i) 140 ꢀC, neat; (ii) Pd(PPh3)4,
dimedone; (iii) R-NH2, EDC, HOBt; (iv) NaOH, (1:1) dioxane/H2O,
50 ꢀC.
10
11
4
Table 2. In vitro liver and muscle enzyme inhibition assay data for
compounds 16–1910
A pharmacodynamic study treating db/db mice with 29
resulted in a dose dependent effect on plasma glucose
concentration that was clearly evident at six h post-
dose13 (Fig. 2). In addition, a 100 mpk dose of 29 yiel-
ded glucose levels statistically similar to lean animals.
Of particular interest was the observed efficacy of 29
despite its sub-optimal pharmacokinetic profile.14
Rough calculations involving cmax and t1/2 combined
with the observed EC50 for 29 do not predict blood
levels commensurate with the observed efficacy at 6 h.
Several potential explanations exist, in addition to the
potential species difference between the hepatocyte
assay and PK/PD models, which will be discussed in a
subsequent communication. Initial tissue analysis indi-
cated that glycogen accumulation occurred exclusively
in the liver and not in peripheral skeletal muscle (data
not shown), the implications of which in relation to
McArdle’s disease15 will be discussed in the aforemen-
tioned future publication. This represents the first
Compd
R
HLGPa
IC50 (mM)
HMGPa
IC50 (mM)
16
17
18
19
iPrNH–
EtNH–
iBuNH–
BnNH–
9.10
12.4
7.88
12.9
19.9
8.28
10.61
10.90
Hantzsch chemistry outlined in Scheme 1. A compar-
ison of mono-substituted chlorobenzyl inhibitors (20–
22) revealed a general preference for meta- and para-
phenyl ring substitution over ortho substitution. The
preference against ortho subsituents was corroborated
by dichlorobenzyl inhibitors 23–25, as the addition of
an o-Cl substituent afforded decreased potency relative
to the parent mono-substituted benzyl inhibitors, 21–22.
Assay of this inhibitor series in rat hepatocytes sup-
ported the general trend established by the in vitro
assay. Noteworthy was the relative activity of p-Cl-ben-
zyl substituted inhibitor 22 and meta-substituted inhib-
itor 21, in which presently unidentified factors related to
the primary hepatocytes differentially impacted inhibitor
potency in a cell assay. A second m-Cl substituent
blunted this effect (27), yielding a 2-fold improvement in
cell potency despite a similar 2-fold reduction in enzyme
potency. para Substitution was still preferred as seen
from a comparison of equipotent GPa inhibitors 26 and
27. Lastly, a comparison of 3,4-Cl2-inhibitor, 26, with
the 3,4-dimethoxybenzyl inhibitor, 29,12 indicated a
preference for the latter substituent that was exception-
ally potent in the cell assay, as well (Table 3).
Figure 2. Blood glucose effect of compound 29 in db/db mice. x-Axis
represents time post-dose or vehicle administration. Each datapoint
represents an n=7.

3408
A. K. Ogawa et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3405–3408
demonstrated glucose normalization by this inhibitor
class in a relevant diabetic animal model.
6. (a) Goldmann, S.; Ahr, H.-J.; Puls, W.; Bischoff, H.; Pet-
zinna, D.; Schlossman, K.; Bender, J. US Patent 4,786,641,
1988. (b) Oikonomakos, N. G.; Tsitsanou, K. E.; Zographos,
S. E.; Skamnaki, V. T.; Goldmann, S.; Bischoff, H. Protein
Sci. 1999, 8, 1930. (c) Bergans, N.; Stalmans, W.; Goldmann,
S.; Vanstapel, F. Diabetes 2000, 49, 1419.
7. Zographos, S. E.; Oikonomakos, N. G.; Tsitsanou, K. E.;
Leonidas, D. D.; Chrysina, E. D.; Skamnaki, V. T.; Bischoff,
H.; Goldmann, S.; Watson, K. A.; Johnson, L. N. Structure
1997, 5, 1413.
In conclusion, a series of dicarboxylate GP inhibitors
provided SAR for the central dihydropyridine scaffold,
which led to the synthesis of compound 29. Oral
administration of 29 to db/db mice led to glucose nor-
malization not previously reported in a diabetic model.
8. (a) Sato, Y. US Patent 4,145,432, 1979. (b) Sato, Y. US
Patent 4,284,634, 1981. (c) Sato, Y.; Ichihashi, M.; Okumura,
K. Chem. Pharm. Bull. 1992, 40, 912.
9. Treatment of the concentrated diacid compound with
2 equiv NaOH(aq) precluded formation of the maleic anhy-
dride.
10. IC50 values derived from inhibiting phosphorolysis of gly-
cogen using recombinant human liver or muscle enzyme. The
standard error for this assay (15%) arose from titration of an
internal standard. Glucose-1-phosphate production was mon-
itored via an enzymatic assay involving phosphoglucomutase/
glucose-6-phosphate dehydrogenase-mediated NADH pro-
duction (ex. 340 nm, em. 465 nm).
Acknowledgements
The authors are grateful to Dr. David E. Moller for
helpful discussions. The authors also wish to thank
Sherrie McCormick for formulation and mass spectral
analysis, and to Judith Fenyk-Melody and Xiaolan
Shen for the in vivo phase of rodent pharmacokinetics.
11. Inhibition of glucagon-stimulated glycogenolysis in pri-
mary rat hepatocytes. The standard error for this assay (15%)
derived from an internal standard. Inhibition of glucose pro-
duction determined by digesting remaining glycogen and
quantifying by monitoring glucose dehydrogenase-mediated
NADPH production (ex. 340 nm, em. 465 nm).
References and Notes
1. Fact Sheet No. 138. World Health Organization, Geneva,
2002.
2. (a) McCormack, J. G.; Westergaard, N.; Kristiansen, M.;
Brand, C. L.; Lau, J. Curr. Pharm. Des. 2001, 7, 1457. (b)
Consoli, A. Diabetes Care 1992, 15, 430.
3. Magnusson, I.; Rothman, D. L.; Katz, L. D.; Shulman,
R. G.; Shulman, G. I. J. Clin. Invest. 1992, 90, 1323.
4. Hellerstein, M. K.; Neese, R. A.; Linfoot, P.; Christiansen,
M.; Turner, S.; Letsher, A. J. Clin. Invest. 1997, 100, 1305.
5. (a) Treadway, J. L.; Mendys, P.; Hoover, D. J. Expert Opin.
Invest. Drugs 2001, 10, 439. (b) Hoover, D. J.; Lefkowitz-
Snow, S.; Burgess-Henry, J. L.; Martin, W. H.; Armento, S. J.;
Stock, I. A.; McPherson, R. K.; Gereroux, P. E.; Gibbs, E. M.;
Treadway, J. L. J. Med. Chem. 1998, 41, 2934. (c) Martin,
W. H.; Hoover, D. J.; Armento, S. J.; Stock, I. A.; McPher-
son, R. K.; Danley, D. E.; Stevenson, R. W.; Barrett, E. J.;
Treadway, J. L. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 1776.
1
12. Characterization for compound 29: H NMR (500 MHz,
CD3OD) d 7.15 (3H, m), 6.96 (1H, ddd, J=7.6, 7.5, 1.5 Hz),
6.92 (1H, m), 6.85 (2H, m), 5.50 (1H, s), 4.87 (2H, m), 4.78
(1H, d, J=16.0 Hz), 3.72 (3H, s), 3.65 (3H, s), 2.36 (3H, s),
1.21 (3H, d, J=6 Hz), 1.05 (3H, d, J=6.5 Hz). LC/MS ESI
(C27H28ClNNaO8) calcd for MNa+: 552.14; found: 552.1.
13. Glucose levels determined at indicated times post-dosing
of either compound 29 or vehicle. Each curve represents a
single group of n=7 animals, and timepoints therein an aver-
age of the group.
14. Pharmacokinetic data from C57BL/6J mice (3/pt), 1 mpk
po: t1/2=1.4 h, Cmax=12 nM, %F=4.5.
15. Stryer, L. Biochemistry; W.H. Freeman and Co.: New
York, 1995; p 598.