Anthranilimide based glycogen phosphorylase inhibitors for the treatment of type 2 diabetes. Part 3: X-ray crystallographic characterization, core and urea optimization and in vivo efficacy

Article abstract of DOI:10.1016/j.bmcl.2008.12.085

Key binding interactions of the anthranilimide based glycogen phosphorylase a (GPa) inhibitor 2 from X-ray crystallography studies are described. This series of compounds bind to the AMP site of GP. Using the binding information the core and the phenyl ur

Full text of DOI:10.1016/j.bmcl.2008.12.085

Bioorganic & Medicinal Chemistry Letters 19 (2009) 1177–1182
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Anthranilimide based glycogen phosphorylase inhibitors for the treatment
of type 2 diabetes. Part 3: X-ray crystallographic characterization, core
and urea optimization and in vivo efficacy
*
Stephen A. Thomson , Pierette Banker, D. Mark Bickett, Joyce A. Boucheron, H. Luke Carter,
Daphne C. Clancy, Joel P. Cooper, Scott H. Dickerson, Dulce M. Garrido, Robert T. Nolte, Andrew J. Peat,
Lauren R. Sheckler, Steven M. Sparks, Francis X. Tavares, Liping Wang, Tony Y. Wang, James E. Weiel
Metabolic Center for Excellence in Drug Discovery, GlaxoSmithKline, 5 Moore Drive, PO Box 13398, Research Triangle Park, NC 27705, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Key binding interactions of the anthranilimide based glycogen phosphorylase a (GPa) inhibitor 2 from
X-ray crystallography studies are described. This series of compounds bind to the AMP site of GP. Using
the binding information the core and the phenyl urea moieties were optimized. This work culminated in
the identification of compounds with single nanomolar potency as well as in vivo efficacy in a diabetic
model.
Received 9 October 2008
Revised 17 December 2008
Accepted 19 December 2008
Available online 25 December 2008
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
Glycogen
Phosphorylase
Inhibitor
Anthranilimide
Glucagon
Challenge
Antidiabetic
Type 2 diabetes has reached near epidemic proportions in most
of the developed world and results in acute and chronic complica-
tions, disability, and death. The degree of this problem is exempli-
fied in the fact that type 2 diabetes is now becoming prevalent in
children and adolescents.1 An increase in the prevalence and de-
gree of obesity in children is a likely driver of the increase in type
2 diabetes in this population.2 Although there are several treat-
ment options for type 2 diabetes, long term efficacy of oral anti-
diabetic agents has been limited. It has recently been reported that
21% of patients will fail on metformin treatment within 5 years.3
Therefore new treatment options are needed that can either be
used as mono-therapy or in combination with other agents. Exces-
sive hepatic glucose production is a major component of hypergly-
cemia and consists of both glycogenolysis and gluconeogenesis. In
the postprandial state glycogenolysis was increased more than
twofold in diabetic subjects compared to non-diabetics. In addition
diabetics have a 55% reduction in liver glycogen content compared
to normal subjects.4 The enzyme glycogen phosphorylase a (GPa) is
responsible for the release of glucose-1-phosphate from glycogen
and is the rate determining step in glycogenolysis. GPa has there-
fore become a target for glucose lowering in type 2 diabetes.
Although several groups have reported small molecule inhibitors
of GPa, definitive clinical results have not been communicated.5
In previous papers, we have outlined some of the properties and
SAR of a novel anthranilimide series of GPa inhibitors exemplified
by compound 1.6
OH
HN
O
O
Cl
NH
O
N
H
1
IC50 = 73 nM
IC50 (cell) =600 nM
CYP2C9 IC50 = 1000 nM
* Corresponding author.
E-mail address: stephen.a.thomson@gsk.com (S.A. Thomson).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.12.085

1178
S. A. Thomson et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1177–1182
We have reported on the SAR of the amino-acid head group and
between the T (inactive) and R (active) state. Therefore binding at
this site could be envisioned to block the shift between the T and R
states.
how modification at this position can affect the potency of CYP450
inhibition.7 Overall low nanomolar potency for GPa inhibition and
decreased P450 inhibition can be obtained by replacement of the
cyclohexyl containing amino-acid head group with a t-butyl thre-
onine head group. In addition replacement of the 2-methyl-6-
chloro phenyl group of 1 with a 2,4,6-trimethylphenyl group im-
proved potency as well, giving compound 2.
The molecule forms a number of important hydrogen bonds to
the protein as shown in Figure 2. The carboxylate of the amino-acid
moiety forms a bidentate interaction with Arg310 of the phos-
phate-recognition site. The amide carbonyl forms a hydrogen bond
with the NH2 of Gln71. The urea moiety forms two hydrogen
bonds with the main-chain of the symmetry-related subunit (of
the homo-dimer); (1) the urea NH with the carbonyl of Val400
and (2) the urea carbonyl with the NH of Asp420 (where the prime
refers to residues from the symmetry-related subunit). These later
interactions are similar to those observed for the acyl urea inhibi-
tors previously reported.11 More than a dozen other residues con-
tribute to binding through van der Waals interactions. The
lipophilic regions of the amino-acid moiety are surrounded by
Phe196, the alkyl chain of Arg309 and Ala313. The polar end of
Arg309 is flipped away from the binding site and is not participat-
ing in hydrogen bonding interactions with the ligand as is often ob-
served in other crystal structures.12 The naphthalene moiety is
surrounded by residues Asp420, Asn440 and Val450 as well as
Ile68, Gln71, Gln72 and Tyr75. The urea phenyl ring is situated
the deepest in the binding site and forms interactions with
Trp67, Gln71, Thr240, Arg242, Asp227, Phe196, Lys41 and
Arg193. Arg193 appears to be forming pi-stacking interactions
with the ring and is in a conformation relative to the inhibitor that
is different than that observed for the other reported AMP site
binders.11,12
Examination of the ligand (compound 2) in the crystal structure
illustrates two interesting features of the binding confirmation.
First is the potential for an internal hydrogen bond between the
internal NH of the urea with the carbonyl of the amide. Second is
the s-cis confirmation of the 2,4,6-dimethylphenylurea. In order
to understand the low energy confirmation of the ligand we stud-
ied the model system shown in Figure 3 (shown in the s-cis config-
uration) with molecular mechanics (MacroModel, Schrodinger LLC)
using the OPLS2005 force field and a constant dielectric of 4.0. The
low energy conformation of this model includes an intra-molecular
hydrogen bond between the amide carbonyl and the internal NH of
the urea moiety. While the difference in energy between the s-
trans and s-cis conformations of the un-substituted phenylurea
analog was ꢀ2.7 kcal/mol, the difference for the 2,6-dimethylphe-
nylurea analog was ꢀ0.5 kcal/mol. This significant difference is
predicted to allow the 2,6-dimethyl substituted analogs to readily
adopt the s-cis conformation and to bind to the protein with a min-
imal amount of strain. These features aid in the understanding of
O
OH
HN
O
O
NH
O
N
H
2
IC50 = 7 nM
IC50 (cell) =139 nM
CYP2C9 IC50 = 10000 nM
One concern with compound 2 is the potential to form reactive
metabolites via the known pathway of epoxidation across the 5,6
or 7,8 positions of the naphthalene ring system.8 Indeed incubation
of compound 2 with human liver microsomes in the presence of
glutathione and analysis by LC/MS/MS gave ions consistent with
such a process occurring.9 In light of this result and the desire to
further understand the SAR of this series of compounds, we inves-
tigated replacements to the naphthalene ring system and explored
the SAR of the phenyl urea. To aid in this effort we were able to ob-
tain an X-ray crystal structure of compound 2 bound to GPa. Herein
we report the general binding interactions observed in the crystal
structure, additional SAR of this series, as well as in vivo efficacy of
these GPa inhibitors.
The X-ray crystal structure of compound 2 bound to GPa is
shown in Figure 1 (view from solvent).10 Compound 2 binds to
GPa at the AMP regulatory site. Binding at this site can result in en-
zyme inhibition either by competing with the allosteric activator,
AMP, or by stabilizing the inactive T conformation of the enzyme.
The AMP binding site is partially solvent exposed at the dimer
interface and lies along one of the protein’s twofold axes of sym-
metry. This axis is the one on which the protein pivots as it shifts
Figure 2. X-ray crystal structure of compound 2 bound to GP, showing key binding
Figure 1. X-ray crystal structure of compound 2 bound to GP at the AMP site.
interactions. Prime residues reside on the symmetry-related subunit shown in blue.

S. A. Thomson et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1177–1182
1179
CO2Me
NO2
b, c, d
CO2Me
a
R
NO2
Cl
O
O
CO2H
HN
O
e, f
HN
CO2Me
O
NH
NH2
R
R
O
NH
Figure 3. Model system used to study the effects of 2,6-dimethyl substitution on
the stability of the s-cis conformation (shown). The s-cis configuration orients the
urea carbonyl and NH toward the main-chain carbonyl and NH of Val400 and
Asp420.
Scheme 1. General syntheses of compounds 5–11. Reagents and conditions: (a) R-
PhB(OH)2, Pd(Cy3P)2Cl2, CsF, CH3CN, H2O, 150 °C, 5 min, microwave; (b) LiOH, H2O,
MeOH, THF; (c) methyl O-(1,1-dimethylethyl)-L-threoninate hydrochloride, HATU,
iPrNEt2, DMF, rt; (d) H2, 5% Pd/C, EtOH; (e) 2,4,6-Me3PhNCO, pyr; (f) LiOH, H2O,
MeOH, THF.
previously reported SAR for this series of compounds in which the
2,6-dimethylphenyl urea was greater than 100 times more potent
than the 2-methylphenyl urea.6
From the crystal structure one can see that the naphthalene
group binds in a narrow lipophilic channel, nearly ending at the
solvent front. This interaction contributes significantly to the activ-
ity of this series of compounds since replacement of the naphthyl
group with a phenyl results in a large loss of activity in at the en-
zyme assay (see Table 1).13 From this structure one could envisage
replacing the naphthyl group with a bi-aryl group, which would al-
low for easier introduction of diversity in this region of the
molecule.
The general synthesis of these bi-aryl anthranilimides is shown
in Scheme 1.14 This synthetic pathway allows for easy variation of
the outboard phenyl ring by employing various phenylboronic
acids. As shown in Table 1 the naphthalene group can be replaced
with substituted bi-phenyl groups with little change in activity.
Unlike compound 2, incubation of compound 9 with human liver
microsomes in the presence of glutathione showed no evidence
of formation of reactive metabolites via oxidation and addition to
the bi-aryl ring system.9 In addition to the enzyme inhibition assay
the compounds were evaluated in a cell based assay to inhibit for-
Figure 4. Crystal structure of compound 12, showing the view of the lipophilic
pocket off the 4 position on the phenyl urea.
skolin stimulated glycogenolysis.15 Although potency in the cell as-
say does not track directly with enzyme inhibition potency, in
general the more potent a compound is at the enzyme the more
potent it will be in the cell assay. It is important to note that sev-
eral of these compounds are reaching the lower limit of the en-
zyme assay and could be more potent than reported.13
Table 1
SAR of naphthyl replacements
O
We next turned to the SAR of the phenyl urea. In our previous
paper, we reported improved potency of the 2,4,6-trimethylphenyl
versus 2,6-dimethylphenyl urea.7 The reason for this improvement
is clear upon closer examination of the crystal structure (see Fig. 4).
Off of the 4 position of the phenyl urea there appears to be a nar-
row pocket defined by residues Tyr155, Thr240, Arg242 and
Arg310, which can accommodate the 4-methyl group. We wished
to explore the boundaries of this pocket and therefore prepared
the compounds listed in Table 2. Here we selected the less active
4-flourophenyl core as the test case so that the SAR of the 4 posi-
tion of the phenyl urea ring can be delineated without reaching
the lower limit of the assay. The general route to these compounds
is shown in Scheme 2, in which the key bromide intermediate can
be functionalized to give various sized alkyl groups via a palladium
mediated coupling reaction and subsequent hydrogenation. For
compound 16 the cyclopropyl ring was constructed via palladium
mediated addition of diazomethane to the olefin of the allyl
intermediate.16
OH
HN
O
O
R
NH
O
N
H
a
Compound
R
GPa inhibition IC50 (nM) GPa (cell)IC50 (nM)
3
4
5
6
7
8
9
10
11
H
F
542 (96)
506 (25)
10 (7)
3 (2)
1680
1820
Phenyl
126 (49)
34 (14)
109
353 (56)
104 (19)
88 (21)
171 (24)
4-MeOphenyl
3-MeOphenyl
2-MeOphenyl
3-Fluorophenyl
4-Fluorophenyl
8 (3)
28 (16)
15 (4)
7 (3)
3,4-Difluorophenyl 13 (6)
a
Values are means of three experiments, standard error is given in parentheses.

1180
S. A. Thomson et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1177–1182
Table 2
Table 3
SAR of 4-alkyl phenyl urea
SAR of 4-alkyl phenyl urea with t-butyl threonine head group
O
O
OH
OH
HN
HN
O
O
O
F
NH
NH
F
O
N
R
O
N
R
H
H
Compound
R
GPa inhibition IC50a (nM) GPa (cell) IC50 (nM)
Compound
R
GPa inhibition IC50a (nM) GPa (cell) IC50 (nM)
12
13
14
15
16
17
H
Me
Et
1067 (230)
143 (23)
87 (23)
8260
1026 (244)
513
229
384
4
18
19
Me
n-Pr
506 (25)
10 (2)
1820
147 (42)
133 (28)
Cyclopropylmethyl 4 (1)
n-Pr
19 (10)
a
Values are means of three experiments, standard error is given in parentheses.
Cyclopropylmethyl 13 (7)
n-Butyl 49 (21)
716
a
Values are means of three experiments, standard error is given in parentheses.
From Table 1 we know that replacement of the fluoro group of
compound 4 with a phenyl group (compound 5) results in a large
increase in potency. Therefore we wished to prepare bi-phenyl
compounds like 5–11 which also contain n-propyl and cyclopro-
pylmethyl substitutions at the 4 position of the phenyl urea (com-
pounds of Table 4). The synthesis of these compounds utilizes
reactions shown in Schemes 1 and 2. Using a reaction sequence
equivalent to that shown in Scheme 1, but employing 5-bromo-
2-isocyanato-1,3-dimethylbenzene instead of the 2-isocyanato-
1,3,5-trimethylbenzene gives the key bromo intermediate which
can then be functionalized at the 4 position of the urea in a manner
equivalent to that outlined in Scheme 2. As shown in Table 4 this
gives compounds that are all within the limit of the enzyme inhi-
bition assay,13 but somewhat surprising is that these compounds
do not show an increase in potency in the cell based assay com-
pared to compounds 5–11. It is possible that a lower limit in po-
tency exist in this cellular assay due to cellular enzyme levels,
although compound 6 (Table 1) would suggest it to be lower than
50 nM.
HN
CO2Me
b, c
CO2H
a
O
H2N
CO2Me
+
NO2
F
NO2
F
I
HN
CO2Me
d, e, g
14
15
16
17
O
h, e, g
h, i, g
j, e, g
F
NH
O
NH
To further understand and differentiate the activity of these
compounds we examined their ability to inhibit the glucose rise in-
duced by iv administered glucagon in an in vivo rat model.17 Com-
pounds were dosed orally 2 h prior to the glucagon challenge and
the data is reported as the % reduction (%R) in the glucose AUC
Br
Scheme 2. General syntheses of compounds 14–17. Reagents and conditions: (a)
HATU, iPrNEt2, DMF, rt, (b) H2, 10% Pd/C, ethyl acetate; (c) 2,6-Me2-4-BrPhNCO, pyr,
rt; (d) vinyltributyltin, Pd(Ph3P)4, CH3CN, 150 °C, 30 min, microwave; (e) H2, Pd/C,
ethyl acetate; (g) LiOH, H2O, MeOH, THF; (h) allyltributyltin, Pd(Ph3P)4, CH3CN,
150 C, 30 min, microwave; (i) CH2N2, Pd(acac)2, Et2O, CH2Cl2, ꢁ5 °C; (j) 2-[(1E)-1-
buten-1-yl]-1,3,2-benzodioxaborole, Pd(Cy3P)2Cl2, CsF, CH3CN, H2O, 150 C, 7 min,
microwave.
Table 4
Biphenyl compounds with 4-alkyl phenyl urea substitutions
O
OH
HN
As can be seen in Table 2 increasing the size of R from H to n-
propyl results in an improvement in potency both at the enzyme
as well as in the cell based assay. In this case there is a good corre-
lation between enzyme potency and cell potency. When R is n-bu-
tyl there appears to be some loss of activity and this maybe
approaching the limit of the pocket size. Incorporation of n-propyl
and cyclopropylmethyl at the 4 position of the phenyl urea was
also carried out with the t-butyl threonine head group in place as
shown in Table 3 (synthetic scheme equivalent to Scheme 2 in
which t-butyl threonine is used). Again in this series increased size
at the 4 position of the urea results in improved potency in both
the enzyme and cell assays. As reported earlier7 the t-butyl threo-
nine head group decreases CYP2C9 inhibition and this trend holds
true when comparing the CYP2C9 activity of compounds 16 and 19
(data not shown).
O
O
NH
R1
O
N
R2
H
a
Compound R1
R2
GPa inhibition IC50
(nM)
GPa (cell) IC50
(nM)
20
21
3,4-F
4-
n-Pr
n-Pr
7 (10)
4 (1)
181 (27)
135 (17)
OMe
3-F
3-F
22
23
n-Pr
4 (1)
156 (15)
145 (18)
Cyclopropylmethyl 3 (1)
a
Values are means of three experiments, standard error is given in parentheses.

S. A. Thomson et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1177–1182
1181
Table 5
pound in this series can acutely lower glucose levels in a diabetic
mouse model (ob/ob). In future publications, we will delineate
the PK properties of these compounds as well as describe addi-
tional in vivo efficacy studies of this series of compounds.
Inhibition of glucagon induced glucose AUC in rats (n = 4)
a
Compound GPa inhibition IC50
(nM)
GPa (cell) IC50
(nM)
%R @ 5 mg/
%R @ 2 mg/
kgb
kgb
2
9
19
23
7 (2)
15 (4)
4 (1)
3 (1)
139 (41)
104 (19)
133 (28)
145 (18)
25
32
36
65
ND
10
22
62
References and notes
1. (a) Reinehr, T.; Schober, E.; Roth, C. L.; Wiegand, S.; Holl, R. Horm. Res. 2008, 69,
107; (b) Weiss, R.; Taksali, S. E.; Caprio, S. Curr. Diab. Rep. 2006, 6, 182; (c)
Soltesz, G. Diabetes Res. Clin. Pract. 2006, 74, S9–S11.
2. (a) Ebbeling, C. B.; Pawlak, D. B.; Ludwig, D. S. Lancet 2002, 360, 473; (b)
Hannon, T. S; Rao, G.; Arslanian, S. A. Pediatrics 2005, 116, 473.
3. (a) Kahn, S. E.; Haffner, S. M.; Heise, M. A.; Herman, W. H.; Holman, R. R.; Jones,
N. P.; Kravitz, B. G.; Lachin, J. M.; O’Neill, C.; Zinman, B.; Viberti, G. N. Eng. J. Med.
2006, 355, 2427; (b) Monami, M.; Lamanna, C.; Marchionni, N.; Mannucci, E.
Diabetes Res. Clin. Pract. 2008, 79, 196.
Compounds were dose orally 2 h prior to the glucagon challenge.
a
Values are means of three experiments, standard error is given in parentheses.
ND = not determined.
b
p < 0.005 for all.
80
60
40
20
4. (a) Woerle, H. J.; Szoke, E.; Meyer, C.; Dostou, J. M.; Wittlin, S. D.; Gosmanov, N.
R.; Welle, S. L.; Gerich, J. E. Am. J. Physiol. 2006, 290, E67; (b) Magnusson, I.;
Rothman, D. L.; Katz, L. D.; Shulman, R. G.; Shulman, G. I. J. Clin. Invest. 1992, 90,
1323.
5. (a) Chen, L.; Li, H.; Liu, J.; Zhang, L.; Liu, H.; Jiang, H. Bioorg. Med. Chem. 2007, 15,
6763; (b) Birch, A. M.; Kenny, P. W.; Oikonomakos, N. G.; Otterbein, L.;
Schofield, P.; Whittamore, P. R. O.; Whalley, D. P. Bioorg. Med. Chem. Lett. 2007,
17, 394; (c) Henke, B. R.; Sparks, S. M. Mini-Rev. Med. Chem. 2006, 6, 845; (d) Li,
Y. H.; Coppo, F. T.; Evans, K. A.; Graybill, T. L.; Patel, M.; Gale, J.; Li, H.; Tavares,
F.; Thomson, S. A. Bioorg. Med. Chem. Lett. 2006, 16, 5892. and references
therein.
23
0
6. Evans, K. A.; Li, Y. H.; Coppo, F. T.; Graybill, T. L.; Cichy-Knight, M.; Patel, M.;
Gale, J.; Li, H.; Thrall, S. H.; Tew, D.; Tavares, F.; Thomson, S. A.; Weiel, J. E.;
Boucheron, J. A.; Clancy, D. C.; Epperly, A. H.; Golden, P. L. Bioorg. Med. Chem.
Lett. 2008, 18, 4068.
Vehicle
-20
7. (a) Sparks, S. M. et al Bioorg. Med. Chem. Lett. 2009, 19, 976; (b) Sparks, S. M.
et al Bioorg. Med. Chem. Lett. 2009, 19, 981.
8. (a) Shultz, M. A.; Choudary, P. V.; Buckpitt, A. R. J. Pharmacol. Exp. Ther. 1999,
290, 281; (b) Buckpitt, A. R.; Warren, D. L. J. Pharmacol. Exp. Ther. 1983, 225,
8.
-40
-60
9. Castro-Perez, J.; Plumb, R.; Liang, L.; Yang, E. Rapid Commun. Mass Spectrom.
2005, 19, 798.
-80
10. GP protein was engineered with an N-term fusion to a hexa his-tag followed by
a TEV protease site and expressed in Escherichia coli. Cells were lysed and
activated to immobilized metal affinity chromatography (IMAC). Fractions
containing GP were pooled, diluted and subjected to anion exchange
chromatography (Q Sepharose HP). Fractions containing GP were pooled and
the enzyme was activated. TEV protease was used to remove the tag and the
activated cleaved GP was activated to another round of IMAC. The GP in the
IMAC flow through was concentrated for crystallography. Crystals were
prepared similar to the method described by Rath et al. for the active
conformation of glycogen phosphorylase (see Rath, V. L.; Ammirati, M.;
LeMotte, P. K.; Fennel, K. F.; Mansour, M. N.; Denley, D. E.; Hynes, T. R.;
Schulte, G. K.; Wasilko, D. J.; Pandit, J. Mol. Cell 2000, 6, 139). The crystallization
method was modified to include the addition of 5.0 mM caffeine and 0.1 mM
ligand to the protein solution 1 h prior to setting up the crystals. Crystals were
frozen in liquid nitrogen directly from the drop and X-ray diffraction data was
collected at 100 K using an ADSC Q210 detector on the Advanced Photon
Source Beam line 17ID at Argonne National Labs. The structure was solved by
molecular replacement using the program Amore from PDB entry 1em6.
Refinement was carried out using Refmac and other CCP4 tools and the final
-100
-120
Treatment
Figure 5. Effect on blood glucose in ob/ob mice by compound 23 dosed at 15 mg/kg
compared to vehicle 3 h postdosing (n = 10).
for 20 min post glucagon treatment. As can be seen in Table 5 these
compounds are able to inhibit the effect of glucagon, which is pre-
sumed to be via the inhibition of GP in the liver. It is of interest that
although these compounds do not differ appreciably in their
in vitro cell activity compound 23 which we would predict to be
the most active based on the herein reported SAR is significantly
more active in vivo. This could be due to a variety of factors, such
as differences in pharmacokinetics,18 tissue distribution (liver
exposure) and enzyme inhibition in the in vivo system.
To further investigate the anti-diabetic effect of this series of
compounds we evaluated the glucose lowering ability of com-
pound 23 in a diabetic mouse model (ob/ob).19 In this model fed
ob/ob mice were orally dosed either with compound 23 (15 mg/
kg) or vehicle and blood glucose was measured before dosing
and 3 h later. As shown in Figure 5 compound 23 gave a robust re-
sponse, lowering glucose levels 103 + 12 mg/dL whereas in the
vehicle treated mice the glucose levels increase 47 + 17 mg/dL over
the 3 h period (p < 0.000001).
model was built using the programs
O and Coot.Compound 2 cocrystals
diffracted to 1.80 Å in resolution with an R-merge of 8.7%. The final model was
refined to an R-factor of 17.4% with a Free R-factor of 19.8%. Compound 12
cocrystals diffracted to 1.90 Å in resolution with an R-merge of 7.8%. The final
model was refined to an R-factor of 15.3%, with a Free R-factor of 18.6%.
Coordinates and statistics are available from the PDB using accession code
3DDS and 3DDW, respectively.
11. Klabunde, T.; Wendt, K. U.; Kadereit, D.; Brachvogel, V.; Burger, H.; Herling,
A. W.; Oikonomakos, N. G.; Kosmopoulou, M. N.; Schmoll, D.; Sarubbi, E.;
Von Roedern, E.; Schoenafinger, K.; Defossa, E. J. Med. Chem. 2005, 48,
6178.
12. Kristiansen, M.; Andersen, B.; Iversen, L. F.; Westergaard, N. J. Med. Chem. 2004,
47, 3537.
13. Inhibitors were tested for human liver glycogen phosphorylase enzymatic
activity using a coupled kinetic fluorescence intensity assay. The change in
fluorescence due to product formation was measured on a fluorescence plate
reader (Molecular Devices SpectraMax M2) using a excitation wavelength of
560 nm and an emission wavelength of 590 nm. The hGPa enzyme IC50 values
given in Tables 1–5 are average values of at least two replicates where
standard deviations are noted, and were measured in the presence of glucose
(10 mM). Due to the specific activity of the enzyme, a concentration of 10–
15 nM glycogen phosphorylase is used in the assay. Therefore, inhibitors with
IC50 determined to be < approximately 5 nM (Kd < enzyme concentration)
cannot be accurately evaluated in this assay format. Inhibitors falling into this
category may have IC50 significantly lower than the estimate. For further
experimental details, see Evans, K. A.; Cichy-Knight, M.; Coppo, F. T.; Dwornik,
K. A.; Gale, J. P.; Garrido, D. M.; Li, Y. H.; Patel, M.; Tavares, F. X.; Thomson, S. A.;
In summary, we have described the key binding interactions
and conformation of this series of anthranilamide based GPa inhib-
itors and by using the knowledge gained from the crystal structure
we have expanded the SAR to deliver extremely potent molecules.
Combining the best features of the SAR has yielded inhibitors
which are able to inhibit GP in vivo at low doses as measured by
a glucagon challenge test. In addition we have shown that a com-

1182
S. A. Thomson et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1177–1182
Dickerson, S. H.; Peat, A. J.; Sparks, S. M.; Banker, P.; Cooper, J. P. WO 2006/
052722 A1.
drug (0.1–15 ml/kg) 2 h prior to the glucagon challenge (GC). A time zero blood
sample (0.4 ml) was collected for determination of glucose and the rats were
14. All novel compounds were characterized by NMR and LCMS, and gave
satisfactory results in agreement with the proposed structure.
dosed via the jugular vein with Sandostatin, 0.5 mg/kg, and glucagon, 10 lg/kg.
Blood samples were collected after 10 and 20 min. Whole blood was allowed to
sit at room temperature for 20 min and then centrifuged (3000g) to obtain
serum. Serum levels of glucose were determined using an Olympus AU640TM
clinical chemistry immuno-analyzer.
15. This full curve assay was designed to detect the inhibition of glycogenolysis
(glycogen breakdown) by test compounds. On the day before the assay the
glycogen in HepG2 cells is pre-labeled by overnight inclusion of 14C-glucose in
the culture medium. To begin the assay, the cells are treated with test
compounds, and glycogenolysis is stimulated by forskolin treatment for
60 min. The cells are then lysed and the radiolabeled glycogen in the cells is
18. For compound 23, rat PK (IV dosed at 1 mg/kg, PO dosed at 5 mg/kg);
Cl = 15 mL/min/kg, T1/2 = 2.3 h,%F = 27, PO Cmax = 536 ng/mL).
19. ob/ob mice, Strain B6.V-Lepob/J Stock No. 000632 (Drasher, 1955) (6–7-wk-old)
were housed 3–4/cage with free access to food and water for 1–2 wk. On the
day prior to analysis, the mice were sorted by blood glucose (tail snip,
glucometer) and body weight into treatment groups (N = 10 for drug, N = 20 for
vehicle). On the following day, the mice were moved to the study room at 6:30
am, and allowed to acclimate for ꢀ1 h. Basal fed glucose levels were measured
by tail-nick and glucometer read and the mice were dosed with 10 ml/kg
vehicle (5% DMSO: 30% Solutol HS15: 20% PEG400: 45% 25 mM N-
methylglucamine) or drug, at 8 am. Blood glucose levels were measured
again by tail-nick and glucometer at 11 am.
quantified. If
a test compound inhibits glycogenolysis, the radiolabeled
glycogen content of the cells will be greater than control (forskolin treated).
The hGPa(cell) IC50 values given in Tables 1–5 are average values of at least two
replicates where standard deviations are noted.
16. Ma, D.; Zhu, W. J. Org. Chem. 2001, 66, 348.
17. Jugular vein cannulated male CD rats (220–260 g) were received 1–2 days after
cannulation, housed individually with free access to food and water for 4 days
prior to the glucagon challenge (GC) studies. The rats were sorted by body
weight into treatment groups (N = 4–5). Rats were dosed orally with vehicle
(5% DMSO: 30% Solutol HS15: 20% PEG400: 45% 25 mM N-methylglucamine) or