Amino acid anthranilamide derivatives as a new class of glycogen phosphorylase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2008.05.102

A series of amino acid anthranilamide derivatives identified from a high-throughput screening campaign as novel, potent, and glucose-sensitive inhibitors of human liver glycogen phosphorylase a are described. A solid-phase synthesis using Wang resin was also developed which provided efficient access to a variety of analogues, and resulted in the identification of key structure-activity relationships, and the discovery of a potent exemplar (IC₅₀ = 80 nM). The SAR scope, synthetic strategy, and in vitro results for this series are presented herein.

Full text of DOI:10.1016/j.bmcl.2008.05.102

Bioorganic & Medicinal Chemistry Letters 18 (2008) 4068–4071
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Amino acid anthranilamide derivatives as a new class of glycogen
phosphorylase inhibitors
a
a
a
a
a
Karen A. Evans a, , Yue H. Li , Frank T. Coppo , Todd L. Graybill , Maria Cichy-Knight , Mehul Patel ,
Jennifer Gale a, Hu Li a, Sara H. Thrall a, David Tew a, Francis Tavares b, Stephen A. Thomson b,
James E. Weiel b, Joyce A. Boucheron b, Daphne C. Clancy b, Andrea H. Epperly b, Pamela L. Golden b
*
a Discovery Research, GlaxoSmithKline Pharmaceuticals, 1250 South Collegeville Road, PO Box 5089, Collegeville, PA 19426-0989, USA
b Metabolic and Viral Diseases Center of Excellence for Drug Discovery, GlaxoSmithKline Pharmaceuticals, Five Moore Drive, PO Box 13398,
Research Triangle Park, NC 27709-3398, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of amino acid anthranilamide derivatives identified from a high-throughput screening campaign
as novel, potent, and glucose-sensitive inhibitors of human liver glycogen phosphorylase a are described.
A solid-phase synthesis using Wang resin was also developed which provided efficient access to a variety
of analogues, and resulted in the identification of key structure–activity relationships, and the discovery
of a potent exemplar (IC50 = 80 nM). The SAR scope, synthetic strategy, and in vitro results for this series
are presented herein.
Received 18 April 2008
Revised 23 May 2008
Accepted 27 May 2008
Available online 12 June 2008
Keywords:
Ó 2008 Elsevier Ltd. All rights reserved.
Human liver glycogen phosphorylase
Glycogen phosphorylase inhibitors
Solid-phase synthesis
According to the International Diabetes Federation (IDF), more
than 230 million people, almost 6% of the world’s adult population,
now live with diabetes. The incidence of diabetes and associated
complications, such as obesity and heart disease, is escalating rap-
idly in various societies around the world, placing a significant eco-
nomic burden on healthcare resources. Despite the use of various
hypoglycemic agents with or without co-therapy with insulin, cur-
rent treatments often fail to achieve significant lowering of serum
glucose to a level where the severity of diabetic complications is
effectively reduced. Thus, there is a clear need for novel hypoglyce-
mic agents targeting the underlying mechanism(s) of elevated ser-
um glucose.
inhibitors of this enzyme.2 Herein, the discovery, synthesis, and
structure–activity relationships (SAR) of a series of amino acid
anthranilamide derivatives as novel glucose-sensitive GPa inhibi-
tors are presented.
O
H
N
HO
Cl
H
R1
O
HN
N
O
Studies have shown that excessive hepatic glucose production
(HGP) is a significant factor contributing to hyperglycemia in pa-
tients with type 2 diabetes (non-insulin dependent diabetes melli-
tus).1 Glucose is produced by both gluconeogenesis and
glycogenolysis, the release of glucose-1-phosphate from glycogen.
Since glycogenolysis is a major component of HGP and human liver
glycogen phosphorylase a (GPa) catalyzes this reaction as the rate-
limiting enzyme, it is thought that inhibiting GPa will limit glyco-
genolysis, reduce HGP, and thus lower blood glucose, thereby pro-
viding a potential new treatment for type 2 diabetes. There are
several examples in the literature of small-molecule allosteric
1a, R1 = CH2Cy
IC50 (+glu) = 2.2 uM
IC50 (-glu) = 51 uM
IC50 (cell) = 7.4 uM
1b, R1 = CH2CO2H
IC50 (+glu) = 0.39 uM
IC50 (-glu) = 1.6 uM
IC50 (cell) = 4.7 uM
Compounds 1a (IC50 = 2.2 lM) and 1b (IC50 = 0.39 lM) were
identified as part of a high-throughput screen of the GSK proprie-
tary compound collection. Both compounds, and compound 1b in
particular, were found to have good enzyme inhibition in the pres-
ence of glucose (+glu), but were observed to be less potent in the
absence of glucose (Àglu), suggesting glucose sensitivity for this
* Corresponding author. Tel.: +1 610 917 7764; fax: +1 610 917 7391.
E-mail address: karen.a.evans@gsk.com (K.A. Evans).
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.05.102

K. A. Evans et al. / Bioorg. Med. Chem. Lett. 18 (2008) 4068–4071
4069
series of compounds.3 In addition, both compounds were active in
Table 1
In vitro GPa inhibition: amino acid side chain variations
cells, demonstrating micromolar potencies. Significant metabolic
advantages should exist for glucose-sensitive GPa inhibitors which
act primarily at the high blood glucose levels typically seen in dia-
betic patients, and then lose potency as glucose levels fall in order
to avoid hypoglycemia.
Initial objectives for this chemical series were thus to increase
the potency, establish key structure–activity relationships, develop
efficient synthetic methods, and evaluate the pharmacokinetic pro-
file. Structural modifications of this template were primarily
achieved through solid-phase synthesis using IroriÒ AccuTag-
100TM (miniKan) technology with Wang resin-bound amino acids
(Scheme 1).
O
H
N
HO
Cl
H
R1
O
HN
N
O
1
Compound
R1
GPa IC50
,
l
M
GPa (cell) IC50, lM
Commercially available Fmoc-protected L- or D-amino acids (2)
1a-(S)
1c-(S)
1c-(R)
1d-(S)
1e-(2S,3S)
1f-(S)
1b-(S)
1g-(S)
1h-(2S,3R)
1i-(S)
CH2Cyclohexyl
Cyclohexyl (Cy)
Cy
2.2 2.9
0.08
4.2
0.27 0.005
0.28
0.36
0.39 0.37
0.42
0.43
0.45
0.52
0.57
0.58
0.60
0.70
1.3
3.0 0.31
7.4 1.8
0.60 0.11
were loaded onto 4-hydroxymethylphenoxymethyl polystyrene
(Wang) resin4 using standard conditions.5 Whenever possible,
commercially available enantiopure Fmoc-protected amino acids
preloaded onto Wang resin (3) were used.6 Compound 3 was trea-
ted with piperidine in DMF to remove the Fmoc protecting group,
and then reacted with substituted unprotected amino benzoic
acids 4 to give resin-bound amides 5. No undesired byproducts
were observed in this step, presumably due to the reactivity differ-
ence between the less reactive aniline and the more reactive pri-
Trp
2.5
1.1
1.7
4.7 0.7
2.1
CH(CH3)CH2CH3
2-Thienyl
CH2CO2H
CH2CONH2
CH(OH)CH3
CH2(4-OHPhenyl)
CH3
1.9
3.0
1j-(S)
1k-(S)
1m-(S)
1n-(S)
1o-(S)
1p-(S)
1q-(S)
H
Pro
mary amine. Lastly, compounds
5 were treated with the
2.9
3.3
>10
CH2CH(CH3)2
CH2CH2CO2H
CH2CH2CONH2
(CH2)4NH2
corresponding isocyanates, and the final products were cleaved
from the resin using a 1:1 solution of trifluoroacetic acid and
dichloromethane. Resin-bound intermediates 5 could also be re-
acted with sulfonyl chlorides, chloroformates, acid chlorides, or
carboxylic acids in a similar fashion.7
>10
Thus, this solid-phase synthesis strategy proved to be a highly
efficient and robust method for rapid SAR identification for this
series, using one synthetic sequence to explore three points of
diversity simultaneously. All compounds were purified by re-
verse-phase HPLC8 to purities of >95% (LCMS, UV 214 nm detec-
tion),9 and then evaluated in the enzyme inhibition assay against
human liver glycogen phosphorylase. Compounds were then se-
lected for further evaluation in the cellular assay.10 The results
are outlined below.
Table 2
In vitro GPa inhibition: phenyl spacer variations
R2
5
6
1
4
3
O
H
N
HO
Cl
2
H
O
HN
N
Initially, the SAR around the amino acid side chain (R1) was
investigated. The S-enantiomer (1c-(S),11 IC50 = 80 nM) was estab-
lished as the preferred isomer compared to the R-enantiomer
O
6
(1c-(R), IC50 = 4.2 lM). A number of S-amino acid derivatives were
Compound
R2
GPa IC50
,
l
M
GPa (cell) IC50, lM
prepared which showed that a variety of substituents were toler-
ated at the R1 position with submicromolar enzyme potency and
micromolar potency in the cellular assay. Compounds with a long-
er alkyl side chain were less potent, as in 1o versus 1b, 1p versus
1g, and 1a versus 1c-(S). The most dramatic result observed was
in shortening the side chain of the cyclohexylmethyl analogue
1a. By simply removing the methylene linker (1c-(S)), the enzyme
1c
6a
6b
6c
6d
6e
4,5-Phenyl (fused)
H
4,5-Dimethoxy
4,5-Difluoro
3,5-Dimethyl
3,6-Dichloro
0.08
100
7.0
0.86
>25
>25
0.60 0.11
>7.5 2.5
6.8 3.2
O
O
(a)
OH
+
NHFmoc
NHFmoc
HO
Pol
O
Pol
R1
R1
O
Wang resin
2
3
R2
R2
O
O
O
H
N
(b),(c)
(d),(e)
H
R3
N
Pol
O
NH2
N
N
HO
R1
O
H
H
HO2C
H2N
R1
O
R2
5
1
4
Scheme 1. Reagents and conditions: (a) 0.3 M 2,6-dimethylbenzoyl chloride; 0.5 M pyridine; 0.3 M R1CHNHFmocCO2H; DMF, rt, 20 h, (b) 20% piperidine/DMF, rt, 20 h, (c)
0.3 M EDC; 0.3 M HOAT; 0.3 M DIEA; 0.3 M R2CO2H; NMP, rt, 24 h, (d) 10 equiv R3NCO, pyridine, rt, 24 h, (e) 50% TFA/DCM, rt, 30 min.

4070
K. A. Evans et al. / Bioorg. Med. Chem. Lett. 18 (2008) 4068–4071
Table 3
With compound 1c-(S) now in hand as a potent lead, variations
on the central aryl ring R2 were investigated to better understand
the role of the naphthyl ring, and therefore a small number of
disubstituted phenyl analogues were prepared (Table 2). Replace-
ment of the naphthyl ring with an unsubstituted phenyl ring (6a)
was unfavorable, as were 3,5- and 3,6-disubstitutions (compounds
6d and 6e, respectively). While 4,5-disubstitutions (6b, 6c) were
tolerated, the naphthyl moiety (1c) remained the most potent.
The relationship of the urea and carbonyl group position on the
naphthyl ring was then studied (Table 3). It was quickly estab-
lished by the preparation of compounds 7a and 7b that the urea
and the carbonyl must be ortho to each other on the aryl ring.
A significant effort was also spent probing changes and require-
ments at the urea position R3 (Table 4). Replacement of the urea
with aryl sulfonamides, substituted phenyl amides (i.e., removal
In vitro GPa inhibition: naphthyl spacer variations
6
5
1
O
H
2
Cl
N
4
H
HO
3
HN
N
O
O
7
Compound
Carbonyl/urea relationship
GPa IC50, lM
1c
7a
7b
2,3-
1,4-
2,6-
0.08
100
100
of G in compound 8) (not shown, GPa IC50 = 100 lM), and ure-
thanes (8n) resulted in loss of potency. In fact, removal of any of
the three hydrogen bond positions in this template by replacement
of NH with NMe resulted in reduction in potency (GPa IC50 > 1 lM,
potency improved by 30-fold and the cellular potency improved to
submicromolar levels.
Table 4
In vitro GPa inhibition: urea variations
O
H
N
HO
R1
O
HN
G
R3
O
8
Compound
R1
R3
G
GPa IC50
,
l
M
GPa (cell) IC50, lM
1c
8a
8b
8c
8d
8e
8f
8g
1b
8h
8i
Cy
Cy
Cy
Cy
Cy
Cy
Cy
Cy
CH2CO2H
CH2CO2H
CH2CO2H
CH2CO2H
CH2CO2H
CH2CO2H
CH2CO2H
2-Cl-6-MePhenyl
2,6-diClPhenyl
2,6-diClPhenyl
2,6-diMePhenyl
2,6-diEtPhenyl
2,6-diFPhenyl
2-MePhenyl
N
N
C
0.08
0.10
6.8
0.12 0.02
0.59 0.28
4.9
30 41
100
0.39
0.23
0.12 0.08
5.0
1.2
1.6 1.5
100
0.60 0.11
0.32
1.6
1.1 0.25
4.5 1.9
N
N
N
N
N
N
N
N
N
N
N
O
>5
2-ClPhenyl
2-Cl-6-MePhenyl
2,6-diClPhenyl
2,6-diMePhenyl
2,6-diFPhenyl
2-MePhenyl
4.7
0.64 0.43
1.1
8j
8k
8m
8n
2-ClPhenyl
2-ClPhenyl
Table 5
In vitro metabolic stability in liver microsomes
O
H
N
HO
R3
H
R1
O
HN
N
6
O
2
6
Compound
R1
R3
T1/2 (min) human
T1/2 (min) rat
Clint (ml/min/kg) human
Clint (ml/min/kg) rat
1c
8a
8h
8i
Cy
Cy
2-Cl,6-Me
2,6-diCl
2,6-diCl
2,6-diMe
2-Cl,6-Me
16
16
>120
>120
>120
<15
<15
>120
>120
19
118
120
<15
<15
<15
1317
857
<23
<23
151
CH2CO2H
CH2CO2H
(CH2)4NH2
1q

K. A. Evans et al. / Bioorg. Med. Chem. Lett. 18 (2008) 4068–4071
4071
pharmacokinetic profile of this novel series will be the subject
of a future communication.
O
H
Acknowledgments
N
HO
HO2C
Cl
H
O
HN
N
The authors thank Walter Johnson for his assistance with the
LCMS spectra and Victoria Magaard for her assistance with com-
pound purification.
O
Cl
8h
References and notes
T1/2 = 4.0 h
AUC (iv) = 45247 ng-hr/mL
AUC (po) = 1671 ng-hr/mL
Clb = 0.92 mL/min/kg
Vss = 0.25 L/kg
1. (a) Treadway, J. L.; Mendys, P.; Hoover, D. J. Exp. Opin. Invest. Drugs 2001, 10,
439. and references therein; (b) Martin, W. H.; Hoover, D. J.; Armento, S. J.;
Stock, I. A.; McPherson, R. K.; Danley, D. E.; Stevenson, R. W.; Barrett, E. J.;
Treadway, J. L. Proc. Natl. Acad. Sci. 1998, 95, 1776.
2. (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.
%F = 1
Figure 1. In vivo rat DMPK of 8h (10 mg/kg po, 2.5 mg/kg iv, vehicle (2:98) DMS-
O:30% solutol in saline).
3. Inhibitors were tested for human liver glycogen phosphorylase enzymatic
not shown). However, modification of the urea to a phenylacetic
amide (G@C) was weakly tolerated (8b vs 8a). The key finding
was that appropriate 2,6-disubstitution on the aryl ring was mini-
mally required for good potency with methyl or chloro substitu-
ents being the most preferred (1c, 8a, 8c, 1b, 8h–i). Simple 2-
substitution showed a decrease in potency (8f–g, k–m), most pro-
nounced with the cyclohexyl side chain (8f–g).
activity using
a fluorescence intensity endpoint assay. To aid in the
identification of glucose-sensitive inhibitors of glycogen phosphorylase, the
assay was performed with and without 10 mM glucose. The change in
fluorescence due to product formation was measured on a fluorescence plate
reader (Viewlux, Perkin-Elmer) using
a 525 nm excitation filter and 595
emission filter. A concentration of 25 nM GPa enzyme was used. Therefore,
inhibitors can be accurately evaluated to an IC50 = 12.5 nM in this assay format.
The hGPa enzyme IC50 values given in Tables 1–4 are average values of at least
2 replicates where standard deviations are noted, and were measured in the
presence of glucose. See Ref. 7 for additional details.
In ADME studies, compound 1c-(S) was rapidly metabolized in
rat liver microsomes, with a half-life of <15 min and displayed very
high intrinsic clearance (CLint) of 1317 mL/min/kg (Table 5). This
observation was consistent with results from in vivo rat pharmaco-
kinetic studies, which showed very rapid clearance following intra-
venous administration (2.5 mg/kg), and very low Cmax of 7.3 ng/mL
when administered orally (10 mg/kg). Compound 8h had a much
better metabolic stability profile, and comparable cellular potency
to 1c-(S), and was therefore also tested in vivo. While compound
8h showed reasonable half-life (4 h) and lower plasma clearance
compared to 1c-(S), it unfortunately had poor oral exposure (Fig.
1). While it is not surprising that the diacid 8h has poor bioavail-
ability, this result demonstrates that by proper modifications of
the amino acid side chain, the clearance and half-life can be dra-
matically improved.
In summary, a series of amino acid anthranilamide derivatives
were successfully identified from high-throughput screening as
novel, potent, glucose-sensitive inhibitors of human liver glyco-
gen phosphorylase a. A solid-phase synthesis strategy was em-
ployed to provide efficient access to a variety of analogues for
rapid SAR generation. A potent exemplar of this series was
quickly identified (1c-(S), GPa IC50 = 80 nM) which appeared to
be rapidly metabolized both in vitro and in vivo, and a related
analogue was identified with comparable enzyme potency (8h),
GPa IC50 = 230 nM) and an improved metabolic stability profile.
The potency of compound 1c-(S) and the improved half-life of
compound 8h rendered this series worthy of continued explora-
tion and development with additional modifications to the ami-
no acid side chain, the phenylurea, and the central aryl ring.
Details regarding the further optimization of the potency and
4. The Wang resin was obtained from Polymer Laboratories, 1.7 mmol/g (PL-
Wang, Product # 1463, 150–300 lm).
5. Sieber, P. Tetrahedron Lett. 1987, 28, 6147.
6. Commercially available resins preloaded with L-amino acids were obtained
from Polymer Labs (www.polymerlabs.com), typically 0.86 mmol/g, >99%
optical purity.
7. 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.; Dickerson, S.H.; Peat, A.J.; Sparks, S.M.; Banker, P.; Cooper, J.P.
WO 2006/052722 A1.
8. Preparative HPLC purifications (reverse-phase) were performed using a Gilson
215 Liquid Handler with Unipoint software, typically with a YMC 20 Â 50
combiprep column (or 30 Â 75) ODS-A, 5u. A 5-min run (25 mL/min, 10% ACN/
H2O, 0.1%TFA to 90% ACN/H2O, 0.1% TFA) with UV detection at 254 nm was
typically used.
9. All novel compounds were characterized by LCMS, and gave satisfactory results
in agreement with the proposed structure. Purity data were determined by a
C18 reverse phase HPLC column [Keystone Aquasil (1 Â 40 mm)] in 10–90%
ACN/H2O containing 0.02% TFA (3.6 min gradient) and monitored by a UV
detector operating at 214 nm and by a SEDEX 75 evaporative light scattering
detector (ELSD) operating at 42 °C. LCMS M+H signals were consistent with
expected molecular weight for all reported products.
10. 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 prelabeled 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
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 1–4 are average values of at least 2
replicates where standard deviations are noted.
11. Spectral data for compound 1c-(S): 1H NMR (400 MHz, MeOH-d4) d: ppm 1.21–
1.36 (m, 5 H), 1.70 (d, J = 11.7 Hz, 1 H), 1.82 (d, J = 10.8 Hz, 1 H), 1.96 (br s, 1 H),
2.36 (s, 3 H), 4.52 (br s, 1 H), 7.21–7.26 (m, 2 H), 7.34 (d,J = 6.6 Hz, 1 H), 7.41 (dt,
J = 7.0, 0.9 Hz, 1 H) 7.51 (dt, J = 8.2, 1.0 Hz, 1 H), 7.76 (d, J = 8.2 Hz, 1 H), 7.88 (d,
J = 8.1 Hz,
1 H), 8.18 (s, 1 H), 8.57 (s, 1 H), 8.75 (br s, 1H). ESMS m/z
[M+H]+ = 494.4.