Novel GSK-3 inhibitors with improved cellular activity

Article abstract of DOI:10.1016/j.bmcl.2004.02.037

A novel series of [1-(1H-benzimidazol-7-yl)-1H-pyrazolo[3,4-d]pyrimidin-4- yl] arylhydrazones was synthesized and shown to potently inhibit glycogen synthase kinase-3 (GSK-3). In light of detailed structure-activity relationships and structural knowledge of the GSK-3 binding pocket, a benzimidazole substituent was incorporated onto the pyrazolopyrimidine core resulting in improved potency over previous analogs. More importantly, these derivatives show low nanomolar efficacy for stimulating glycogen synthesis in vitro and therefore may be useful in the treatment of type 2 diabetes mellitus.

Full text of DOI:10.1016/j.bmcl.2004.02.037

Bioorganic & Medicinal Chemistry Letters 14 (2004) 2127–2130
Novel GSK-3 inhibitors with improved cellular activity
Andrew J. Peat,* Dulce Garrido, Joyce A. Boucheron, Stephanie L. Schweiker,
Scott H. Dickerson, Jayme R. Wilson, Tony Y. Wang and Stephen A. Thomson
GlaxoSmithKline Research and Development, 5 Moore Drive, Research Triangle Park, NC 27709, USA
Received 7 January 2004; revised 6 February 2004; accepted 9 February 2004
Abstract—A novel series of [1-(1H-benzimidazol-7-yl)-1H-pyrazolo[3,4-d]pyrimidin-4-yl] arylhydrazones was synthesized and
shown to potently inhibit glycogen synthase kinase-3 (GSK-3). In light of detailed structure–activity relationships and structural
knowledge of the GSK-3 binding pocket, a benzimidazole substituent was incorporated onto the pyrazolopyrimidine core resulting
in improved potency over previous analogs. More importantly, these derivatives show low nanomolar efficacy for stimulating
glycogen synthesis in vitro and therefore may be useful in the treatment of type 2 diabetes mellitus.
Ó 2004 Elsevier Ltd. All rights reserved.
Glycogen synthase kinase-3 (GSK-3) is a ubiquitously
expressed serine–threonine kinase known to phosphor-
OMe
Ar1
ylate glycogen synthase.1 Glycogen synthase (GS)
H
N
N
H
N
N
N
N
functions as the rate-limiting enzyme for glycogen for-
mation and is inactivated upon phosphorylation by
GSK-3.2 Since GSK-3 is constitutively active in vivo,
GS typically exists in the inactive state. Insulin, signaling
via the PI-3 kinase/PKB pathway, inhibits GSK-3,
resulting in GS activation, and subsequent formation of
glycogen.3 The conversion of plasma glucose into gly-
cogen has been shown to be decreased in the liver and
skeletal muscle of diabetic patients.4 Analysis of skeletal
muscle from type 2 diabetic humans showed GSK-3
activity and expression levels were significantly higher
than those in healthy subjects.5 Therefore, small mole-
cules that mimic the action of insulin by inhibiting GSK-
3 may be useful for the treatment of type 2 diabetes
mellitus.
N
N
N
H
N
H
H
H
N
N
Ar2
template
N
1A
Figure 1. Pyrazolopyrimidine template.
pound 1A, which potently inhibits GSK-3
(pIC50 ¼ 8.2).8 Combining the SAR data with structural
information gained from the X-ray crystal structure of
GSK-3b,9 we have designed a second-generation com-
pound set with increased potency. More importantly,
these derivatives show significant improvements in cel-
lular efficacy for stimulating glycogen synthesis in vitro
as compared to our earlier series.
A number of publications have emerged describing
molecules that inhibit GSK-3.6 Several compounds have
also been reported to stimulate glycogen synthesis in
vitro6a;b as well as lower plasma glucose in diabetic
animals.7 We recently described a novel class of pyraz-
olopyrimidine analogs that inhibit GSK-3 (Fig. 1).
Chemical modifications to this template yielded detailed
structure–activity relationships (SAR) as well as com-
The binding orientation for the pyrazolopyrimidine
series was originally based on in silico docking of
compound 2A into a model of the ATP-binding pocket
generated from the crystal structure of GSK-3b. As
shown in Figure 2, the pyrazolopyrimidine makes two
hydrogen-bond contacts to the hinge region of the
enzyme through the N–H of the hydrazone and the N-5
nitrogen of the pyrimidine core. The hydrazone is ori-
ented toward the opening of the binding pocket where
the aryl group Ar2 interacts with water and possibly the
Keywords: Glycogen synthase kinase-3 (GSK-3); Glycogen synthesis;
Kinase inhibitor.
* Corresponding author. Tel.: +1-919-483-6137; fax: +1-919-483-6053;
e-mail: ajp25551@gsk.com
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.02.037

2128
A. J. Peat et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2127–2130
Figure 2. Proposed binding of compound 2A into the GSK-3 ATP-binding pocket.
positively charged arginine (Arg 141). As expected, a
variety of polar substituents are tolerated on Ar2 as
illustrated by compounds 1A–E, 2A–C, and 3A–E.
The aryl group Ar1 of the pyrazolopyrimidine protrudes
deeper into the enzyme toward a lysine residue (Lys 85).
In silico models and X-ray crystal data show this ring
occupies a narrow region within the enzyme and needs
to adopt a nearly coplanar geometry with the pyrazolo-
pyrimidine core for optimal binding. Therefore, we
hypothesized that introduction of substituents that
favor the co-planar arrangement of Ar1 with the pyra-
zolopyrimidine core would result in optimal potency.
The desired planar geometry might be achieved via
formation of an intramolecular hydrogen bond between
a hydrogen-bond donor on the aryl ring and the nitro-
gen at the 2-position of the pyrazole core. Ultimately we
decided that benzimidazole-containing analogs such as 3
(Scheme 1) would serve as good substrates to test this
idea. This ring system also places another heavy atom in
the narrow region of the enzyme that may enhance
binding (Fig. 3). In addition, the distal nitrogen atom in
the benzimidazole ring might interact as a hydrogen-
bond acceptor with the positively-charged Lys 85 in a
similar manner as the –OMe substituent.
Figure 3. In silico surface model of the GSK-3 ATP-binding pocket
with compound 3A.
steps from commercially available 3-nitrophenylene-
diamine. Treatment of aniline I with BF3–Et2O and
isoamyl nitrite produced the intermediate diazonium
salt that was subsequently reduced to the aryl hydrazine
II. A solution of II and ethoxymalononitrile in refluxing
ethanol gave the pyrazole-adduct III, which upon
heating with trimethyl ortho-formate yielded the imino-
ether IV. Formation of the pyrazolopyrimidine core was
affected via a Dimroth rearrangement by heating IV in
the presence of excess hydrazine monohydrate in etha-
nol to give the desired intermediate V.10 The final target
compounds 3A–E were synthesized via condensation of
V with various aldehydes and a catalytic amount of
pyrrolidine in refluxing ethanol.11
The novel benzimidazole–pyrazolopyrimidine deriva-
tives 3A–E were prepared as depicted in Scheme 1 from
4-amino-benzimidazole, which is readily obtained in two
N
CN
N
N
N
N
1. BF3-Et2O
Isoamyl nitrite
N
H
EtO
CH(OMe)3
H
N
CN
N
EtOH
reflux
(50%)
N
100 oC
(73%)
H
2. SnCl2/HCl
(48%)
H
H
H
N
N
H
H
H N
H
H
NC
III
I
II
N
N
N
N
N
H
O
H
N
N
N
N
H
H
N
N
N
N
Ar2
N
N
H
H2NNH2
N
N
N
cat. pyrrolidine
H-bond
N
H
MeO
EtOH
(60%)
H
EtOH
H
N
H
NH2
reflux
NC
Ar2
(33-81%)
V
IV
3A-E
Scheme 1. Syntheses of benzimidazole containing pyrazolopyrimidines.

A. J. Peat et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2127–2130
Table 1. Enzyme and cellular potency of pyrazolopyrimidines
2129
a
b;c
Compounds
Ar1 ¼
Ar2 ¼
4-Pyridyl
p-F-phenyl
pIC50
EC50 (lM)
%Maxc;d
Cell permeation (nM/s)
1A
1B
1C
1D
1E
2A
2B
2C
3A
3B
3C
3D
3E
3-OMe-phenyl
3-OMe-phenyl
3-OMe-phenyl
3-OMe-phenyl
3-OMe-phenyl
Phenyl
8.2
8.1
7.8
8.6
7.5
7.0
6.2
7.7
8.4
8.2
8.8
8.3
8.3
0.22 ( 0.03)
Inactive
37 ( 2)
Inactive
26 ( 2)
19 ( 0)
46 ( 2)
48 ( 3)
NT
127
5
p-C(O)NH(CH2)2NMe2-phenyl
p-SO2Me-phenyl
p-CO2H-phenyl
4-Pyridyl
4.00 ( 0.51)
0.31 ( 0.04)
1.88 ( 0.49)
1.22 ( 0.16)
NT
15
168
65
99
5
Phenyl
Phenyl
p-F-phenyl
p-C(O)NH(CH2)2NMe2-phenyl
4-Pyridyl
p-F-phenyl
NT
NT
98
212
92
15
61
17
Benzimidazole
Benzimidazole
Benzimidazole
Benzimidazole
Benzimidazole
0.11 ( 0.01)
0.23 ( 0.06)
2.33 ( 0.38)
0.24 ( 0.07)
2.73 ( 0.23)
45 ( 3)
57 ( 4)
76 ( 2)
54 ( 3)
42 ( 2)
p-C(O)NH(CH2)2NMe2-phenyl
p-SO2Me-phenyl
p-CO2H-phenyl
a Concentration, which inhibits 50% of the activity of GSK-3b.
b Concentration, which elicits 50% of the maximal response.
c Values are means of three experiments.
d Maximum glycogen produced by the compound as a percentage of the maximum glycogen produced by 1 lM insulin.
Table 1 highlights the potency of GSK-3 inhibition
(pIC50) and glycogen synthesis stimulation (EC50) of the
various analogs. The potency of the compound was
determined via a scintillation-proximity assay (SPA)
in which the ability of the compound to inhibit
The benzimidazole derivatives 3A–D also showed better
cellular responses for both EC50 and %max values as
compared to the 3-methoxyphenyl compounds 1A–D,
even though the respective potencies in the enzyme
inhibition assay are similar. The most dramatic case was
observed for the analogs in which Ar2 ¼ 4-fluorophenyl
(1B and 3B). Although both compounds are equipotent
against the enzyme (pIC50Õs ꢀ 8.1), the 3-methoxyphenyl
derivative 1B showed no effect on stimulating glycogen
synthesis, whereas the benzimidazole 3B produced a
robust response (EC50 ¼ 0.23 lM). The large difference in
observed efficacy might be a result of the ability of each
compound to cross the cell membrane. Since GSK-3 is an
intracellular kinase, compounds must enter the cell to
affect glycogen synthesis. We used a MDCK cell line
assay in an attempt to qualitatively estimate a com-
poundÕs ability to permeate the cell.14 As shown in Table
1, compound 1B gave a very low permeation value
(5 nM/s) and it is likely that this compound does not
readily enter the cell. Interestingly, replacement of the
3-methoxyphenyl group with the benzimidazole moiety
(compound 3B) led to a significant increase in the per-
meation value (92 nM/s). Therefore we propose the large
difference in efficacy between compounds 1B and 3B is a
result of their respective abilities to cross the cell mem-
brane. In general, for compounds with similar pIC50
values for GSK-3 inhibition and MDCK values >20 nM/
s, there is a consistent degree of efficacy for glycogen
production as illustrated among compounds 1A, 1D, 3A,
3B, and 3D. These compounds all have pIC50 values in
the range of 8.2–8.6, MDCK values >20 nM/s, and EC50
values between 0.11 and 0.31 lM. For compounds with
MDCK values <20 nM/s, the cellular efficacy was rela-
tively weak (EC50>1 lM) regardless of enzyme potency.
For example, 3A and 3E both contain a benzimidazole
and are equipotent at GSK-3. But while 3A shows good
cellular efficacy, 3E does not (EC50 ¼ 2.73 lM). Again,
based on the MDCK data, it appears that cell perme-
ation may account for the significant difference in effi-
cacy since 3E yielded a low permeation value (17 nM/s)
and 3A was quite high (212 nM/s). Therefore, within this
compound class we believe MDCK data is a useful tool
in predicting cell permeation and cellular efficacy.
the phosphorylation of
a synthetic GS peptide
by GSK-3b was measured.12 As hypothesized, the
benzimidazole-containing analogs 3A–E were quite
potent in inhibiting GSK-3 with pIC50Õs > 8. In all
examples studied, the benzimidazole series (see examples
3A–C) led to a significant (greater than 10-fold)
improvement in potency as compared with the unsub-
stituted phenyl derivatives (examples 2A–C8). In addi-
tion, it appears the benzimidazoles 3A–E were also
more potent than the corresponding 3-methoxyphenyl
compounds 1A–E8 though several are not statistically
significant.
We next looked at the effect of these compounds on the
synthesis of 14C-containing glycogen from 14C-labeled
glucose in a rat skeletal muscle cell line (L6 cells).13 The
cellular efficacy of each compound was derived from a
10-point dose–response curve and reported as the
effective concentration that elicits the half-maximal re-
sponse (EC50). The %max refers to the maximal glyco-
gen produced in response to the compound expressed as
a percentage of the maximal glycogen synthesized in
response to a fully efficacious dose of insulin (1 lM). In
this assay, compound 3A (EC50 ¼ 0.11 lM) showed a
10-fold improvement in efficacy as compared to the
unsubstituted analog 2A (EC50 ¼ 1.22 lM), which is in
good agreement with the relative potencies of these two
compounds for inhibition of GSK-3b (3A is 25-fold
more potent than 2A). The maximal amount of glycogen
produced in response to these two compounds was
essentially the same, ca. 45% of the maximal insulin
response. None of the compounds synthesized were able
to achieve 100% of the insulin value and analog 3C gave
the highest response at 76%. This is not surprising
considering that insulin likely stimulates other signaling
pathways, which also increase glycogen synthesis such as
upregulating glucose transport and inhibiting glycogen
breakdown.

2130
A. J. Peat et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2127–2130
Kurtzhals, P.; Bowler, A. N.; Ehrbar, U.; Hansen, B. F.
J. Med. Chem. 2003, 46, 3333; (c) Witherington, J.;
Bordas, V.; Garland, S. L.; Hickey, D. M. B.; Ife, R. J.;
Liddle, J.; Saunders, M.; Smith, D. G.; Ward, R. W.
Bioorg. Med. Chem. Lett. 2003, 13, 1577.
In conclusion, we have used previous SAR data and
structural knowledge of the GSK-3 binding pocket to
design a novel set of GSK-3 inhibitors. The benzimid-
azole substituent forms an intramolecular hydrogen
bond to the N-2 nitrogen of the pyrazole core, resulting
in a coplanar geometry that allows enhanced binding to
the kinase active site and increased potency as compared
to the unsubstituted-phenyl analogs. In addition, the
benzimidazole compounds demonstrate improved cel-
lular efficacy in stimulating glycogen synthesis over
previous series. We also found that within this set of
compounds, MDCK data is a useful tool in predicting
cell permeation and compound efficacy. The in vivo
effects of these compounds in animal models of type 2
diabetes will be reported shortly.
7. Henriksen, E. J.; Kinnick, T. R.; Teachey, M. K.; OÕKeefe,
M. P.; Ring, D.; Johnson, K. W.; Harrison, S. D. Am. J.
Physiol. Endocrinol. Metab. 2003, 284, E892.
8. Peat, A. J.; Boucheron, J. A.; Dickerson, S. H.; Garrido,
D.; Mills, W.; Peckham, J.; Preugschat, F.; Smalley, T.;
Schweiker, S. L.; Wilson, J. R.; Wang, T. Y.; Zhou, H. Q.;
Thomson, S. A. Bioorg. Med. Chem. Lett. 2004, 14, 2121–
2125.
9. Bax, B.; Carter, P. S.; Lewis, C.; Guy, A. R.; Bridges, A.;
Tanner, R.; Pettman, G.; Mannix, C.; Culbert, A. A.;
Brown, M. J. B.; Smith, D. G.; Reith, A. D. Structure
2001, 9, 1143–1152.
10. Hosmane, R. S.; Lim, B. B.; Burnett, F. N. J. Org. Chem.
1988, 53, 382.
References and notes
11. Compound purity was determined to be >95% by both 1H
NMR and LC/MS spectroscopy; the spectral data were
consistent with the reported structure.
1. (a) Cohen, P. In The Enzymes; Academic: New York,
1986; Vol. XVII, pp 461–497; (b) Woodgett, J. R.; Cohen,
P. Biochim. Biophys. Acta 1984, 788, 339.
2. For recent reviews see: (a) Eldar-Finkelman, H.; Ilouz, R.
Expert Opin. Invest. Drugs 2003, 12, 1511; (b) Eldar-
Finkelman, H. Trends Mol. Med. 2002, 8, 126.
3. (a) Grimes, C. A.; Jope, R. S. Prog. Neurobiol. 2001, 65,
391; (b) Frame, S.; Cohen, P. Biochem. J. 2001, 359, 1;
(c) Cross, D. A.; Alessi, D. R.; Vandenheede, J. R.;
McDowell, H. E.; Hundal, H. S.; Cohen, P. Biochem. J.
1994, 303, 21.
12. GSK-3b was assayed in 96-well microtitre plates at a final
concentration of 20 nM in 100 mL HEPES at pH 7.2
containing 10 mM MgCl2, 0.1 mg/mL bovine serum albu-
min, 1 mM dithiothreitol, 0.3 mg/mL heparin, 2.8 lM
peptide substrate (Biotin––Ahx-AAAKRREILSRRP-
S(PO3)YR-amide), 2.5 lM ATP and 0.2 lCi/well [c33P]-
ATP. After 40 min, the reaction was stopped by addition
of 100 mM EDTA and 1 mM solution in 100 mM HEPES
followed by a solution of diluted streptavidin coated SPA
beads (Amersham) in PBS to give a final concentration of
0.25 mg of beads per assay well. The wells were counted on
a Packard TopCount NXT microplate counter.
13. Glycogen accumulation in differentiated L6 cells was
determined by an adaptation of the method decribed by:
Thomas, J. A.; Schlender, K. K.; Larner, J. Anal. Biochem.
1968, 25, 486. Our exact protocol will be reported shortly.
14. Irvine, J. D.; Takahashi, L.; Lockhart, K.; Cheong, J.;
Tolan, J. W.; Selick, H. E.; Grove, J. R. J. Pharm. Sci.
1999, 88, 28.
4. DeFronzo, R. A.; Bonadonna, R. C.; Ferrannini, E.
Diabetes Care 1992, 15, 318.
5. Nikoulina, S. E.; Ciaraldi, T. P.; Mudailar, S.; Mohideen,
P.; Carter, L.; Henry, R. R. Diabetes 2000, 49, 263.
6. (a) Kuo, G.-H.; Prouty, C.; DeAngelis, A.; Shen, L.;
OÕNeill, D. J.; Shah, C.; Connolly, P. J.; Murray, W. V.;
Conway, B. R.; Cheung, P.; Westover, L.; Xu, J. Z.; Look,
R. A.; Demarest, K. T.; Emanuel, S.; Middleton, S. A.;
Jolliffe, L.; Beavers, M. P.; Chen, X. J. Med. Chem. 2003,
46, 4021; (b) Olesen, P. H.; Sorensen, A. R.; Urso, B.;