Synthesis and evaluation of novel heterocyclic inhibitors of GSK-3

Article abstract of DOI:10.1016/j.bmcl.2006.01.057

A set of novel heterocyclic pyrimidyl hydrazones has been synthesized as inhibitors of glycogen synthase kinase-3 (GSK-3) with the most active exhibiting low nanomolar activity. Quantum mechanical calculations indicate that of the conformational factors t

Full text of DOI:10.1016/j.bmcl.2006.01.057

Bioorganic & Medicinal Chemistry Letters 16 (2006) 2091–2094
Synthesis and evaluation of novel heterocyclic inhibitors of GSK-3
Terrence L. Smalley, Jr.,* Andrew J. Peat, Joyce A. Boucheron, Scott Dickerson,
Dulce Garrido, Frank Preugschat, Stephanie L. Schweiker,
Stephen A. Thomson and Tony Y. Wang
Research & Development, GlaxoSmithKline, Inc., 5 Moore Drive, Research Triangle Park, NC 27709, USA
Received 16 December 2005; revised 16 January 2006; accepted 17 January 2006
Available online 7 February 2006
Abstract—A set of novel heterocyclic pyrimidyl hydrazones has been synthesized as inhibitors of glycogen synthase kinase-3
(GSK-3) with the most active exhibiting low nanomolar activity. Quantum mechanical calculations indicate that of the
conformational factors that could determine binding affinity, the planarity of the phenyl ring in relation to the central core and
the conformation of the hydrazone chain may be the most influential.
Ó 2006 Elsevier Ltd. All rights reserved.
Glycogen synthase kinase-3 (GSK-3), a protein in the
N
1 X=N, Y=N, Z=CH
2 X=C, Y=CH, Z=NH
3 X=C, Y=CH, Z=NMe
4 X=N, Y=CH, Z=CH
5 X=N, Y=CH, Z=N
6 X=C, Y=N, Z=NH
serine/threonine kinase family, is broadly expressed
and serves many functions within the human body.1
Among some of the diseases that GSK-3 may affect
are Alzheimer’s disease, diabetes, various cancer types,
and neurological disorders.2 One function of GSK-3 is
to mediate the conversion of glycogen to glucose and
is regulated in part by insulin signaling. In patients with
insulin resistance, GSK-3 is constituently active, which
leads to an increase in plasma glucose levels and hyper-
glycemia.3 Inhibitors of GSK-3 could reduce glucose
levels by mimicking the effect of insulin signaling on
GSK-3 and thus could be used as anti-diabetic treat-
ments. As part of our research efforts to discover effec-
tive medicines for the treatment of metabolic diseases,
we became interested in inhibitors of GSK-3 as a poten-
tial treatment for diabetes.
H
N
N
N
N
Z
X
Y
OMe
Figure 1.
the pyrazole portion of the core structure was varied,
as exemplified by structures 2–6 (Fig. 1).5
The binding of 1 to the ATP-binding pocket of GSK-3
has been described and several key interactions were ob-
served (Fig. 2).6 Hydrogen bonding interactions of 1
with the b-strand of GSK-3 are required for binding,
and for this reason we opted to retain the hydrazinopyr-
imidine portion of the core. The hydrazone moiety was
shown to adopt an S-cis conformation which provides a
desirable binding interaction with the protein back-
bone.4 Previous structure–activity relationships (SAR)
suggested that the phenyl ring at N-1 is required to exist
in a co-planar arrangement with the pyrazolopyrimidine
core in order to access a narrow pocket of the protein.
We became interested in pyrazole modifications as a
way to further enhance the co-planar relationship of
We recently disclosed a series of GSK-3 inhibitors de-
signed around
a pyrazolopyrimidine template (1,
Fig. 1).4 While the compounds described therein were
interesting inhibitors that showed a broad spectrum of
activity against GSK-3, our efforts were mainly focused
on peripheral modifications but retained a pyrazolo-
pyrimidine core. This communication describes our
efforts to discover potent GSK-3 inhibitors in which
Keywords: GSK-3; Heterocycles; Diabetes; Inhibitors.
*
Corresponding author. Tel.: +1 919 483 1054; fax: +1 919 315
0430; e-mail: terry.l.smalley@gsk.com
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.01.057

2092
T. L. Smalley, Jr. et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2091–2094
the pyrrole. Ring closure with formic acid and subse-
quent chlorination as previously described4 provided
the desired pyrrolopyrimidine 9. The chloro group was
displaced with hydrazine followed by condensation with
pyridine-4-carboxaldehyde to provide 2. Alternatively,
the N-methyl derivative 3 could be accessed by methyl-
ation of 9. Displacement of the chloro group and con-
densation as above gave 3.
The preparation of pyrrolopyrimidine 4 is depicted in
Scheme 2. Aldehyde 119 was protected as the 1,3-dioxo-
lane prior to chemoselective displacement of only one of
the chloro groups to provide the pyrrole precursor 12.
Acid catalyzed hydrolysis of the 1,3-dioxolane and sub-
sequent cyclization provided 4-chloropyrrolopyrimidine
13. Compound 4 was obtained upon displacement of the
chloro group with hydrazine hydrate and condensation
with the appropriate aldehyde.
Figure 2. Model of 1 docked into the active site of GSK-3.7 This
representation depicts the S-cis conformation of the hydrazone and
also indicates the importance of ring planarity of the 3-methoxyphenyl
ring with the pyrazolopyrimidine core.
Imidazopyrimidine 5 was synthesized according to the
outline depicted in Scheme 3. 5-Amino-4,6-dichloropyr-
imidine10 was treated with 3-methoxyaniline to afford a
monosubstituted product, which was subsequently treat-
ed with diethoxymethyl acetate11 to afford ring closure
and provide 15. From that point the route described
above was followed to install the hydrazone.
the phenyl ring with the core, as well as strengthen the S-
cis conformation of the hydrazone.
The syntheses of the new inhibitors are described in
Schemes 1–4.4,5 The synthesis of 2 began using a modi-
fied literature procedure to access 8 (Scheme 1).8 With
the desired functionality present, 8 was treated with
Na2CO3 to remove the carbamate protecting group on
The synthesis of pyrazolopyrimidine 6 is shown in
Scheme 4. (3-Methoxyphenyl)acetonitrile was treated
with ethyl diazoacetate followed by ring closure with
N
CO2Et
N
CN
Cl
N
H
N
N
R
N
N
N
NC
R
N
h,i
a,b,c
d,e,f
N
H2N
OMe
OMe
OMe
OMe
9 R=H
8
7
2 R=H
3 R=Me
g
10 R=Me
Scheme 1. Reagents and conditions: (a) HCO2Et, NaH; (b) H2NCH2CN, NaOAc, 80%; (c) ethyl chloroformate, DBN, 0 °C, then DBN, 0 °C to rt;
(d) Na2CO3, MeOH, 71%; (e) HCO2H, reflux, 75%; (f) POCl3, 100 °C; (g) (MeO)2SO2, K2CO3, 75%; (h) hydrazine hydrate, EtOH, 80 °C, 82%; (i)
pyridine-4-carboxaldehyde, pyrrolidine (cat.), EtOH, 80 °C, 80%.
Cl
N
Cl
N
Cl
N
O
a,b
c
d,e
N
N
CHO
N
4
O
H
N
Cl
N
OMe
OMe
11
12
13
Scheme 2. Reagents and conditions: (a) ethylene glycol, p-TsOH, PhH, 100 °C, 64%; (b) m-anisidine, p-TsOH (cat.), THF, sealed tube, 130 °C, 39%;
(c) 1.5 M aq HCl, THF, 80 °C, 61%; (d) hydrazine hydrate, EtOH, 80 °C, 44%; (e) pyridine-4-carboxaldehyde, pyrrolidine (cat.), EtOH, 80 °C, 78%.

T. L. Smalley, Jr. et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2091–2094
2093
N
H
N
N
N
N
Cl
N
Cl
N
NH2
Cl
N
N
N
N
a,b
c,d
N
N
OMe
OMe
15
5
Scheme 3. Reagents and conditions: (a) 3-methoxyaniline, EtOH, 98%; (b) diethoxymethyl acetate, D, 54%; (c) hydrazine hydrate, EtOH, 80 °C; (d)
pyridine-4-carboxaldehyde, pyrrolidine (cat.), EtOH, 80 °C.
N
H
N
N
OH
N
N
N
OMe
H
N
N
H
N
N
a,b
c,d,e
N
NC
OMe
OMe
14
6
Scheme 4. Reagents and conditions: (a) EtO2CCHN2, NaOEt, NaOH, 29%; (b) formamidine acetate; (c) POCl3; (d) hydrazine hydrate, EtOH, 80 °C;
(e) pyridine-4-carboxaldehyde, pyrrolidine (cat.), EtOH, 80 °C.
formamidine acetate to provide 14. Treatment with
Table 1. Conformational energy and GSK-3 inhibitory activity of
novel heterocyclic compounds 2–6
POCl3 provided the chloro derivative, which was prose-
cuted in a similar manner as previously described to af-
ford the desired product 6.
N
Compounds 2–6 were evaluated as inhibitors of GSK-3
and the results are shown in Table 1.12 The pyrrole
analog 2 was equipotent to 1 (pIC50 = 8.3), however
N-methylation (3) gave a 100-fold loss of activity. The
isomeric pyrrole (4) was shown to be an order of magni-
tude less active than both 1 and 2. Imidazole 5 was also
significantly less active (pIC50 = 6.2), but the isomeric
pyrazole (6) was shown to be equipotent with 1 and 2.
Thus, we observed that both 2 and 6 were comparable
in activity to 1.
H
N
N
N
N
Z
X
Y
OMe
Compound
X
Y
N
Z
pIC50 Cis versus
Planarity
trans energya energyb
While the discovery of potent core templates was exciting,
we sought to explain the difference in activities of our
inhibitors. We compared the conformations of 2–6 with
that of 1 to determine the existence of any favorable bind-
ing interactions. One conformational aspect that had
been previously observed was the preference for planarity
of the phenyl ring at N-1.4 The planarity can be modulat-
ed by ortho substituents on the phenyl ring, as these are in
close proximity to N-2. One explanation of this effect is
that with nitrogen at the 2-position (e.g, 1), only aminimal
steric interaction with the hydrogens on the ortho-posi-
tion of the phenyl ring exists (Fig. 3). To assess the poten-
(kcal/mol)
(kcal/mol)
1
2
3
4
5
6
N
CH
8.2
8.3
À5.7
À8.3
ND
0.13
0.61
ND
3.2
C
CH NH
C
CH NMe 6.3
N
CH CH
7.5
6.2
8.3
À4.7
À0.06
À8.3
N
CH
CH
N
N
NH
2.6
0.0
a The cis versus trans energy refers to the conformation of the hydra-
zone moiety.
b Planarity energy refers to the energy required to force a co-planar
conformation between the phenyl ring and the core.

2094
T. L. Smalley, Jr. et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2091–2094
3. (a) Maeda, Y.; Nakano, M.; Sato, H.; Miyazaki, Y.;
N
N
Schweiker, S. L.; Smith, J. L.; Truesdale, A. T. Bioorg.
Med. Chem. Lett. 2004, 14, 3907; (b) Wagman, A. S.;
Johnson, K. W.; Bussiere, D. E. Curr. Pharm. Design
2004, 10, 1105; (c) Desai, M.; Ni, Z.-J.; Ng, S.; Pfister, K.
B.; Ramurthy, S.; Subramanian, S.; Wagman, A. S. Patent
#WO2004018455; Chem. Abstr., 2004, 140, 235595; (d)
Eldar-Finkelman, H.; Ilouz, R. Expert Opin. Invest. Drugs
2003, 12, 1511; (e) Ring, D. B.; Johnson, K. W.;
Henriksen, E. J.; Nuss, J. M.; Goff, D.; Kinnick, T. R.;
Ma, S. T.; Reeder, J. W.; Samuels, I.; Slabiak, T.;
Wagman, A. S.; Hammond, M.-E. W.; Harrison, S. D.
Diabetes 2003, 52, 588.
H
N
N
lone pair
repulsion
H
N
N
N
N
N
H
N
N
N
H
twist
N
N
H
H
OMe
OMe
4. (a) Brown, M. L.; Cheung, M.; Dickerson, S. H.; Garrido,
D. M.; Mills, W. Y.; Miyazaki, Y.; Peat, A. J.; Peckham,
J. P.; Smalley, T. L.; Thomson, S. A.; Veal, J. M.; Wilson,
J. L. R. Patent#WO 2004009602; Chem. Abstr., 2004, 140,
128436; (b) Peat, A. J.; Boucheron, J.; Dickerson, S.;
Garrido, D.; Mills, W.; Peckham, J.; Preugschat, F.;
Smalley, T.; Schweiker, S.; Wilson, J.; Wang, T.; Zhou,
H.-Q.; Thomson, S. A. Bioorg. Med. Chem. Lett. 2004, 14,
2121.
5
1
Figure 3.
tial importance of this interaction, ab initio calculations
were performed on the inhibitors and are shown in
Table 1.13 The results suggest that increasing the twist of
the phenyl ring could clearly account for the observed
modulation in activity of our inhibitors. In addition, the
conformation of the hydrazone chain is another factor
influencing binding as seen in 2 and 6. In both of these
inhibitors, the S-cis hydrazone conformation is preferred
due to an intramolecular hydrogen bond. Methylation of
2 removed the hydrogen bond (3), dramatically decreas-
ing the activity.
1
5. All novel compounds exhibited acceptable H NMR and
mass spectra, and were of >95% purity as determined by
analytical HPLC.
6. Image was created with PyMol.
7. Structural coordinates for the GSK-3 structure were
similar to those reported in: 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.
When the substitution pattern is that as in 4, the steric
interactions of the phenyl and hydrogen at C-2 were
amplified and the activity was reduced significantly.
The trend was observed with 5 as well, as an order of
magnitude decrease in binding was observed. However,
this trend was not followed with pyrrole 2 which also
contains a C–H at the 2-position. We believe that the in-
creased bond length between C-1 and the phenyl ring in
2 lessens the non-covalent interaction between C-2 and
the ortho hydrogens on the phenyl ring, thus reducing
any potential contributions to the activity. In addition
to the twist of the phenyl ring in compound 5, N-3 inter-
acts unfavorably with the nitrogen of the hydrazone
(Fig. 3). This interaction could lower the preference of
cis versus trans orientation, which in turn could be par-
tially accountable for the observed decrease in activity.
8. Buchanan, J. G.; Craven, D. A.; Wightman, R. H.;
Harnden, M. R. J. Chem. Soc., Perkin Trans. 1 1991, 195.
9. (a) Talekar, R. R.; Wightman, R. H. Tetrahedron 1997, 53,
3831; (b) Cristalli, G.; Franchetti, P.; Grifantini, M.;
Vittori, S.; Lupidi, G.; Riva, F.; Bordoni, T.; Geroni, C.;
Verini, M. A. J. Med. Chem. 1988, 31, 390; (c) Mont-
gomery, J. A.; Hewson, K. J. Med. C.hem. 1967, 10, 665.
10. Commercially available from Aldrich Chemical Company
and was used as received.
11. (a) Montgomery, J. A.; Holum, L. B. J. Am. Chem. Soc.
1958, 40, 404; (b) Prasad, R. N.; Noell, C. W.; Robins, R.
K. J. Am. Chem. Soc. 1959, 81, 193.
12. GSK-3 was assayed in 96-well microtiter 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 ATP, solution in
100 mM Hepes followed by a solution of streptavidin
coated SPA beads (Amersham) in PBS to give a final
concentration of 0.25 mg of beads per assay well. The
plates were counted on a Packard TopCount NXT
microplate counter.
13. The energies were calculated at the HF/6-31G(d) level of
theory with Gaussian 98 from Gaussian, Inc. The values
reported reflect the energy difference of the trans hydra-
zone conformation relative to the cis hydrazone, with a
value of 0 kcal/mol indicating that no preference between
the two conformations exists. A positive number suggests
that a preference for the trans hydrazone conformation
exists and a negative number suggests that a preference for
the cis hydrazone conformation exists.
In conclusion, we have synthesized novel heterocyclic
inhibitors of GSK-3 based on a template previously dis-
covered. The activity can be rationalized by computa-
tional analysis and thus can be modulated based on
the heterocycle and the substitution pattern about the
ring. The most active compounds were currently being
further evaluated as potential anti-diabetic treatments
and these results will be reported in due course.
References and notes
1. Cohen, P.; Frame, S. Nat. Rev. 2001, 2, 1.
2. Ali, A.; Hoeflich, K. P.; Woodgett, J. R. Chem. Rev. 2001,
101, 2527.