The discovery of a novel series of glucokinase activators based on a pyrazolopyrimidine scaffold

Article abstract of DOI:10.1016/j.bmcl.2012.10.090

Glucokinase is a key enzyme in glucose homeostasis since it phosphorylates glucose to give glucose-6-phosphate, which is the first step in glycolysis. GK activators have been proven to lower blood-glucose, and therefore have potential as treatments for type 2 diabetes. Here the discovery of pyrazolopyrimidine GKAs is reported. An original singleton hit from a high-throughput screen with micromolar levels of potency was optimised to give compounds with nanomolar activities. Key steps in this success were the introduction of an extra side-chain, which increased potency, and changing the linking functionality from a thioether to an ether, which led to improved potency and lipophilic ligand efficiency. This also led to more stable compounds with improved profiles in biological assays.

Full text of DOI:10.1016/j.bmcl.2012.10.090

Bioorganic & Medicinal Chemistry Letters 22 (2012) 7302–7305
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
The discovery of a novel series of glucokinase activators based
on a pyrazolopyrimidine scaffold
Peter Bonn a, D. Mikael Brink a, , Jonas Fägerhag a, Ulrik Jurva a, Graeme R. Robb b, Volker Schnecke a,
⇑
a
Anette Svensson Henriksson a, Michael J. Waring b, , Christer Westerlund
⇑
a Cardiovascular and Gastrointestinal Innovative Medicines, AstraZeneca, S-431 83 Mölndal, Sweden
b Cardiovascular and Gastrointestinal Innovative Medicines, AstraZeneca, Mereside, Alderley Park, Macclesfield, Cheshire SK10 4TG, United Kingdom
a r t i c l e i n f o
a b s t r a c t
Article history:
Glucokinase is a key enzyme in glucose homeostasis since it phosphorylates glucose to give glucose-6-
phosphate, which is the first step in glycolysis. GK activators have been proven to lower blood-glucose,
and therefore have potential as treatments for type 2 diabetes. Here the discovery of pyrazolopyrimidine
GKAs is reported. An original singleton hit from a high-throughput screen with micromolar levels of
potency was optimised to give compounds with nanomolar activities. Key steps in this success were
the introduction of an extra side-chain, which increased potency, and changing the linking functionality
from a thioether to an ether, which led to improved potency and lipophilic ligand efficiency. This also led
to more stable compounds with improved profiles in biological assays.
Received 24 September 2012
Revised 16 October 2012
Accepted 19 October 2012
Available online 27 October 2012
Keywords:
Glucokinase
Pyrazolopyrimidine
Enzyme activators
Ó 2012 Elsevier Ltd. All rights reserved.
Type 2 diabetes (T2D) is a complex disease which affects 150
million people worldwide, with its prevalence expected to double
by the year 2025.1 Current therapies do not achieve adequate
glycaemic control2 and consequently there is an unmet need for
therapeutic agents which deliver greater efficacy.
moderately potent activator of human GK (pEC50 5.6) albeit with
sub-optimal lipophilicity (HPLC logD 3.8). Whilst this activity
was significantly weaker than that of 1 and 2, it was considered
to be sufficiently differentiated in terms of chemical structure to
warrant an optimisation programme.
Glucokinase (GK, also known as hexokinase IV or D) is
expressed predominantly in the liver, pancreas, brain and gut. Its
function is to catalyse the conversion of glucose to glucose-6-
phosphate. In both the liver, where the action of GK promotes gly-
cogen synthesis, and the pancreas, where the action of GK results
in glucose-sensitive insulin release,3 GK is the rate limiting enzyme
in glucose utilisation.4 Activation of GK is therefore expected to im-
prove glycaemic control by modulating hepatic glucose balance
and decreasing the threshold for insulin secretion.5–7
Several groups have described small molecule glucokinase
activators (GKAs), which act by binding to an allosteric site.8 Astra-
Zeneca has described previously the identification and subsequent
optimisation of a series of GKAs resulting in the discovery of the
In exploring the SAR around 3, particular attention was given to
variants of the thiazole (termed the R1 substituent), introduction of
an additional substituent a
-to the amide carbonyl (termed R2) and
variants of the 2-chlorophenyl-N substituent of the pyrazolopyr-
imidine ring (Fig. 2). From examining the output from the screen,
it was interesting to note that an analogue without the 2-chloro
substituent in the pendant benzene ring 4 was considerably less
active (pEC50 <4) suggesting that a 2-substituent should be main-
tained. Hence, structural variation in this region was restricted to
varying the o-substituent (R3) with groups of similar size to chloro.
Within the structure of 3, the geminal donor–acceptor motif
formed by the nitrogen of the thiazole ring and the NH of the
amide substituent was of note. Motifs of this type are common
to a significant number of GKAs and have been observed by
X-ray crystallography to form an interaction with the protein back-
bone at Arg63.15,16 A series of analogues replacing the thiazole ring
with other cyclic aromatic systems were supportive of the same
binding mode for these compounds (Table 1). For example, the
phenyl derivative 5 lost all measurable activity, consistent with
the requirement for an acceptor in the ring. The 4-methylpyridyl
derivative 6 showed increased potency relative to 3, which was
greater than the corresponding increase in logD [0.3 increase in
lipophilic ligand efficiency (LLE)].17 More polar heterocycles such
as pyrimidine 7 and methylpyrazole 8 were less potent. Compound
development candidates
1 and 2 (AZD1092 and AZD1656,
Fig. 1).9–14 The clinical promise of GKAs but close structural simi-
larity of the development compounds gave rise to the need for a
structurally differentiated compound which would mitigate
unforeseen issues which might have been encountered in the pro-
gression of 1 and 2.
A high-throughput screen of the AstraZeneca compound collec-
tion revealed the singleton hit pyrazolopyrimidine 3 (Fig. 2) as a
⇑
Corresponding authors.
E-mail address: mike.j.waring@astrazeneca.com (M.J. Waring).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.10.090

P. Bonn et al. / Bioorg. Med. Chem. Lett. 22 (2012) 7302–7305
7303
N
O
O
Trp99
N
O
O
N
N
N
N
HO
N
O
N
H
H
O
N
N
O
1 - AZD1092
2 - AZD1656
O
O
Tyr215
Figure 1. AZD1092, AZD1656.
Arg63
O
S
R1
S
N
3: R1 =
,R2 = H, R3=Cl
,R2 = H, R3=H
N
N
N
H
R2
N
S
4: R1 =
N
N
Figure 3. Docking of 3 into a protein crystal structure of glucokinase. The back
R3
surface of the binding site is shown in gold. Dotted green lines indicate hydrogen-
bonds; dotted orange lines indicate edge-face p-interactions.
Figure 2. Generic structure of the series and high throughput screening hits.
Table 2
7 exhibited no measurable activity and 8 showed a 0.8 unit drop in
pEC50 relative to 3, albeit with a significant drop in logD (0.5 unit
increase in LLE).
Computational docking of 3 in a GK crystal structure was con-
sistent with the hypothesised hydrogen bonding interactions with
Arg63 and suggested that the pendant o-chlorophenyl ring could
Structure–activity relationships for the R2 position
O
S
N
N
S
N
N
R2
H
N
N
form edge-face p-interactions with Trp99 and Tyr215 (Fig. 3). This
Cl
also highlighted that substitution in the R2 position would allow
access to an unoccupied pocket in the protein and, hence, might
lead to an increase in potency (Fig. 3). Accordingly, R2 alkyl substit-
uents such as n-butyl 9 (Table 2) showed an increase in potency
but this was outbalanced by the increase in logD resulting in re-
duced LLE. Methoxyethyl derivative 10 was found to be equipotent
with 9 but with significantly lower logD suggesting that the ether
oxygen can form a specific interaction with the protein. Compound
10 was also more potent than 3 (0.5 units) but was still less attrac-
tive in terms of LLE (0.5 units). Larger substituents such as
(methoxyethoxy)ethyl 11 and phenyl 12 showed reduced potency
and LLE. Nevertheless, this series of analogues showed that po-
tency could be modulated by substitution in the R2 position.
At this stage, stability studies revealed that the thioether link-
age contained in these compounds was unstable in incubations
with glutathione, forming SNAr-type adducts (Scheme 1). The ex-
tent of reaction observed was found to be 5 times that of clozapine
R2
GK pEC50
HPLC logD
LLE
1.8
À0.2
1.3
3
9
–H
–n-Bu
5.6
6.2
3.8
6.4
O
O
10
11
6.1
5.5
4.8
4.7
0.8
O
12
<4.0
5.6
<À1.6
SG
O
S
O
S
GSH
N
N
N
S
N
N
N
N
HS
+
N
H
N
N
N
Cl
H
N
Cl
Scheme 1. Hydrolytic instability of thioethers.
Table 1
Structure–activity relationships for the R1 position
O
in incubations with human liver microsomes supplemented with
glutathione, indicating that this was a significant issue for the com-
pounds.18 This issue might have been expected given the known
chemical reactivity of thioether substituted electron deficient sys-
tems of this type19 and it was anticipated that the corresponding
ether analogues would be more stable. Such analogues 13 and 14
possessed significantly improved potency relative to the thioether
matched molecular pairs (Table 3) and, combined with the ex-
pected reduction in logD, led to large increases in LLE. Interest-
ingly, these changes were not consistent in magnitude, with the
improvement observed with the R2 n-butyl substituted 14 (pEC50
+0.7 relative to 10) being greater than that observed with the
unsubstituted 13 (pEC50 +0.2 relative to 3). Gratifyingly, these
compounds were established to be stable with respect to glutathi-
one reactivity.
R1
N
S
N
H
N
N
N
Cl
R1
GK pEC50
5.6
HPLC logD
3.8
LLE
1.8
S
3
5
6
N
<4.0
4.3
<À0.3
6.3
4.2
2.1
N
N
With these encouraging results in hand, the R3 position re-
mained to be explored. The best combination of groups identified
thus far was R1 = 4-methylpyridine, R2 = methoxyethyl in the ether
series. R3 changes were therefore explored with this more optimal
7
8
<4.0
4.8
2.2
2.5
<1.8
2.3
N
N
N

7304
P. Bonn et al. / Bioorg. Med. Chem. Lett. 22 (2012) 7302–7305
Table 3
Table 5
Ether analogues, figures in parentheses show change from thioether matched
molecular pair
Technical profiles of single enantiomers
O
O
O
S
N
O
N
N
N
N
O
H
N
N
N
H
R2
N
N
N
N
R3
Cl
R3
S-15
S-17
S-18
R2
GK pEC50
HPLC logD
LLE
–Cl
–CF3
–OMe
13
14
–H
–n-Bu
5.8 (+0.2)
6.9 (+0.7)
3.2 (À0.6)
5.7 (À0.7)
2.6 (+0.8)
1.2 (+1.4)
a
GK pEC50
logD
LLE
7.3 (5.0)
3.9
3.4
95
7.2 (5.3)
4.3
2.9
6.9 (<4.0)
3.5
3.4
Solubility (
lM)
78
86
hERG IC50
(
l
M)
22
35
34
3.2
53
26
27
>33
132
Rat hep. Clint(mL minÀ1 Â 106 cells)
Table 4
Structure–activity relationships for the R3 position
Rat Cl (mL minÀ1 kgÀ1
)
Rat Vss (L kgÀ1
Rat F (%)
)
O
N
O
N
N
a
H
The GK pEC50 values for the corresponding R enantiomers were 5.0 for R-15, 5.3
N
for R-17 and <4.0 for R-18.
N
O
N
R3
which in turn were derived from the corresponding aryl hydra-
zines as exemplified for the R3 chloro derivatives (Scheme 2).21
This synthetic sequence started with condensation of 2-chloro-
phenylhydrazine 19 with 2-(ethoxymethylene)malononitrile to af-
ford the 5-amino-4-cyanopyrazole 20. The nitrile of 20 was
converted to the corresponding primary amide by acidic hydrolysis
in concentrated sulphuric acid. The pyrimidine ring was then
annulated by treatment with formamide under microwave heating
to form the pyrazolopyrimidinone 21. This could be converted to
the corresponding thiol 22 by treatment with Lawesson’s reagent
or to the chloride 23 by reaction with phosphorous oxychloride.
The thioether derivatives could be prepared by alkylation of thi-
ols such as 22 with an appropriate 2-bromoacetamide (Scheme 3).
For example, 9 was prepared by coupling of 2-bromohexanoic acid
24 with 2-amino-4-methylthiazole to afford the amide 25 followed
by reaction with 22 to give 9.
R3
GK pEC50
HPLC logD
LLE
15
16
17
18
–Cl
7.0
6.5
7.0
6.7
4.1
4.2
4.4
3.5
2.9
2.3
2.6
3.2
–Me
–CF3
–OMe
combination. In order to investigate this, the comparator R3 chloro
derivative 15 was prepared (Table 4). This compound proved that
the observed benefits of the changes to R1, R2 and the linker were
additive with 15 proving the most active (and highest LLE) com-
pound to date. R3 variations failed to improve on this compound
suggesting that chloro may be the optimal R3 substituent. For
example, methyl derivative 16 showed reduced potency with
equivalent logD and trifluoromethyl derivative 17 was equipotent
but with higher logD and reduced LLE. The methoxy derivative 18
showed a reduction in potency although with a larger drop in logD,
leading to increased LLE.
The ethers were prepared by SNAr displacement of the chlorides
such as 23 with an appropriate 2-hydroxyacetamide (Scheme 4).
For example, 15 was prepared by initial bis-silyl protection of
2-hydroxy-4-methoxybutanoic acid 26. The resulting bis-silyl
derivative 27 was coupled to 2-amino-5-methylpyridine by treat-
ment with oxalyl chloride to effect selective deprotection of the
Up until this point, R2 substituted analogues had been assessed
in racemic form. Given the encouraging potencies and LLE values
observed with compounds such as 15, 17 and 18, separation of
the enantiomers was undertaken. This revealed, as would be ex-
pected, that one enantiomer was significantly more active than
the other. The active enantiomers were subsequently established
by X-ray crystallography to possess the S-configuration.20 Further
profiling of the active enantiomers revealed that despite their logD
values, which were still relatively high, they exhibited good solu-
bility and low hERG activity (Table 5). In vitro intrinsic clearance
was moderate for S-15 and S-17 but was higher for the methoxy
derivative S-18. This apparent discrepancy could be rationalised
by metabolite identification studies, which revealed that cleavage
of the methoxy ether is a significant metabolic pathway (data
not shown). Compound S-15 was progressed to in vivo rat pharma-
cokinetic studies and showed clearance levels which were also
moderate in vivo and a reasonable bioavailability (53%) indicative
of very high absorption. Overall, these data showed encouraging
technical profiles, indicating that the series was worthy of further
optimisation, in particular if lipophilicity could be reduced.
The synthesis of these compounds could be achieved in a
modular fashion allowing the structural variations described to
be explored efficiently. The intermediates containing the pyrazolo-
pyrimidine core structure were prepared from pyrazole precursors
NH2
NC
H2N
HN
CN
CN
(a)
+
Cl
N
EtO
N
Cl
19
SH
N
20
N
N
(b)
N
Cl
O
O
(d)
(e)
22
H2N
H2N
(c)
Cl
HN
Cl
N
N
N
N
N
N
N
Cl
N
N
21
Cl
23
Scheme 2. Synthesis of pyrazolopyrimidine intermediates. Reagents and condi-
tions: (a) EtOH, reflux, 1 h, 80%; (b) H2SO4 (concn), rt, 0.5 h, 95%; (c) formamide,
microwave, 210 °C, 0.5 h, 92%; (d) THF, Lawesson’s reagent, reflux, 4 h, 67%; (e)
POCl3, reflux, 12 h, 81%.

P. Bonn et al. / Bioorg. Med. Chem. Lett. 22 (2012) 7302–7305
7305
O
O
S
Acknowledgments
(a)
Br
Br
N
OH
N
H
O
S
n-Bu
24
n-Bu
The authors thank Robert Garcia for the testing of compounds
against glucokinase and colleagues in the DMPK department for
the provision of secondary ADMET data.
N
S
N
25
N
H
n-Bu
N
(b)
SH
N
N
N
9
N
N
N
Cl
Cl
Supplementary data
22
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.bmcl.2012.
10.090.
Scheme 3. Reagents and conditions: (a) i—SOCl2, reflux 0.5 h; ii—DCM, DIPEA, 2-
amino-4-methylthiazole; (b) DMF, K2CO3, 3 h, rt, 39% over 3 steps.
References and notes
1. Tadayyon, M.; Smith, S. A. Expert Opin. Investig. Drugs 2003, 12, 307.
2. Turner, R. C.; Cull, C. A.; Frighi, V.; Holmann, R. J. Am. Med. Assoc. 2005, 1999,
281.
O
O
O
O
O
HO
TBDMSO
TBDMSO
(b)
(a)
OH
OTBDMS
N
N
H
3. Matschinsky, F. M. Diabetes 2002, 51, S394.
4. Matschinsky, F. M.; Magnuson, M. A. In Frontiers in Diabetes; Matschinsky, F. M.,
Magnuson, M. A., Eds.; Karger: Basel, 2004; Vol. 16, pp 1–17.
5. Gloyn, A.; Odili, S.; Zelent, D.; Buettger, C.; Castleden, H. A.; Steele, A. M.; Stride,
A.; Shiota, C.; Magnuson, M. A.; Lorini, R.; d’Annunzio, G.; Stanley, C. A.; Kwagh,
J.; van Schaftingen, E.; Veigha da Cunha, M.; Barbetti, F.; Dunten, P.; Han, Y.;
Grimsby, J.; Taub, R.; Ellard, S.; Hattersley, A. T.; Matschinsky, F. M. J. Biol. Chem.
2005, 280, 14105.
6. Leighton, B.; Atkinson, A.; Coghlan, M. P. Biochem. Soc. Trans. 2005, 33, 371.
7. Zelent, D.; Najafi, H.; Odili, S.; Buettger, C.; Weik-Collins, H.; Li, C.; Doliba, N.;
Grimsby, J.; Matschinsky, F. M. Biochem. Soc. Trans. 2005, 33, 306.
8. For comprehensive reviews of GKAs see: Fyfe, M. C. T.; Procter, M. P. Drugs
Future 2009, 34, 641; Matschinsky, F. M. Nat. Rev. Drug Disc. 2009, 8, 399;
Coghlan, M.; Leighton, B. Expert Opin. Investig. Drugs 2008, 17, 145.
9. McKerrecher, D.; Allen, J. V.; Bowker, S.; Boyd, S.; Caulkett, P. W. R.; Currie, G.
S.; Davies, C. D.; Fenwick, M. L.; Gaskin, H.; Grange, E.; Hargreaves, R. B.; Hayter,
B. R.; James, R.; Johnson, K. M.; Johnstone, C.; Jones, C. D.; Lackie, S.; Rayner, J.
W.; Walker, R. P. Bioorg. Med. Chem. Lett. 2005, 15, 2103.
28
27
26
O
(c)
O
HO
N
N
O
H
N
O
N
N
29
H
(d)
O
N
Cl
N
O
N
N
N
15
Cl
N
N
Cl
23
10. McKerrecher, D.; Allen, J. V.; Caulkett, P. W. R.; Donald, C. S.; Fenwick, M. L.;
Grange, E.; Johnson, K. M.; Johnstone, C.; Jones, C. D.; Rayner, J. W.; Walker, R. P.
Bioorg. Med. Chem. Lett. 2006, 16, 2705.
11. Pike, K. G.; Allen, J. V.; Caulkett, P. W. R.; Clarke, D. S.; Donald, C. S.; Fenwick, M.
L.; Johnson, K. M.; Johnstone, C.; McKerrecher, D.; Rayner, J. W.; Walker, R. P.;
Wilson, I. Bioorg. Med. Chem. Lett. 2011, 21, 3467.
Scheme 4. Reagents and conditions: (a) DMF, TBDMSCl, imidazole, rt, 20 h, 66%; (b)
i—DCM, oxalyl chloride, DMF (cat.), rt, 1 h; ii—DCM, pyridine, 2-amino-5-methyl-
pyridine, rt, 2 h, 40%; (c) THF, (n-Bu)4NFÁ3H2O, rt, 2 h, 69%; (d) THF, LHMDS, 50 °C,
20 h, 62%.
12. Waring, M. J.; Brogan, I. J.; Coghlan, M.; Johnstone, C.; Jones, H. B.; Leighton, B.;
McKerrecher, D.; Pike, K. G.; Robb, G. R. Med. Chem. Commun. 2011, 2, 771.
13. Waring, M. J.; Johnstone, C.; McKerrecher, D.; Pike, K. G.; Robb, G. R. Med. Chem.
Commun. 2011, 2, 775.
14. Waring, M. J.; Clarke, D. S.; Fenwick, M. D.; Godfrey, L.; Groombridge, S. D.;
Johnstone, C.; McKerrecher, D.; Pike, K. G.; Rayner, J. W.; Robb, G. R.; Wilson, I.
Med. Chem. Commun. 2012, 3, 1077.
carboxylate and subsequent acid chloride formation, followed by
interception with the amine to give 28. Fluoride mediated silyl
deprotection afforded the alcohol 29. The final SNAr reaction with
23 was carried out on the lithium alkoxide of 29 formed by treat-
ment with LHMDS to afford 15.
15. Kamata, K.; Mitsuya, M.; Nishimura, T.; Eiki, J. I.; Nagata, Y. Structure 2004, 12,
429.
16. Dunten, P.; Swain, A.; Kammlott, U.; Crowther, R.; Lukacs, C. M.; Levin, W.;
Reik, L.; Grimsby, J.; Corbett, W. L.; Magnuson, M. A.; Matschinsky, F. M.;
Grippo, J. F. In Frontiers in Diabetes; Matschinsky, F. M., Magnuson, M. A., Eds.;
Karger: Basel, 2004; Vol. 16, pp 145–154.
17. Hopkins, A. L.; Groom, C. R.; Alex, A. Drug Discovery Today 2004, 9, 430.
18. Ernster, L.; Siekevitz, P.; Palade, G. E. J. Cell Biol. 1962, 15, 541.
19. Taha, M. L.; Lazrek, H. B. Bull. Soc. Chim. Belg. 1995, 104, 647.
20. Waring, M. J.; Bennett, S. N. L.; Boyd, S.; Campbell, L.; Davies, R. D. M.; Martin,
N. G.; Robb, G. R.; Wilkinson, G. Med. Chem. Commun., in preparation.
21. Cheng, C. C.; Robins, R. K. J. Org. Chem. 1956, 21, 1240.
In conclusion, a new series of glucokinase activators has been
identified. Incorporation of an additional substituent in the R2 po-
sition and switching from a thioether to an ether linkage coupled
with exploration of the SAR in other regions of the molecule have
developed the initial screening hit 3 into compounds showing the
potential for optimisation into clinical candidates. The results of
subsequent work to further develop the series will be the subject
of further communication.20