Imidazo[1,2-a]pyridines: A potent and selective class of cyclin-dependent kinase inhibitors identified through structure-based hybridisation

Article abstract of DOI:10.1016/S0960-894X(03)00638-3

High-throughput screening identified the imidazo[1,2-a]pyridine and bisanilinopyrimidine series as inhibitors of the cyclin-dependent kinase CDK4. Comparison of their experimentally-determined binding modes and emerging structure-activity trends led to th

Full text of DOI:10.1016/S0960-894X(03)00638-3

Bioorganic & Medicinal Chemistry Letters 13 (2003) 3021–3026
Imidazo[1,2-a]pyridines: A Potent and Selective Class of
Cyclin-Dependent Kinase Inhibitors Identified Through
Structure-Based Hybridisation
Malcolm Anderson, John F. Beattie, Gloria A. Breault,* Jason Breed, Kate F. Byth,
Janet D. Culshaw, Rebecca P. A. Ellston, Stephen Green, Claire A. Minshull,
Richard A. Norman, Richard A. Pauptit, Judith Stanway, Andrew P. Thomas*
and Philip J. Jewsbury*
AstraZeneca, Alderley Park, Macclesfield, Cheshire SK10 4TG, UK
Received 12 March 2003; accepted 24 March 2003
Abstract—High-throughput screening identified the imidazo[1,2-a]pyridine and bisanilinopyrimidine series as inhibitors of the
cyclin-dependent kinase CDK4. Comparison of their experimentally-determined binding modes and emerging structure–activity
trends led to the development of potent and selective imidazo[1,2-a]pyridine inhibitors for CDK4 and in particular CDK2.
# 2003 Elsevier Ltd. All rights reserved.
Cancer has long been recognized as a disease of aber-
rant cellular proliferation. It is now known that proteins
that regulate proliferation are frequently mutated,
deleted or over-expressed in cancer cell lines, and that
this leads to the deregulated growth of the tumour cell.
Such proteins therefore represent potential targets for
therapeutic intervention.
The cyclin D1-CDK4/6-pRb signalling axis is a key
mediator of growth control in normal cells and a fre-
quent target for mutation in tumours, implicating
CDK4 as an important biochemical target for small
molecule inhibition of cell cycling. However, since cyclin
D-dependent kinases are dispensable in cells lacking
functional Rb, some tumours may be refractory to spe-
cific inhibition of CDK4. Recently, CDK2 has become
an attractive target with the hypothesis that pharmaco-
logical inhibition of CDK2 might selectively kill tumour
cells in which inactivation or loss of pRB serves to
deregulate E2F activity.3 The development of CDK
inhibitors with different selectivity profiles therefore
potentially offers the opportunity to treat a wide range
of tumour types.
The cyclin-dependent kinases (CDKs) are a family of
serine/threonine kinases that are important in control-
ling entry into and transition through each phase of the
cell cycle.1 Their cellular activity is tightly regulated
through a number of mechanisms:2 phosphorylation by
kinases and dephosphorylation by phosphatases; the
binding of specific activating proteins called cyclins; and
the binding of inhibitory peptides such as members of
the p16INK4A and p21CIP/p27KIP families. These
inhibitory peptides must be either sequestered or
destroyed, and the CDK appropriately phosphorylated
and bound to its cyclin, before the enzyme is fully
active. The synthesis and degradation of the activating
cyclins is tightly controlled; this, together with their
specificity for a particular CDK partner, causes first
one, and then another member of the CDK family to
become active and drive progression through the cell
cycle.
A number of groups have identified CDK inhibitors.4À18
In this paper we describe how comparison of the bind-
ing mode of a weak inhibitor identified through high-
throughput screening with that of a more potent series
described in earlier work17,18 led to the identification of
the imidazo[1,2-a]pyridines as a potent and selective
class of CDK inhibitors.
A high-throughput screen of the AstraZeneca com-
pound collection using an in vitro Scintillation Proxi-
mity Assay (SPA) to detect the inhibition of CDK4-
cyclin D catalysed [g-33-P]-phosphorylation of a GST-
Rb substrate15 identified the imidazo[1,2-a]pyridine 1 as
*Corresponding author. E-mail: philip.jewsbury@astrazeneca.com
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00638-3

3022
M. Anderson et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3021–3026
an inhibitor of CDK4 (IC50 8 mM). This prompted us to
initiate chemistry to explore the activity of this class of
compounds against CDK2 and CDK4.
Acetylation of the amino group was tolerated, but with
little increase in potency over the parent compound 1
(Table 1).
The most significant compound in this series was 2:
although still relatively weak, it was sufficiently potent
and soluble to attempt determination of the imi-
dazo[1,2-a]pyridine binding mode by X-ray crystal-
lography of the complex with CDK2.20 In these studies,
inactive CDK2 monomer was used as a structural sur-
rogate for activated CDK–cyclin complexes. While the
detailed use of a structural surrogate for molecular
design will be limited, studies demonstrate that a series’
bioactive conformation is typically conserved between
the activated CDK2–cyclinA complex and inactive
CDK2 monomer,21,22 and between CDK2 and other
kinases with low sequence homology though counter
examples do exist.22
Our initial synthetic goals were to improve potency and
demonstrate SAR by varying the substituents at the
amino group. The synthetic route to compound 119 is
shown in Scheme 1: reaction of 2-amino pyridine with
3-bromo 2,4-pentanedione gave 3-acetyl imidazo[1,2-a]-
pyridine. Condensation of the methyl ketone with DMF
dimethyl acetal gave aminopropenone A, which could
be cyclised with guanidine to give the desired 3-(pyr-
imid-4-yl)imidazo[1,2-a]pyridine product.
The crystal structure of CDK2 complexed with 2
revealed the imidazo[1,2-a]pyridine binding mode (Fig.
1). The hydrogen-bonding interactions with the protein
are between the pyrimidine N1 and Leu83 backbone
NH, the 2-amido NH and Leu83 backbone carbonyl O,
and the imidazo[1,2-a]pyridine N1 and the Lys33 amino
group. A water molecule bridges between the 2-amido
carbonyl O and the Asp86 side chain (Fig. 1a). The first
two hydrogen bonds are also found in the bisanilino-
pyrimidine binding mode (Fig. 1b) and are consistent
with the accepted kinase purine-mimetic pharmaco-
phore. The third had not been observed previously in
our studies, but an equivalent interaction is found
between an imidazole ring and the analogous lysine in
the complexes of pyridinyl-imidazoles with p38.23
An earlier study had identified the bisanilinopyrimidines
as potent and selective CDK inhibitors.17,18 A compar-
Scheme 1. Synthesis of imidazo[1,2-a]pyridine 1.
Figure 1. Crystal structure of CDK2 complexed with 2.24 (a) Final 2Fo-Fc electron density for the inhibitor 2 (1s level). (b) Superposition (by
matching Ca-carbon atoms of residues 80–87) of 2 and a bisanilinopyridine.17 The bisanilinopyrimidine is modelled as a racemic mixture but this
portion of the molecule is poorly ordered in the crystal structure.17 The figures were prepared using Bobscript and Raster3D.32,33

M. Anderson et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3021–3026
3023
Table 1. Structures and enzyme activity for amino-substituted imi-
dazo[1,2-a]pyridines
SAR in the bisanilinopyrimidine series suggested that
CDK2 potency would be improved by a para-sulphon-
amoyl substituted aniline; whereas CDK4 selectivity
would require a para-(3-dimethylamino-2-hydroxy-
propoxy) substituted aniline. The synthesis of com-
pounds 6 and 7 was prioritised.
Compd
CDK4
IC50, mMa
CDK2
IC50, mMa
R
1
2
3
4
8
8.9
40
>10
4
2.9
95
>10
H
COCH3
COCF3
CONHPh
Synthesis of compound 634 began with alkylation of 4-
nitrophenol with epibromohydrin. Opening of the
epoxide with dimethylamine followed by hydrogenation
of the nitro group gave the required aniline. Reaction of
this aniline with cyanamide gave the corresponding
guanidine, which was reacted in situ with the amino-
propenone of the imidazo[1,2-a]pyridine B. Compound
735 was prepared by reaction of the same imidazo[1,2-
a]pyridine B with phenylguanidine followed by chloro-
sulphonation of the aniline and reaction with ammonia
in methanol (Scheme 2, Table 2).
aAverage of at least two measurements; enzyme protocol.15
ison of the bisanilinopyrimidine and imidazo[1,2-a]pyr-
idine binding modes (Fig. 1b) suggested a hybridisation
strategy: the addition of a 2-aniline group to the pyr-
imidine ring to replace the 2-amido group. Emerging
Addition of the 2-anilino group (5) led to only a small
increase in potency against CDK4 but a 100-fold
increase against CDK2. Varying the anilino para-sub-
stituent led to enhanced activity against CDK4 (6), or
particularly CDK2 (7). This trend was analogous to
that found in the 2,4-bisanilinopyrimidine series (8 and
9), suggesting the binding modes of the anilino groups
were similar in the two series. X-ray structure determi-
nation of the complexes of 6 and a 2-des methyl ana-
logue36 of 7 with CDK2 confirmed this interpretation
(Fig. 2).
Table 2. Structures and enzyme activity for 2-anilino imidazo[1,2-a]-
pyridines (IP) and their 2,4-bisanilinopyrimidine (BAP) analogues
Compd Series
R
CDK4
CDK2
IC50, mMa IC50, mMa
1
5
6
7
8
9
IP
IP
IP
IP
8
4
H
3.6
0.15
2.5
2
0.036
0.032
<0.003
32
OCH2CH(OH)CH2N(Me)2
SO2NH2
BAP OCH2CH(OH)CH2N(Me)2
BAP SO2NH2
2.7
0.006
Scheme 2. Synthesis of 2-anilino substituted imidazo[1,2-a]pyridines 6
and 7.
aAverage of at least two measurements; enzyme protocol.15

3024
M. Anderson et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3021–3026
The hydrogen bonding interactions of the pyrimidine
core to the peptide backbone remain the same in these
structures as those seen in Figure 1a. In the CDK2
complex with 6, the direct interaction between the imi-
dazo[1,2-a]pyridine ring and the Lys33 amino group is
replaced by a water-mediated interaction with the
Asp145 side chain, possibly due to the carboxylation of
the lysine side chain (Fig. 2a).
In these CDK2 structures, the hydrophobic anilino
rings lie against a hydrophobic surface formed by the
Phe82, Leu83, Leu134 and Ile10 side chains, and against
Figure 2. Crystal structures37 of CDK2 complexed with 6 and a 2-des methyl analogue of 7. (a) Final 2Fo-Fc electron density for the inhibitor 6 (1s
level). In this structure, the Lys33 side chain has become carboxylated. 6 is a racemate (one enantiomer shown) but the molecule is poorly ordered
beyond the hydroxyl. (b) Initial 2Fo-Fc electron density for the inhibitor 2-des methyl 7 (1.3s level). (c) Superposition (matching Ca-carbon atoms
of residues 80–87) of 2 and 6. The figures were prepared using Bobscript and Raster3D.32,33

M. Anderson et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3021–3026
3025
the peptide backbone of residues 84–86. The desolvation
of these hydrophobic surfaces possibly accounts for the
increased potency found for compound 5. The smaller
increase in CDK4 activity may be due to differences in
the hydrophobic residues packing the aniline ring:
Phe82 in CDK2 is replaced by His95 in CDK4, and
Leu83 by Val96.
Grafstrom, R.; Harrison, B. A.; Harrison, K.; Akamike, E.;
Markwalder, J. A.; Nakano, Y.; Seitz, S. P.; Sharp, D. M.;
Trainor, G. L.; Sielecki, T. M. Bioorg. Med. Chem. Lett. 2001,
11, 2209.
8. Honma, T.; Hayashi, K.; Aoyama, T.; Hashimoto, N.;
Machida, T.; Fukasawa, K.; Iwama, T.; Ikeura, C.; Suzuki-
Takahashi, I. J. Med. Chem. 2001, 44, 4615.
9. Fry, D. W.; Bedford, D. C.; Harvey, P. H.; Fritsch, A.;
Keller, P. R.; Wu, Z.; Dobrusin, E.; Leopold, W. R.; Fattaey,
A.; Garrett, M. D. J. Biol. Chem. 2001, 276, 16617.
10. Soni, R.; O’Reilly, T.; Furet, P.; Muller, L.; Stephan, C.;
Zumstein-Mecker, S.; Fretz, H.; Fabbro, D.; Chaudhuri, B. J.
Natl. Cancer Inst. 2001, 93, 436.
11. Honma, T.; Yoshizumi, T.; Hashimoto, N.; Hayashi, K.;
Kawanishi, N.; Fukasawa, K.; Takaki, T.; Ikeura, C.; Ikuta,
M.; Suzuki-Takahashi, I.; Takashi, H.; Nishimura, S.; Mor-
ishima, H. J. Med. Chem. 2001, 44, 4628.
12. Soni, R.; Muller, L.; Furet, P.; Schoepfer, J.; Stephan, C.;
Zumstein-Mecker, S.; Fretz, H.; Chaudhuri, B. Biochem. Bio-
phys. Res. Commun. 2000, 275, 877.
13. Jeong, H.-W.; Kim, M.-R.; Son, K.-H.; Young Han, M.;
Ha, J.-H.; Garnier, M.; Meijer, L.; Kwon, B.-M. Bioorg. Med.
Chem. Lett. 2000, 10, 1819.
14. Ryu, C.-K.; Kang, H.-Y.; Lee, S. K.; Nam, K. A.; Hong,
C. Y.; Ko, W.-G.; Lee, B.-H. Bioorg. Med. Chem. Lett. 2000,
10, 461.
15. Breault, G. A., Pease, J. E. PCT Int. Application WO
0012485 A1 20000309.
16. Davies, T. G.; Bentley, J.; Arris, C. E.; Boyle, F. T.; Cur-
tin, N. J.; Endicott, J. A.; Gibson, A. E.; Golding, B. T.;
Griffin, R. J.; Hardcastle, I. R.; Jewsbury, P.; Johnson, L. N.;
Mesguiche, V.; Newell, D. R.; Noble, M. E. M.; Tucker, J. A.;
Wang, L.; Whitfield, H. J. Nat. Struct. Biol. 2002, 9, 745.
17. Beattie, J. F.; Breault G. A.; Ellston, R. P. A.; Green S.;
Jewsbury, P. J.; Midgely C. J.; Naven, R. T.; Pauptit, R.A.;
Tucker, J. A.; Pease, J. E. Bioorg. Med. Chem. Lett. See paper
in this issue. doi: 10.1016/S0960-894X(03)00202-6.
18. Breault, G. A.; Ellston, R. A; Green, S.; James, S. R.;
Jewsbury, P. J.; Midgley, C. J.; Pauptit, R. A.; Minshull, C.
A.; Tucker, J. A.; Pease, J. E. Bioorg. Med. Chem. Lett. See
paper in this issue. doi: 10.1016/S0960-894X(03)00203-8.
19. Compd 1: NMR (DMSO-d6) d 2.62 (d, 1H), 6.74 (s, 2H),
6.85 (d, 1H), 7.01 (td, 1H), 7.4 (m, 1H), 7.60 (d, 1H), 8.30 (d,
1H), 9.85 (d, 1H); MS (ES+) 226 [MH]+.
The para-sulphonamide group in Figure 2b forms
hydrogen bonds with the Asp86 backbone NH and with
its carboxylic side chain. Davies et al. have described
how the aniline packing and sulphonamide binding are
optimally arranged in their NU6102 series.16 Similar
interaction geometries are seen in our structure and may
account for the increased potency seen for 7 over 5.
Asp86 is conserved between CDK2 and CDK4 suggest-
ing that the improved CDK2 selectivity seen for 7
results from differences in the residues packing the ani-
line ring.
A number of basic residues lining the CDK2 binding
site are replaced by acidic or neutral residues in CDK4.
In particular Lys89, which is known to coordinate an
acidic substituent in purvalanol B,40 is replaced by
Thr102 in CDK4. The lack of an equivalent lysine side
chain at this position, and other differences at the sur-
face of the binding site (CDK2/Lys9->CDK4/Glu11
and His84->Asp97) may account for the preference for
a basic para-substituent by CDK4 (5 cf 6).
In conclusion, knowledge of the imidazo[1,2-a]pyridine
binding mode and its comparison with the bisanilino-
pyridimidine SAR led to the development of the imi-
dazo[1,2-a]pyridine series of CDK inhibitors. Examples
from this series have been identified as potent and
selective inhibitors of CDK4, while other compounds
show particularly potent and selective inhibition of
CDK2.
Acknowledgements
20. Schulze-Gahmen, U.; DeBondt, H. L.; Kim, S.-H. J. Med.
Chem. 1996, 39, 4540.
21. Davies, T. G.; Tunnah, P.; Meijer, L.; Marko, D.; Eisen-
brand, G.; Endicott, J. A.; Noble, M. E. M. Structure 2001, 9,
389.
We are grateful to Drs. J. Endicott and M. Noble, Uni-
versity of Oxford, for useful discussions and initial supply
of reagents.
22. Our unpublished work.
23. Tong, L.; Pav, S.; White, D.; Rogers, S.; Crane, K. M.;
Cywin, C. L.; Brown, M. L.; Pargellis, C. A. Nat. Struct. Biol.
1997, 4, 311.
References and Notes
24. Protein and crystals were obtained according to estab-
lished procedures.25,26 Diffraction data were collected on
beamline 9.6 at SRS, Daresbury, at 100 K. Data processing,
data reduction and structure solution by molecular replace-
ment were carried out using programs from the CCP4 suite.27
Compound 2 was modelled into electron density using
QUANTA.28 The protein complex model was refined using
CNX29 and Refmac5,30 and the final structure31 has been
deposited in the Protein Data Bank with deposition code 1ioq
together with structure factors and detailed experimental con-
ditions.
25. Lawrie, A. M.; Noble, M. E.; Tunnah, P.; Brown, N. R.;
Johnson, L. N.; Endicott, J. A. Nat. Struct. Biol. 1997, 4, 796.
26. Legraverend, M.; Tunnah, P.; Noble, M.; Ducrot, P.;
Ludwig, O.; Grierson, D. S.; Leost, M.; Meijer, L.; Endicott,
J. J. Med. Chem. 2000, 43, 1282.
1. Sherr, C. J. Science 1996, 274, 1672.
2. Shapiro, G. I.; Harper, J. W. J. Clin. Invest. 1999, 104,
1645.
3. Chen, Y.-N. P.; Sharma, S. K.; Ramsey, T. M.; Jiang, L.;
Martin, M. S.; Baker, K.; Adams, P. D.; Bair, K. W.; Kaelin,
W. G., Jr. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 4325.
4. Webster, K. R.; Kimball, S. D. Emerg. Drugs 2000, 5, 45.
5. Kimball, S. D.; Webster, K. R. In Annual Reports in
Medicinal Chemistry; Doherty, A. M., Ed.; Academic: San
Diego, 2001: Vol. 36, p 139, and references therein.
6. Ikuta, M.; Kamata, K.; Fukasawa, K.; Honma, T.;
Machida, T.; Hirai, H.; Suzuki-Takahashi, I.; Hayama, T.;
Nishimura, S. J. Biol. Chem. 2001, 276, 27548.
7. Carini, D. J.; Kaltenbach, R. F.; Liu, J.; Benfield, P. A.;
Boylan, J.; Boisclair, M.; Brizuela, L.; Burton, C. R.; Cox, S.;

3026
M. Anderson et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3021–3026
27. CCP4 Acta Crystallogr. 1994, D50, 760.
28. Quanta2000, Accelrys.
29. CNX version 2000.1, Accelrys.
30. Refmac5, CCP4.
31. Crystallographic statistics for 2 are: space group P212121,
unit cell 54.6, 72.7, 72.9 A, resolution 2.31 A, 12637 unique
reflections from 60244 observations give 93.8% completeness
with Rmerge=13.2% and mean I/s(I) of 8.0. The final model
containing 2099 protein, 32 water and 20 inhibitor atoms has
an R-factor of 23.3% (Rfree using 5% of data is 27.9%). Mean
temperature factor for protein is 36.3 and for ligand is 32.4 A2.
32. Esnouf, R. J. Mol. Graphics 1997, 156, 132.
lished procedures.25,26 Diffraction data were collected on
beamline 9.6 at SRS, Daresbury, at 100 K. Data processing,
data reduction and structure solution by molecular replace-
ment were carried out using programs from the CCP4 suite.27
Compound 6 is a racemic mixture making the conformation
ill-defined beyond the hydroxyl. A single isomer was modelled
into electron density using QUANTA.28 The protein complex
models was refined using CNX29 and Refmac5,30 and the final
structures38,39 have been deposited in the Protein Data Bank
with deposition codes 1oir and 1oit together with structure
factors and detailed experimental conditions.
38. Crystallographic statistics for 6 are: space group P212121,
unit cell 53.0, 70.7, 72.9 A, resolution 1.91 A, 20394 unique
reflections from 95201 observations give 84.8% completeness
with Rmerge=5.2% and mean I/s(I) of 21.7. The final model
containing 2284 protein, 124 water and 31 inhibitor atoms has
an R-factor of 19.7% (Rfree using 5% of data is 23.7%). Mean
temperature factor for protein is 14.6 and for ligand is 22.1 A2.
39. Crystallographic statistics for 2-des methyl analogue of 7
are: space group P212121, unit cell 53.5, 72.2, 72.2 A, resolu-
tion 1.60 A, 33885 unique reflections from 179548 observa-
tions give 80.7% completeness with Rmerge=3.5% and mean I/
s(I) of 17.8. The final model containing 2059 protein, 272
water and 26 inhibitor atoms has an R-factor of 26.3% (Rfree
using 5% of data is 25.5%). Mean temperature factor for
protein is 19.4 and for ligand is 19.6 A2.
33. Merritt, E. A.; Murphy, M. E. P. Acta Crystallogr. 1994,
D50, 869.
34. Compd 6: NMR (DMSO-d6) d 2.35 (s, 6H), 2.40–2.63 (m,
2H), 3.82–4.02 (m, 3H), 6.90 (d, 2H), 7.06 (dd, 1H), 7.30 (d,
1H), 7.50 (dd, 1H), 7.59 (s, 2H), 7.74 (d, 1H), 8.38 (d, 1H),
8.58 (s, 1H), 9.42 (s, 1H); MS (ES+) 405 [MH]+.
35. Compd 7: NMR (DMSO-d6) d 2.64 (s, 3H), 7.05 (dd, 1H),
7.15–7.20 (m, 3H), 7.44 (dd, 1H), 7.64 (d, 1H), 7.74 (d, 2H),
7.92 (d, 2H), 8.68 (d, 1H), 9.75 (d, 1H); MS (ES+) 381
[MH]+.
36. Unexpectedly, soaking with 7 cracked the cdk2 crystals
whereas this did not happen with the 2-des methyl analogue.
No adverse contacts between the methyl group and the protein
could be identified by modelling. The synthetic route and
characterisation of the 2-des methyl imidazo[1,2-a]pyridine
series is the subject of a paper in preparation by the authors;
details will be made available on request.
40. Gray, N. S.; Wodicka, L.; Thunnissen, A. W.; Norman,
T. C.; Kwon, S.; Espinoza, F. H.; Morgan, D. O.; Barnes, G.;
LeClerc, S.; Meijer, L.; Kim, S.; Lockhart, D. J.; Schultz, P. G.
Science 1998, 281, 533.
37. Protein and crystals were obtained according to estab-