Structure-guided design of pyrazolo[1,5-a]pyrimidines as inhibitors of human cyclin-dependent kinase 2

Article abstract of DOI:10.1016/j.bmcl.2004.12.073

The protein structure guided design of a series of pyrazolo[1,5-a] pyrimidines with high potency for human cyclin-dependent kinase 2 (CDK2) is described. Some examples were shown to inhibit the growth of human colon tumour cells, were equipotent for CDK1 and were selective against GSK-3β and other kinases.

Full text of DOI:10.1016/j.bmcl.2004.12.073

Bioorganic & Medicinal Chemistry Letters 15 (2005) 863–867
Structure-guided design of pyrazolo[1,5-a]pyrimidines
as inhibitors of human cyclin-dependent kinase 2
Douglas S. Williamson,* Martin J. Parratt, Justin F. Bower, Jonathan D. Moore,
Christine M. Richardson, Pawel Dokurno, Andrew D. Cansfield, Geraint L. Francis,
Richard J. Hebdon, Rob Howes, Philip S. Jackson, Andrea M. Lockie, James B. Murray,
Claire L. Nunns, Jenifer Powles, Alan Robertson, Allan E. Surgenor
and Christopher J. Torrance
Vernalis (R&D) Ltd, Granta Park, Great Abington, Cambridge CB1 6GB, United Kingdom
Received 12 November 2004; revised 17 December 2004; accepted 23 December 2004
Available online 19 January 2005
Abstract—The protein structure guided design of a series of pyrazolo[1,5-a]pyrimidines with high potency for human cyclin-depen-
dent kinase 2 (CDK2) is described. Some examples were shown to inhibit the growth of human colon tumour cells, were equipotent
for CDK1 and were selective against GSK-3b and other kinases.
ꢀ 2005 Elsevier Ltd. All rights reserved.
Cyclin-dependent kinases (CDKs) play a central role in
mammalian cell division.1 Recent knock-out mouse
experiments, however, suggest significant functional
redundancy within this family of kinases.2,3 Many ques-
tions have therefore been raised regarding the optimal
selectivity profile for CDK inhibitors aimed at yielding
safe and effective anti-cancer agents. Pragmatically,
absolute selectivity has yet to be achieved and most ef-
forts have focused on using CDK2 as a template, given
its ready applicability to protein structure guided drug
design.4
the related but therapeutically antagonistic glycogen
synthase kinase-3b (GSK-3b) at concentrations only
modestly in excess of those required to inhibit CDK2.8
BMS-387032 (3) is also subject to PGP-mediated drug
resistance.9 Given these issues, there is still scope for fur-
ther research towards finding the ideal CDK inhibitor
for cancer therapy. We describe herein the discovery
of a series of compounds with high potency for
CDK2, which in some cases were shown to inhibit the
growth of human colon tumour cells and were equipo-
tent against CDK1, a kinase that partially overlaps in
function with CDK2.10 Many examples were also shown
to be selective against GSK-3b and other kinases.
Inhibitors with varying CDK selectivity profiles are now
being evaluated in clinical trials. Flavopiridol (1)5 (phase
I/II) displays modest selectivity for CDKs over other ki-
nases and inhibits many members within the CDK fam-
ily, including those without cell-cycle roles. Roscovitine/
CYC-202 (2)6 (phase I/II) and BMS-387032 (3)7 (phase
I) represent bona fide CDK inhibitors currently in clini-
cal development. Both compounds 2 and 3 inhibit
CDK2 more effectively than CDK1, roscovitine (2) has
limited potency on cells and BMS-387032 (3) inhibits
Ph
N
NH
OH
O
N
N
HO
HO
N
N
H
N
Cl
1
2
OH
O
NH
N
H
N
Keywords: Pyrazolo[1,5-a]pyrimidine; CDK2; Cyclin-dependent
kinase.
S
S
O
N
O
*
Corresponding author. Tel.: +44 (0)1223 895555; fax: +44 (0)1223
895556; e-mail: d.williamson@vernalis.com
3
0960-894X/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.12.073

864
D. S. Williamson et al. / Bioorg. Med. Chem. Lett. 15 (2005) 863–867
The pyrazolo[1,5-a]pyrimidine 4 arose as a hit from a
high-throughput screen of a commercially available ki-
nase-directed library. It was found to have an IC50 of
1.8 lM against CDK2, but inhibited GSK-3b with sim-
ilar potency (IC50 0.96 lM).11,12 Compound 4 was also
shown to inhibit the growth of human colon tumour
cells (HCT116) in a sulforhodamine B (SRB) growth
inhibition assay13 (GI50 7.9 lM).14 Compound 4 is
structurally similar to roscovitine (2), and other pur-
ine-based competitor compounds such as NU-6102
(5)15 and H717 (6).16 It was decided to explore the
SAR around the 3-, 5- and 7-positions of the pyraz-
olo[1,5-a]pyrimidine template, with a view to optimizing
potency and selectivity for CDK2 and tumour cell
growth inhibition.
N
O
N
O
N
N
N
N
N
N
N
H
HN
S
HN
HN
7
N
3
N
5
Br
N
O
NH2
H2N
O
5
4
6
The 7-Cl of 5,7-dichloropyrazolo[1,5-a]pyrimidines of
type 717 underwent facile displacement with a range of
anilines to afford compounds of type 8 (Scheme 1).
The 5-Cl could be displaced with aliphatic amines by
heating with microwaves in a Smith SynthesizerTM at
180 ꢁC in 1,4-dioxane and acetonitrile in the presence
of triethylamine, resulting in compounds of type 9
(X = NH). Alternatively, ethers of type 9 (X = O) could
be synthesized by treating a 1,4-dioxane–acetonitrile
solution of the chloride 8 with an alcohol in the presence
of sodium hydride and triethylamine, and then heating
with microwaves to 180 ꢁC as described previously.
Figure 1. Compound 8a bound to inactive monomeric human CDK2,
with key kinase donor–acceptor–donor motif interactions indicated,
and potential for elaboration to improve contact with Phe80 side
chain.
bonyl, a pyrazole N and Leu83 backbone NH, and the
aromatic C2-H and the Glu81 backbone carbonyl was
observed. This type of interaction motif is a well-recog-
nized feature of a number of ATP competitive kinase
ligands, and the importance of C–HÆÆÆO interactions is
now documented.20 In addition, a hydrogen bond be-
tween a sulfone oxygen and Asp86 [analogous to the sul-
fonamide O atom of NU-6102 (5)] was present, and the
compound showed good steric complimentarity with the
active site cleft. Overlaying this structure with that of
roscovitine (2) bound to CDK2 indicated that elabora-
tion at the 3-position of the pyrazolopyrimidine ring
would allow interaction with the Phe80 residue. Com-
pound 8b incorporated an isopropyl group at the 3-po-
sition and, encouragingly, this increased the CDK2
potency approximately 10-fold to 0.23 lM, and reduced
GSK-3b potency to 11 lM. This resulted in a molecule,
which was now almost 50-fold selective over GSK-3b,
more potent than roscovitine (2) (CDK2 IC50 0.65 lM,
GSK-3b IC50 34 lM)8 and more selective over GSK-
3b than NU-6102 (5) (CDK2 IC50 0.045 lM, GSK-3b
IC50 0.040 lM).8
The sulfonylphenylamino analogue 8a was initially syn-
thesized, and this inhibited CDK2 with an IC50 of
2.0 lM and had a similar potency (IC50 3.1 lM) against
GSK-3b. An X-ray crystallographic structure18 was ob-
tained, following co-crystallization of the compound
with inactive monomeric human CDK2 prepared in-
house19 (Fig. 1). The key ꢀdonor–acceptor–donorꢁ bind-
ing motif between the 7-NH and Leu83 backbone car-
R2
R2
Cl
N
NH
NH
N
N
N
N
N
N
(i)
(ii)
R3
Cl
Cl
N
X
N
R1
R1
R1
9
7
8
Replacement of the methyl sulfone with a dimethylsulf-
onamide and incorporation of the trans-4-amin-
ocyclohexylamino moiety, as found in H717 (6), into
the 5-position of the pyrazolopyrimidine ring afforded
compound 9a. This compound had improved potency
Scheme 1. Reagents and conditions: (i) R2NH2, EtOH, 70–90%; (ii)
(for X = NH) R3NH2, Et3N, MeCN, 1,4-dioxane, lWave @ 180 ꢁC,
3 h, 18–82%; (for X = O) R3OH, Et3N, NaH, Et3N, MeCN, 1,4-
dioxane, lWave @ 180 ꢁC, 3 h, 7–20%.

D. S. Williamson et al. / Bioorg. Med. Chem. Lett. 15 (2005) 863–867
865
for CDK2 (IC50 0.059 lM), possibly as a result of an
additional interaction between the charged amino group
attached to the cyclohexyl ring, which interacts via a salt
bridge to Asp145, and a hydrogen bond with Asn132, as
observed in an X-ray structure (Fig. 2).18 In addition,
the crystal structure of compound 9a in complex with
CDK2 clearly shows the improved interaction between
Phe80 and the isopropyl group, in an analogous fashion
to the roscovitine complex. These changes also benefited
the selectivity against GSK-3b (IC50 40 lM) and the cel-
lular activity. Compound 9a had a GI50 of 0.58 lM, a
considerable improvement in comparison to that of ros-
covitine (2) (GI50 15 lM).8 In terms of broader specific-
ity versus kinases, 9a inhibited CDK5 with equal
potency to CDK2, but was 3-fold selective versus
CDK7, >20-fold selective versus CDK6 and >50-fold
selective versus Chk1, PKA, IKKa, MEK1, CK2,
JNK1a, PKCa, MAPK1 and PDGFR. High concentra-
tions of 9a (32 lM), however, strongly inhibited CK1,
Rsk1, Chk2 and Flt3.
Figure 2. Compound 9a bound to inactive monomeric human CDK2,
showing the kinase motif interaction and improved steric complemen-
tarity with the Phe80 side chain. The salt bridge from the ligand
primary amine to Asp145 is also highlighted.
Table 1 contains the compounds synthesized, along with
details of potency for CDK2, selectivity against GSK-3b
Table 1. Enzyme activity (CDK2 and GSK-3b) and cell-based inhibition in HCT116 (colon) cell line for pyrazolo[1,5-a]pyrimidines 9a–na
R2
NH
N
N
R3
X
N
R1
No.
X
R1
iPr
R2
R3
CDK2 IC50 lMa,b
GSK-3b IC50 lMc
HCT 116 GI50 lM
O
S
H2N
O
9a
NH
0.059
0.004
0.007
0.008
0.010
0.061
0.092
0.017
2.1
40
0.58
Me2N
O
S
H2N
H2N
H2N
H2N
O
Me2N
9b
9c
9d
9e
9f
NH
NH
NH
NH
NH
NH
O
22
0.46
5.1
0.54
11
O
S
O
Me2N
CN
Cl
O
S
O
Me2N
2.7
O
S
O
Me2N
Br
3.4
0.91
1.2
Et
HN
O
S
O
Me2N
iPr
iPr
iPr
iPr
>50
>50
28
O
S
Et
O
Me2N
N
9g
9h
9i
16
Et
O
S
H2N
O
Me2N
0.71
3.4
O
S
O
Me2N
NH
>50
H2N
O
S
O
Me2N
9j
O
O
iPr
iPr
Me2CHCH2
1.2
1.2
50
12
7.6
O
S
O
Me
9k
Me2CHCH2
7.8
(continued on next page)

866
D. S. Williamson et al. / Bioorg. Med. Chem. Lett. 15 (2005) 863–867
Table 1 (continued)
No.
X
R1
Br
R2
R3
CDK2 IC50 lMa,b
GSK-3b IC50 lMc
HCT 116 GI50 lM
O
S
Cl
O
H2N
9l
NH
0.002
8.7
0.34
Me2N
H2N
H2N
S
O
9m
9n
NH
NH
Cl
Br
0.011
0.016
4.0
5.3
0.12
0.08
S
O
SO2Ph
a All IC50 and GI50 values are the mean of at least two determinations, and are rounded to two significant figures where appropriate.
b [ATP] 100 lM (Km 50 lM).
c [ATP] 10 lM (Km 10 lM).
and inhibition of tumour cell growth. All compounds
were also screened against Chk1 and PKA, and shown
to be >50-fold selective for CDK2.
Compounds 9m (GI50 0.12 lM) and 9n (GI50
0.08 lM), which incorporated an arylsulfonylphenyl-
amino substituent at the 7-position and a 3-Cl or Br
atom, respectively, had the greatest cellular activity.
Replacement of the 3-isopropyl substituent with cyclo-
pentyl (9b), nitrile (9c), chloro (9d) or bromo (9e) all en-
hanced the enzyme potency significantly, but did not
enhance cellular activity. We gathered additional infor-
mation on the selectivity of compound 9e, which was
>12-fold less potent versus CDK5 and >30-fold less po-
tent versus CDK1, CDK6, CDK7, Abl1, IKKb, MEK1,
FRFR-3, p70S6Kinase, PKCc, c-Raf, CK1d, MAPK1,
MAPKAP2, PAK2 and NEK2. Compound 9e, how-
ever, retained the ability of compound 9a to inhibit
Chk2 (90% inhibition @ 300 nM).
Selected compounds were also assayed for their ability
to inhibit human CDK1. Compounds 9a and 9n inhib-
ited CDK1 and CDK2 with approximately equal po-
tency, whilst compound 9e was around 60 times less
potent versus CDK1. Compounds 9a, 9e and 9n were
capable of downregulating phosphorylation on threo-
nine 821 of the retinoblastoma protein in HCT116 cells
(a site phosphorylated by CDK2), when added at a low
multiple (4·) of the GI50. Moreover, at 1 · GI50, com-
pound 9n caused an accumulation of cells in the G2/M
phase of the cell cycle, a result consistent with the ex-
pected mode of action for a dual CDK1/CDK2 inhibi-
tor. The effects of our compounds on cells will be fully
described in a forthcoming paper.
A number of modifications were carried out on the pri-
mary amine group of compound 9a, which was consid-
ered to be a potential metabolic liability. These
generally retained the enzyme potency, but reduced the
cellular activity dramatically. For example, mono-ethy-
lation of the primary amine21 to give compound 9f re-
sulted in the same potency against CDK2 (0.061 lM),
but the GI50 dropped to 1.2 lM. Diethylation to give
9g resulted in a slight drop in CDK2 potency
(0.092 lM), but the GI50 fell to 16 lM.
Compounds 9a and 9f were administered intravenously
and orally to male SD rats at a dose of 1 mg/kg. These
compounds displayed poor bioavailability and were
not progressed.
In conclusion, knowledge of the binding mode of a series
of pyrazolo[1,5-a]pyrimidines to CDK2 was gained by
X-ray structures, which assisted the design of highly po-
tent inhibitors for this kinase. Some examples were
shown to inhibit the growth of human colon tumour
cells, were equipotent against CDK1 and were selective
against GSK-3b and other kinases.
The linker at the 5-position was also varied to an O
atom and, in comparison with 9a, compound 9h dis-
played a marginally increased potency against CDK2
(0.017 lM), but made no significant difference to the cel-
lular potency (GI50 0.71 lM).
Minor changes, such as moving the primary amine of
compound 9a to the 3-position of the cyclohexyl ring
(9i), or total replacement with a 2-methylpropanyloxy
side chain (9j), were detrimental to CDK2 and cellular
potency. The dimethyl sulfonamide group of compound
9j could be replaced with a methyl sulfone (9k) with no
significant effect on CDK2 potency or cellular activity.
Acknowledgements
The authors would like to thank Peter Kierstan, Joanne
Wayne and Angela Wheatley for running enzyme and
cell-based assays, Steven Athwal and Mhairi Donohoe
for production of CDK2, Adam Hold and Heather
Simmonite for analytical support, Angela Merrett and
Tony Padfield for pharmacokinetic studies and
Rod Hubbard for his valued comments on this
manuscript.
Incorporation of a chlorine atom into the 2-position of
the benzenesulfonamide ring of 9e gave the most potent
CDK2 inhibitor 9l (CDK2 IC50 2 nM, GI50 0.34 lM).

D. S. Williamson et al. / Bioorg. Med. Chem. Lett. 15 (2005) 863–867
867
13. Skehan, P.; Storeng, R.; Scudiero, D.; Monks, A.;
McMahon, J.; Vistica, D.; Warren, J. T.; Bokesch, H.;
Kenney, S.; Boyd, M. R. J. Natl. Cancer Inst. 1990, 82,
1107–1112.
References and notes
1. Murray, A. W. Cell 2004, 116, 221–234.
´
2. Mendez, J. Cell 2003, 114, 398–399.
14. GI50 is the inhibitor concentration which reduced tumour
cell growth by 50% over 72 h.
3. Pagano, M.; Jackson, P. K. Cell 2004, 118, 535–538.
4. Davies, T. G.; Pratt, D. J.; Endicott, J. A.; Johnson, L. N.;
Noble, M. E. M. Pharmacol. Ther. 2002, 93, 125–133.
5. Zhai, S.; Senderowicz, A. M.; Sausville, E. A.; Figg, W. D.
Ann. Pharmacother. 2002, 36, 905–911.
15. Davies, T. G.; Bentley, J.; Arris, C. E.; Boyle, F. T.;
Curtin, 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, 745–749.
16. Dreyer, M. K.; Borcherding, D. R.; Dumont, J. A.; Peet,
N. P.; Tsay, J. T.; Wright, P. S.; Bitonti, A. J.; Shen, J.;
Kim, S.-H. J. Med. Chem. 2001, 44, 524–530.
17. Novinson, T.; Bhooshan, B.; Okabe, T.; Revankar, G. R.;
Robins, R. K.; Senga, K.; Wilson, H. R. J. Med. Chem.
1976, 19, 512–516.
18. The PDB entry codes are 1Y8Y (CDK2/8a) and 1Y91
(CDK2/9a) Berman, H. M.; Westbrook, J.; Feng, Z.;
Gilliland, G.; Bhat, T. N.; Weissig, H.; Shindyalov, I. N.;
Bourne, P. E. Nucleic Acids Res. 2000, 28, 235–
242.
6. Meijer, L.; Raymond, E. Acc. Chem. Res. 2003, 36, 417–
425.
7. Misra, R. N.; Xiao, H.-y.; Kim, K. S.; Lu, S.; Han, W.-C.;
Barbosa, S. A.; Hunt, J. T.; Rawlins, D. B.; Shan, W.;
Ahmed, S. Z.; Qian, L.; Chen, B.-C.; Zhao, R.; Bednarz, M.
S.; Kellar, K. A.; Mulheron, J. G.; Batorsky, R.; Roongta,
U.; Kamath, A.; Marathe, P.; Ranadive, S. A.; Sack, J. S.;
Tokarski, J. S.; Pavletich, N. P.; Lee, F. Y. F.; Webster, K.
R.; Kimball, S. D. J. Med. Chem. 2004, 47, 1719–1728.
8. In-house data.
9. Kamath, A. V.; Chong, S.; Chang, M.; Marathe, P. H.
Cancer Chemother. Pharmacol. 2005, 55, 110–116.
10. Pacek, M.; Prokhorova, T. A.; Walter, J. C. Cell Cycle
2004, 3, 401–403.
19. Rosenblatt, J.; De Bondt, H.; Jancarik, J.; Morgan, D. O.;
Kim, S.-H. J. Mol. Biol. 1993, 230, 1317–1319.
20. Pierce, A. C.; Sandretto, K. L.; Bemis, G. W. Proteins
2002, 49, 567–576.
21. Trova, M. P. PCT Int. Appl. WO 03/022805, 2003; Chem.
Abstr. 2003, 221651.
11. All kinases from Upstate Ltd.; http://www.upstate.com.
12. Kinase assays were adapted from the following generic
protocol, with ATP concentrations stated in the footnotes
to Table 1 Lane, M. E.; Yu, B.; Rice, A.; Lipson, K. E.;
Liang, C.; Sun, L.; Tang, C.; McMahon, G.; Pestell, R.
G.; Wadler, S. Cancer Res. 2001, 61, 6170–6177.