4-Arylazo-3,5-diamino-1H-pyrazole CDK inhibitors: SAR study, crystal structure in complex with CDK2, selectivity, and cellular effects

Article abstract of DOI:10.1021/jm0605740

In a routine screening of our small-molecule compound collection we recently identified 4-arylazo-3,5-diamino-1H-pyrazoles as a novel group of ATP antagonists with moderate potency against CDK2-cyclin E. A preliminary SAR study based on 35 analogues suggests ways in which the pharmacophore could be further optimized, for example, via substitutions in the 4-aryl ring. Enzyme kinetics studies with the lead compound and X-ray crystallography of an inhibitor-CDK2 complex demonstrated that its mode of inhibition is competitive. Functional kinase assays confirmed the selectivity toward CDKs, with a preference for CDK9-cyclin T1. The most potent inhibitor, 4-[(3,5-diamino-1H-pyrazol-4-yl) diazenyl]phenol 31b (CAN508), reduced the frequency of S-phase cells of the cancer cell line HT-29 in antiproliferation assays. Further observed cellular effects included decreased phosphorylation of the retinoblastoma protein and the C-terminal domain of RNA polymerase II, inhibition of mRNA synthesis, and induction of the tumor suppressor protein p53, all of which are consistent with inhibition of CDK9.

Full text of DOI:10.1021/jm0605740

6500
J. Med. Chem. 2006, 49, 6500-6509
4-Arylazo-3,5-diamino-1H-pyrazole CDK Inhibitors: SAR Study, Crystal Structure in Complex
with CDK2, Selectivity, and Cellular Effects†
Vladim´ır Krysˇtof,‡ Petr Cankarˇ,§ Iveta Frysˇova´,§ Jan Slouka,§ George Kontopidis,|, Petr Dzˇuba´k,# Maria´n Hajdu´ch,#
Josef Srovnal,# Walter F. de Azevedo Jr.,X Martin Orsa´g,‡ Martina Paprska´ˇrova´,‡ Jakub Rolcˇ´ık,‡ Alesˇ La´tr,‡
Peter M. Fischer,|,O and Miroslav Strnad*,‡
Laboratory of Growth Regulators, Faculty of Science, Palacky´ UniVersity and Institute of Experimental Botany, Sˇlechtitelu˚ 11, 783 71
Olomouc, Czech Republic, Department of Organic Chemistry, Faculty of Science, Palacky´ University, trˇ. Svobody 8,
772 00 Olomouc, Czech Republic, Structure Based Design Group, Cyclacel Pharmaceuticals, Incorporated, James Lindsay
Place, Dundee, DD1 5JJ, Scotland, United Kingdom, Laboratory of Experimental Medicine, Department of Pediatrics,
Faculty of Medicine, Palacky´ University and Faculty Hospital, Pusˇkinova 6, 775 20 Olomouc, Czech Republic, and Faculdade
de Biocieˆncias, Pontifı´cia Universidade Cato´lica do Rio Grande do Sul, Av. Ipiranga 6681, Porto Alegre, RS 90619-900, Brazil
ReceiVed May 15, 2006
In a routine screening of our small-molecule compound collection we recently identified 4-arylazo-3,5-
diamino-1H-pyrazoles as a novel group of ATP antagonists with moderate potency against CDK2-cyclin E.
A preliminary SAR study based on 35 analogues suggests ways in which the pharmacophore could be
further optimized, for example, via substitutions in the 4-aryl ring. Enzyme kinetics studies with the lead
compound and X-ray crystallography of an inhibitor-CDK2 complex demonstrated that its mode of inhibition
is competitive. Functional kinase assays confirmed the selectivity toward CDKs, with a preference for CDK9-
cyclin T1. The most potent inhibitor, 4-[(3,5-diamino-1H-pyrazol-4-yl)diazenyl]phenol 31b (CAN508),
reduced the frequency of S-phase cells of the cancer cell line HT-29 in antiproliferation assays. Further
observed cellular effects included decreased phosphorylation of the retinoblastoma protein and the C-terminal
domain of RNA polymerase II, inhibition of mRNA synthesis, and induction of the tumor suppressor protein
p53, all of which are consistent with inhibition of CDK9.
Introduction
During the past decade various drug discovery programs have
led to the identification of many potent and selective ATP-
antagonist inhibitors of CDKs, despite the fact that the active
site in protein kinases is highly conserved (reviewed recently
in refs 7-9). Generally, such inhibitors consist of structurally
distinct, flat heterocyclic molecules, which enter the active site
and compete with ATP, a feature usually demonstrated by
enzyme kinetics and inhibitor-CDK2 cocrystal analyses. Anti-
CDK drugs possess prominent inhibitory properties against
cancer cells both in vitro and in vivo, and several are currently
being evaluated in clinical trials as a new generation of
anticancer chemotherapeutics.10,11
The cyclin-dependent kinases (CDKsa) are a family of serine/
threonine protein kinases that play essential roles in the
regulation of cell division. Individual CDKs phosphorylate
distinct substrates in different phases of the cell cycle and are,
therefore, classified as G1 (CDK4 and CDK6-D cyclins, CDK2-
cyclin E), S (CDK2-cyclin A, CDK1-cyclin A), and G2/M
(CDK1-cyclin B) phase-specific CDKs.1,2 The activity of these
kinases is tightly regulated at several levels: through interactions
with negative and positive regulatory partners, activation by
phosphorylation or dephosphorylation, and shifts in their sub-
cellular localization. Cell cycle deregulation is frequently
accompanied by altered CDK activity in many cancers, caused
either by changed expression or activation of CDKs and/or their
interacting proteins. For these reasons, cell cycle CDKs,
especially CDK2 and CDK4, have been actively pursued as
pharmacological targets for novel anticancer agents.3-6
Apart from roles in cell cycle regulation, several CDKs also
participate in other physiological processes such as neuronal
functions (CDK5, CDK11), apoptosis (CDK1, CDK5), and
transcription (CDK2, CDK7, CDK8, CDK9, CDK11); these
processes have also been shown to be affected by pharmaco-
logical inhibitors of CDKs.6,12 In particular, CDKs implicated
in the regulation of transcription at the level of RNA polymerase
II (RNAP-II) have recently attracted interest among medicinal
chemists because of the roles of these CDKs during viral
infection.13,14 Small DNA viruses rely on host cell CDKs for
their replication; some require the host expression apparatus,
while others encode their own cyclins, which activate cellular
CDKs or contain proteins that directly recruit CDKs to the
nascent viral transcripts.13 Some CDK inhibitors, such as
flavopiridol and roscovitine, have already been shown to
interfere with the replication of several viruses, and this activity
is attributed to inhibition of transcription.15-18
* Corresponding author. Tel.: +420 585 634 850. Fax: +420 585 634
870. E-mail: strnad@aix.upol.cz.
† The structure has been deposited with the Protein Data Bank (PDB)
under the accession code 2CLX.
‡ Palacky´ University and Institute of Experimental Botany.
§ Palacky´ University.
| Cyclacel Pharmaceuticals, Inc.
Current address: Veterinary Faculty, University of Thessaly, P. O. Box
199, 43100 Karditsa, Greece.
# Palacky´ University and Faculty Hospital.
X Pontif´ıcia Universidade Cato´lica do Rio Grande do Sul.
O Current address: Centre for Biomolecular Sciences, School of Phar-
macy, University of Nottingham, University Park, Nottingham NG7 2RD,
U.K.
a Abbreviations: CDK, cyclin-dependent kinase; RNAP-II, RNA poly-
merase II; CTD, C-terminal domain; DRB, 5,6-dichloro-1-(â)-D-ribofura-
nosylbenzimidazole; BrdU, 5-bromo-2′-deoxyuridine; CEA, carcinoembry-
onic antigen; EpCAM, epithelial cell adhesion molecule; GAPDH,
glyceraldehyde-3-phosphate dehydrogenase.
These transcriptional CDK inhibitors may also have onco-
logical applications. Until recently, the antiproliferative effects
of CDK inhibitors (which are mainly oligospecific) against
tumor cells were thought to be due to cellular attenuation of
10.1021/jm0605740 CCC: $33.50 © 2006 American Chemical Society
Published on Web 09/29/2006

Arylazopyrazole CDK9 Inhibitors
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 22 6501
Scheme 1. Synthesis of 4-Arylazo-3,5-diamino-1H-pyrazoles
CDK2 activity.19 However, recent target validation studies have
shown that CDK2 knockout mice are fully viable and, even,
that cancer cells can proliferate in the absence of CDK2.20,21
The exact reasons for the undoubted antitumor activity, both in
vitro and in vivo, of many CDK inhibitors remain unclear, but
probably more than one CDK must be targeted at the same time
to achieve an antiproliferative effect. In particular, inhibition
of the CDKs that mediate phosphorylation of the C-terminal
domain (CTD) of RNAP-II is thought to contribute markedly
to the cytotoxic potency of such compounds, as shown for
instance with flavopiridol and roscovitine.22-25 A rapid decline
in levels of key proteins, due to inhibition of transcription, may
be the ultimate cause of the cytoxicity of pharmacological
inhibitors of CDKs, especially when both the affected protein
and the corresponding mRNA have short half-lives. One such
gene product is the antiapoptotic factor Mcl-1, which is crucial
for the survival of a range of cell types. Its down-regulation
either by roscovitine or siRNA is sufficient to induce apopto-
sis.26,27 More particularly, selective CDK9 inhibition could serve
as a potential therapeutic strategy against tumor invasion and
metastasis, as recently demonstrated by the finding that inflam-
matory cytokine tumor necrosis factor-R promotes tumor
progression through activation of matrix metalloproteinases.28
During the course of our ongoing CDK inhibitor discovery
and development program, we have already identified several
series of structurally diverse compounds that possess the ability
to block CDK phosphorylating activities.29-34 More recently,
we discovered that the screening hit 4-phenylazo-3,5-diamino-
1H-pyrazole compound 1b inhibits CDK2-cyclin E. Although
this inhibition is comparatively weak, the simple two-step
synthetic assembly of the 4-phenylazo-3,5-diamino-1H-pyrazole
system prompted us to develop a novel group of protein kinase
inhibitors. We thus went on to synthesize a group of 35 3,5-
diamino-1H-pyrazole derivatives substituted with different aryl-
azo groups at position 4, to analyze their structure-activity
relationships and to determine the in vitro antiproliferative
properties of these compounds against cancer cell lines.
Chemistry
The 4-arylazo-3,5-diamino-1H-pyrazole CDK inhibitory com-
pounds were synthesized as outlined in Scheme 1. The precursor
hydrazones a were obtained by diazotization of the appropriate
arylamines, followed by condensation with malonodinitrile.
Finally, cyclocondensation of these dicyanohydrazones with
hydrazine afforded the 4-arylazo-3,5-diamino-1H-pyrazoles b.
In total, 12 hydrazones and 18 4-arylazo-3,5-diamino-1H-
pyrazoles were synthesized for the first time. Reaction yields
and compound characterization data are included in the Sup-
porting Information, as well as relevant literature references for
compounds already described (Supplementary Tables 1-3,
Supporting Information).
nonpolar substituents also retained the inhibitory activity at the
level of 1b, for example, the comparatively bulky naphthalen-
1-yl derivative 18b (but not the isomeric 19b) or the partially
saturated congener 20b. Although these data suggest that large
aromatic substituents may be compatible with CDK2 inhibitory
activity, when other bulky arylamines were used as starting
materials, the resulting diaminopyrazoles 8b-17b did not
display significant CDK2 inhibitory activity. The explanation
for these findings could be related to the nature of the para
substituent in the arylazo group: a large substituent causing a
detrimental effect, probably through steric hindrance. Such an
effect is also indicated by the reduced inhibitory activity of the
p-aminomethyl analogue 7b. The next subgroup of compounds,
21b-24b obtained from aryldiamines, contained dimeric 3,5-
diamino-1H-pyrazole systems, so the resulting molecules were
much bulkier, and as expected, these derivatives exhibited
reduced and insignificant CDK inhibitory activity.
Results and Discussion
CDK2 Kinase Inhibitory Activity. Routine screening of our
compound libraries for protein kinase inhibitors led to the
identification of 1b as the first representative of a novel group
of compounds that can diminish the catalytic activity of CDK2-
cyclin E. In an effort to identify basic relationships between
their structure and activity, a number of derivatives differing
in the 4-arylazo function were synthesized and evaluated. As
shown in Figure 1, the 4-halogenophenyl derivatives 3b-6b
have similar inhibitory activities to 1b (i.e., the analogue with
an unsubstituted phenyl ring), while the nonpolar 4-methyl-
phenyl derivative 2b shows a small decrease in activity. Other
Following the finding that the majority of compounds with
at least some activity against CDK2 possessed a small 4-arylazo
group, we turned our attention to the effects of various small
substituents on the phenyl ring. Specifically, we prepared and
tested a series of 4-phenylazo derivatives with an additional

6502 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 22
Krysˇtof et al.
Figure 1. Inhibition of CDK2-cyclin E by 4-arylazo-3,5-diamino-1H-pyrazoles at 10 µM concentration; olomoucine (OC) was included as a
control. Average values from three determinations (SD.
Table 1. Inhibition of CDK2 and in Vitro Antiproliferative Activity of Selected Derivatives
IC50a (µM; or % of viable cells)
No
R
CDK2
MCF7
HOS
G361
K562
1b
H
22 ( 5
>100 (74 ( 5%)
>100 (56 ( 1%)
81 ( 9 (39 ( 10%)
>100 (62 ( 9%)
65 ( 7 (44 ( 6%)
>100 (60 ( 4%)
>100 (72 ( 6%)
>100 (71 ( 9%)
33 ( 5 (20 ( 6%)
>100 (97 ( 7%)
>100 (102 ( 8%)
>100 (97 ( 9%)
>100 (51 ( 10%)
134 ( 5
>100 (83 ( 8%)
>100 (79 ( 8%)
92 ( 5 (32 ( 6%)
>100 (57 ( 6%)
>100 (95 ( 5%)
>100 (80 ( 8%)
>100 (84 ( 7%)
>100 (83 ( 8%)
49 ( 9 (22 ( 8%)
>100 (91 ( 10%)
>100 (94 ( 11%)
>100 (92 ( 3%)
>100 (87 ( 8%)
144 ( 15
>100 (75 ( 2%)
>100 (55 ( 14%)
77 ( 4 (27 ( 5%)
>100 (58 ( 2%)
>100 (69 ( 8%)
>100 (61 ( 10%)
>100 (74 ( 12%)
>100 (70 ( 10%)
64 ( 5 (8.3 ( 4%)
>100 (96 ( 12%)
>100 (93 ( 7%)
>100 (86 ( 14%)
>100 (84 ( 11%)
147 ( 14
>100 (90 ( 14%)
50 ( 3 (7.7 ( 6%)
55 ( 1 (26 ( 7%)
100 ( 10 (48 ( 4%)
79 ( 6 (28 ( 5%)
71 ( 6 (25 ( 3%)
>100 (72 ( 8%)
>100 (53 ( 6%)
62 ( 8 (23 ( 7%)
>100 (89 ( 6%)
>100 (84 ( 8%)
>100 (96 ( 9%)
>100 (86 ( 2%)
145 ( 22
18b
20b
26b
27b
28b
29b
30b
31b
32b
33b
34b
35b
2-CHdCH-CHdCH-3
2-CH2CH2CH2CH2-3
2-NO2
3-NO2
4-NO2
2-OH
3-OH
4-OH
2-COOH
3-COOH
4-COOH
2-CH2OH
10 ( 3
15 ( 6
6.1 ( 1.7
23 ( 2
40 ( 9
2.5 ( 0.1
6.0 ( 0.2
3.5 ( 0.7
92 ( 12
28 ( 9
>100
17 ( 5
olomoucine
5.0 ( 1.0
a IC50 measured in the presence of 15 µM ATP, % of viable cells in the presence of 100 µM of the test compound. Average values from three determinations
(SD.
small polar phenyl substituent (derivatives 26b-35b in Table
1). The most effective members of the series were hydroxy and
nitro compounds. In the case of the phenol analogues, the ortho
(29b) and para (31b) isomers showed the lowest CDK2 IC50
values, whereas the meta-hydroxy derivative 30b was about half
as potent. Further modification, by insertion of a methylene
group between the 2-OH and the phenyl groups (35b), led to a
decline in activity compared with the homologue 29b. Interest-
ingly, nitrated compounds 26b-28b displayed a different
positional effect. Here, the position of the nitro group affected
the CDK2 inhibitory activity in the order ortho > meta > para,
with a 2- to 4-fold drop of potency between the different
isomers. All carboxylic acids (32b-34b) drastically lost inhibi-
tory activity.
Mechanism of CDK2 Inhibition. To further investigate the
mechanism of CDK2-cyclin E inhibition by the 4-phenylazo-
3,5-diamino-1H-pyrazole compounds, kinase assays were per-
formed with one of the more potent analogues, 31b, in which
the concentrations of both ATP and the inhibitor were varied.
Double reciprocal plots of the data (Figure 2) show that
compound 31b exhibited pure competitive inhibition with
respect to ATP, with an apparent Ki value of 13.3 µM, similar
to that found with most other known pharmacological CDK
inhibitors.
Figure 2. Double reciprocal plot of CDK2-cyclin E activity in the
presence of 31b as a function of ATP concentration. The inhibitor
concentrations were 1 µM (filled triangles), 5 µM (empty circles), 10
µM (filled circles), 15 µM (empty squares), and 20 µM (filled squares).
The inset shows the secondary replot of slopes vs concentrations of
31b (Ki ) 13.3 µM).
Kinase Selectivity. The selectivity of the most active
compounds against CDK2 (i.e., 1b, 26b, 27b, 30b, 31b, and
35b) was then tested against six purified CDKs, including
CDK1, CDK4, CDK7, and CDK9 (Table 2). These assays
confirmed that phenylazo-3,5-diamino-1H-pyrazoles are efficient
inhibitors of members of the CDK family other than CDK2. In
addition to CDK2-cyclin E, another G1-phase specific kinase,

Arylazopyrazole CDK9 Inhibitors
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 22 6503
Table 2. Selectivity of 4-Arylazo-3,5-diaminopyrazoles against Purified
Cyclin-Dependent Kinases
Crystal Structure of 31b with CDK2. The crystal structure
of the monomeric CDK2-31b complex was determined at 1.80
Å resolution (Supplementary Table 4, Supporting Information)
to determine the inhibitor’s binding mode within the active site
of the kinase and to verify the competitive character of the
inhibition. As expected, 31b binds to the ATP site located in
the cleft between the N- and C-terminal kinase domains (Figure
3a). The inhibitor is oriented with the 3,5-diaminopyrazole
system facing the CDK2 hinge region polypeptide backbone,
where it interacts with Glu81 and Leu83 through three H-bonds
(Figure 3b), in a manner similar to that of other heterocyclic
CDK inhibitors.43 The phenylazo moiety faces the Phe80 phenyl
ring, forming a π interaction and the phenol function H-bonding
to the carboxyl of Asp145 (Figure 3b). The orientation and the
H-bond pattern of 31b in the active site corresponds to those
of PNU-292137, recently described as a selective CDK2
inhibitor.35,36 Interestingly, 31b is observed to adopt a confor-
mation close to a coplanar cis-diaryldiazine (Figure 3c). The
energetically less-favorable cis ligand conformation, compared
to the trans arrangement, appears to be preferred because it
enables simultaneous interaction with the diaminopyrazole and
phenol moieties described above. Diaryldiazines can undergo
facile cis-trans isomerization under a variety of conditions and
by several mechanisms; whereas trans-diaryldiazines exist
predominantly in planar conformations, this is not possible in
the cis geometry, where the two aryl systems are not coplanar.44
The electron density for 31b in the CDK2 complex is well-
defined apart from the phenol substituent, where weak density
is observed for the benzene ring and density for the OH group
is absent. We suggest that this situation is due to the fact that
the electron density represents an average picture of two binding
modes, in which the ligand adopts cis conformations with the
phenol group either above or below the plane of the remaining
aryldiazine system.
a
IC50
compd CDK1/B CDK2/E CDK2/A CDK4/D1 CDK7/H CDK9/T1
1b
>100
42 ( 3
>100
>100
44.0 ( 7 20 ( 6
>100 77 ( 5
100
22 ( 3
68 ( 6
>100
88 ( 1
>100
>100
>100
77 ( 5
>100
15 ( 3
26b
27b
30b
31b
35b
16 ( 1
70 ( 12
56 (13
5.6 ( 5
12.4 ( 3.2
82 ( 15 1.5 ( 0.5
51 ( 10 >100
69 ( 1
>100
13.5 ( 3.1 26 ( 13 0.35 ( 0.04
>100 >100 26 ( 1
a Kinase activity assayed in the presence of 100 µM ATP. Average values
from three determinations (SD.
Table 3. Selectivity of 31b against Various Protein Kinases
kinase activitya
protein kinase
(%)
CDK9/cyclin T1
CDK2/cyclin E
c-Abl
CHK1
CK2
6.9
31
84
89
65
81
99
51
67
100
82
GSK3â
MAPK1
p70S6K
PKA
PKBR
SAPK2R
a In the presence of 10 µM test compound.
CDK4-cyclin D1, was inhibited approximately to the same
degree by 26b, 27b, 30b, and 31b. Structurally related 3-ami-
nopyrazole CDK inhibitors (e.g., PNU-292137), which were
identified during the course of our investigation, appear to
possess different selectivity profiles and do not inhibit CDK4.35,36
Interestingly, CDK2 in complex with cyclin A proved to be
about 4-fold less-sensitive to the pyrazole derivatives 26b and
31b than CDK2-cyclin E. Surprisingly, all of these derivatives
were most potent against CDK9-cyclin T1, unlike any other
known classes of CDK inhibitors, with 31b being the most
effective inhibitor, with submicromolar potency (IC50 ) 0.35
µM). The selectivity ratios with respect to CDK9 of the
derivative 31b ranged from 40 to 200 for various other CDKs,
suggesting that significant selectivity, even within the CDK
family, can be attained with the diaminopyrazole pharmaco-
phore.
To further characterize the pyrazole derivatives, CDK inhibi-
tor 31b was assayed (at a fixed concentration of 10 µM) against
nine other protein kinases, including CDK9 and CDK2 as
reference controls (Table 3). Most of the kinases tested were
inhibited poorly or not at all, corroborating the selectivity of
pyrazole derivatives toward CDKs. Surprisingly, 31b showed
a distinct selectivity profile when compared with other CDK
inhibitors. For example, purines also inhibit ERK2 and
CDK7,37-39 whereas paullones and aloisines also inhibit
GSK3â.39,40 Another compound that also reduces CTD phos-
phorylation of RNAP-II via CDK9 inhibition, 5,6-dichloro-1-
â-D-ribofuranosylbenzimidazole (DRB), inhibits casein kinases
too,41,42 an effect not observed with 31b.
Model of 31b Complexed to CDK9. The selectivity of 31b
toward CDK9-cyclin T1 found during kinase assays prompted
an investigation into the structural basis for this selectivity.
Analysis of the modeled interactions between 31b and CDK9
(Figure 3b) indicates the participation of Asp167 (which
corresponds to Asp145 in CDK2) in an H-bond with the phenol-
OH of 31b. H-bond interactions corresponding to those with
Glu81 and Leu83 observed in the CDK2-31b complex appear
to be absent in the CDK9-31b model. This is the first CDK
inhibitor to show no participation of the molecular fork of CDK
in intermolecular hydrogen bonds, previously described to be
present in intermolecular hydrogen bonds in the structures of
all CDK-inhibitor complexes studied so far.45 In the CDK9
complex structure, CDK9 interactions with 31b are characterized
by predominantly hydrophobic and van der Waals interactions
between the protein and the 4-phenylazo moiety. Most of the
intermolecular contacts between CDK9 and 31b involve a
hydrophobic pocket formed by the residues Ile25, Val33,
Phe103, Cys116, Leu156, Ala166, and Asp167. A direct
structural comparison of both binary complexes indicates that
the high specificity of 31b for CDK9 does not stem from the
establishment of H-bonds, because the CDK9-31b complex
presents a smaller number of H-bonds than the CDK2-31b
complex. The most striking difference between the two com-
plexes concerns the intermolecular contact area, which is
significantly higher in the CDK9 complex (190 Å2) than in the
CDK2 complex (174 Å2; Figure 3d,e). Structural inspection of
the interaction between 31b and CDK9 suggests that the
inhibitor’s specificity against CDK9 is probably due to the
4-phenylazo moiety, which affords a higher intermolecular
Of the kinases tested, only the structurally distinct ribosomal
p70S6 kinase was found to be slightly sensitive to 31b, with
an IC50 value of about 10 µM. Although this kinase is one of
the potential intracellular targets of the CDK inhibitor purvalanol
B, as shown by affinity chromatography,37 the selectivity profile
of 31b is not reminiscent of any other pharmacological CDK
inhibitor, although its cellular kinase specificity profile remains
to be determined.

6504 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 22
Krysˇtof et al.
Figure 3. Crystal structure of CDK2 complex with 31b. (a) The experimentally determined 3D structure of CDK2 (cyan) aligned with the modeled
structure of CDK9 (green); the ligand is shown as a space-filling CPK representation. (b) Details of the interactions between ligand and residues
of the ATP-binding pocket, indicated with broken lines. (c) Observed electron density around the bound inhibitor 31b. The protein surfaces within
4 Å of the ligand are shown for CDK2 (d) and CDK9 (e), respectively.
contact area with CDK9 than that observed for the CDK2-
31b complex. Especially interesting is the presence of two salt
bridges (Glu107-Lys35 and Glu107-Lys44) in the CDK9-
31b structure; these were also observed in the CDK9-
flavopiridol complex and involve residues from two different
lobes of the CDK9 structure.45 This strong electrostatic interac-
tion brings the two lobes of CDK9 closer together, which
increases the contact area between the enzyme and the inhibitor.
These salt bridges are not observed in the CDK2-31b complex.
A tighter ATP-binding pocket observed in the CDK9-31b
complex makes possible a higher number of intermolecular van
der Waals contacts between CDK9 and 31b.
Cellular Activity. The 4-phenylazo-3,5-diamino-1H-pyra-
zoles with the most potent CDK inhibitory activity were tested
for antiproliferative activity against four human tumor cell lines
of different histological origins. The results are summarized in
Table 1. All compounds except derivatives 27b, 28b, 31b, and
20b showed only marginal effects, with IC50 values exceeding
100 µM. This weak activity is comparable with that of
olomoucine, a well-known but weak inhibitor of CDK1 and
CDK2.46 In general, derivatives substituted with bulky aryl
moieties (e.g., sulfonamides 8b-13b or bis-azo derivatives
22b-25b, data not shown) showed either very low or no
antiproliferative activity, while some pyrazoles with smaller
4-aryl side chains displayed measurable IC50 values.
Surprisingly, the observed effects of the position of the
hydroxyl- and nitro-phenyl substituents on the activity did not
simply reflect the CDK inhibition pattern. Hydroxy compounds
29b and 31b, among the most potent CDK2 inhibitors, showed
significant differences in cell proliferation assays. Compound
31b caused significant reductions in cell number in all four cell
lines used. In contrast, 29b and 30b, despite their potent CDK
inhibition, caused much weaker responses. Conversely, the nitro
derivatives 26b and 28b, which differed approximately 7-fold
in CDK inhibition potency, induced cell death at comparable
concentrations.
membrane penetration and alter its enzyme binding parameters.
In summary, the observed discrepancies between the results of
the kinase and cellular assays suggest that the pyrazole
compounds’ effects on targets other than CDKs may contribute
to their antiproliferative activity and/or that their physicochem-
ical properties affect their cellular activities.
Cell Cycle Effects. It is well known that CDK2 and CDK4
play critical roles in the G1-S transition of the cell cycle by
phosphorylating the retinoblastoma protein (pRb), which in turn
activates E2F-mediated transcription of S-phase specific genes.47
Inhibition of cellular CDK activity is, therefore, expected to
result in inhibition of pRb phosphorylation and cell cycle arrest
in the G1 phase. To investigate further the cellular effects of
the 4-arylazo-3,5-diamino-1H-pyrazoles, we therefore evaluated
compound 31b (an analogue with strong anti-CDK and growth
inhibition activities and high selectivity) using cell cycle analysis
and expression-phosphorylation assays in cancer cells. The
antiproliferative activity of 31b was verified by flow cytometric
analysis of subconfluent and asynchronously growing HT-29
cells, which were doubly stained with propidium iodide and
5-bromo-2′-deoxyuridine (BrdU). In accordance with the se-
lectivity of 31b toward CDK9-cyclin T1 when measured in vitro,
this compound did not influence the cell cycle, unlike the less-
selective CDK inhibitor roscovitine, which blocked both G1-S
and G2-M transitions (Figure 4). With 31b, we only observed
a difference in the S-phase cell population, which showed
substantially reduced intensity of the BrdU signal compared with
control cells. Subsequently, increases in the sub-G1 cell
population, usually considered apoptotic, were observed in cells
treated with either 31b or roscovitine for periods longer than 3
h.
When MCF7 cells were treated with varying concentrations
of 31b for 24 h, dose-dependent inhibition of pRb phosphor-
ylation at Ser807 and Ser780 (both of which are specifically
phosphorylated during the G1 phase of the cell cycle by CDK4-
cyclin D) and Thr821 (specifically phosphorylated by CDK2-
cyclin E) was observed in immunoblots of total proteins probed
with phosphospecific antibodies (Figure 5). The results dem-
onstrate the ability of 31b to affect the activities of CDK4 and
CDK2 in cells. In addition, a drop in pRb phosphorylation
Notably, compound 20b, which has a weak impact on CDK
activity, also showed antiproliferative activity, exceeding that
of the olomoucine control on all tested cell lines, probably due
to its relatively low polar surface area, which may facilitate

Arylazopyrazole CDK9 Inhibitors
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 22 6505
Figure 4. Cell cycle analysis and BrdU incorporation in HT-29 cells treated with 31b (50 µM) or roscovitine (70 µM). Drug concentrations were
equivalent to IC50 values of the compounds against HT-29 cells.
of CDK4 and CDK2 (blocked activation by CDK7, absence of
cyclins) rather than by direct inhibition.
Inhibition of Transcription. Pharmacological CDK inhibi-
tors usually increase the cellular level of the tumor suppressor
protein p53 and, consequently, transcription of p53-activated
genes.24,48 This effect probably results from direct inhibition of
CDKs involved in the regulation of transcription, that is, CDK7,
CDK8, and CDK9, which leads to decreased levels of short-
lived mRNA and protein gene product. One such mRNA
encodes the negative regulator p53-specific Hdm2 ubiquitin E3
ligase. Insufficient elongation of the ligase triggers stabilization
of the p53 protein, as demonstrated by cell treatment with certain
CDK inhibitors.24,32,48 Paradoxically, a block in transcription
can thus be accompanied by increased levels of certain proteins,
like p53 and its downstream regulated tumor suppressor
p21WAF1. We, therefore, evaluated the effect of the derivative
31b at various concentrations on the levels of p53 and p21WAF1
proteins in MCF7 cells and found that 24 h incubation with
31b at concentrations of 50 and 100 µM resulted in strongly
increased levels of p53 (Figure 5). Accumulated p53 then
transactivated the cell cycle inhibitory protein p21WAF1, levels
of which were maximal at 50 µM 31b. In addition, the activity
of 31b was significant at concentrations close to the IC50 value
for MCF7 growth inhibition. Thus, increased levels of p21WAF1
may also contribute to inhibition of CDKs in cells. Olomoucine,
tested in parallel as a control, was effective at doses of about
100 µM (data not shown).
Figure 5. Immunoblot analysis of MCF7 cells treated for 24 h with
the indicated concentrations of 31b. Phosphorylation of pRb appears
to be reduced at Ser780 and Ser807 (indicative of decreased activity
of CDK4) and Thr821 (indicative of decreased activity of CDK2) by
31b. In addition, the compound appears to induce increases in the levels
of the tumor suppressor protein p53 and p21WAF1 and to reduce the
phosphorylation of CTD of RNA polymerase II, indicating that it
induces blocks in transcription.
occurred at inhibitor concentrations between 25 and 50 µM, in
accordance with the observed IC50 value for MCF7 growth
inhibition (Table 1). However, due to the marked selectivity of
31b toward CDK7 and CDK9, we surmise that decreased
phosphorylation of pRb may be caused by insufficient activity

6506 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 22
Krysˇtof et al.
Table 4. Inhibition of mRNA Synthesis by Compound 31b (50 µM) or
Roscovitine (70 µM) in HT-29 Cellsa
(probably via combined effects on CDK2, CDK7, and CDK9)
may be responsible for the antiproliferative and pro-apoptotic
potency of CDK inhibitors. However, the cellular activity of
several 4-phenylazo-3,5-diamino-1H-pyrazole derivatives sig-
nificantly exceeded that of olomoucine, thus, they represent good
starting points for further structural optimization of specific
CDK9 inhibitors, with potential pharmacological applications
in oncology and virology.
GAPDH copies/106 cellsb
0 h
3 h
6 h
12 h
control
roscovitine
31b
2 124 368
(399 147
ND
ND
ND
4 096 055
(906 931
3 945 095
(446 126
3 342 533
(334 107
1 395 413
(233 769
3 146 605
(898 073
3 785 908
(479 616
3 030 440
(230 709
3 926 015
(965 476
1 811 960
(581 162
ND
Experimental Section
CEA copies/106 cellsb
General. All compounds prepared were homogeneous, according
to thin-layer chromatographic analysis. Elemental analyses were
performed with an EA elemental analyzer (Fisons Instruments).
Melting points were determined using a Boetius stage apparatus
and are not corrected. All reaction solvents and reagents were
purchased from Sigma-Aldrich, unless specified otherwise. Specific
antibodies were purchased from Cell Signaling (total pRb and pRb
phosphorylated at Ser780 and Ser807), Biosource (pRb phos-
phorylated at Thr821), Abcam (RNA polymerase II phosphorylated
at Ser5), and Becton-Dickinson (anti-BrdU-FITC conjugate) or
were generous gifts from B. Vojteˇsˇek (p53, PCNA, p21WAF1).
General Method for Preparing Hydrazones a. The appropriate
amine (5 mmol) was dissolved in a mixture of water (30 mL) and
hydrochloric acid (4.5 mL of 37% w/v soln). A solution of NaNO2
(0.35 g, 5 mmol) in ice-cold water (5 mL) was added dropwise
with stirring to the cooled amine solution (0-5 °C). The diazonium
salt solution so formed was then added dropwise to a solution of
malononitrile (0.5 g, 7.5 mmol) and NaOAc (12.5 g) in water (50
mL) with continuous stirring and cooling. After adding the
diazonium salt, the reaction mixture was stirred and cooled for an
additional 30 min before being placed in a refrigerator overnight.
The following day, the precipitated hydrazone product was filtered,
washed with water, and dried. Yields were between 90 and 100%
(for details see the Supporting Information). For the preparation of
compounds 32a-35a, twice the above quantities of hydrochloric
acid, NaNO2, malononitrile, and NaOAc were used.
General Method for Preparing Pyrazoles b. A solution of
hydrazine hydrate (0.5 g, 9 mmol) in MeOH (30 mL) was added
to the hydrazone a (3 mmol). The reaction mixture was heated under
reflux for 4 h and was then evaporated to dryness. The solid residue
was recrystallized from appropriate solvents (for details see
Supplementary Table 2, Supporting Information). Compounds 13b
and 14b were precipitated from aqueous solution by acidification
with hydrochloric acid to pH ) 1.
Enzyme Inhibition Assay. CDK2-cyclin E kinase was produced
in Sf9 insect cells co-infected with appropriate baculoviral con-
structs, as previously described.29 The enzyme was purified on a
NiNTA column (Qiagen) and assayed with 1 mg/mL histone H1
in the presence of 15 µM ATP, 0.05 µCi [γ-33P]ATP and of the
test compound in a final volume of 10 µL, all in reaction buffer
(50 mM HEPES, 10 mM MgCl2, 5 mM EGTA, 10 mM 2-glycer-
olphosphate, 1 mM NaF, 1 mM DTT, pH 7.4). After a 10 min
incubation, reactions were stopped by adding 5 µL of 3% aq H3-
PO4. Aliquots were spotted onto P-81 phosphocellulose (Whatman),
which was subsequently washed 3× with 5% aq H3PO4 and finally
air-dried. To quantify kinase inhibition, a BAS-1800 digital image
analyzer (Fujifilm) was employed. Kinase activity was expressed
as a percentage of maximum activity. The concentration of the test
compounds required to decrease the CDK activity by 50% was
determined from dose-response curves and designated IC50. For
the kinetic analyses, the activity was expressed as arbitrary units
(AU) of digital image analyzer signal.
0 h
3 h
6 h
12 h
control
roscovitine
31b
26 030
(3236
ND
34 053
(2409
30 568
(3005
27 900
(3690
14 235
(1568
19 150
(4801
22 620
(816
18 630
(2506
24 350
(8988
12 005
(4123
ND
ND
ND
EpCAM copies/106 cellsb
3 h
0 h
6 h
12 h
control
roscovitine
31b
4 238 025
(162 292
ND
8 732 195
(398 172
8 258 813
(788 357
6 047 138
(485 595
2 609 440
(362 215
6 151 085
(404 186
5 069 600
(471 663
4 888 343
(531 734
6 863 365
(16 636
2 951 620
(121 527
ND
ND
ND
a After the indicated period, total RNA was extracted and absolute
numbers of GAPDH, CEA, and EpCAM gene transcripts were quantified
by real-time PCR. Drug concentrations were equivalent to IC50 values for
the inhibitors against HT-29 cells. b Average values from four determinations
(SD.
The CDK inhibitor roscovitine has also been reported to
inhibit mRNA synthesis, possibly through inhibition of CDK7
and CDK9, which phosphorylate the CTD of RNAP-II and thus
activate transcription.22,25 The possibility that 31b could lead
to a loss of phosphorylation of RNAP-II in cells was assessed
by Western blot analysis with phosphospecific CTD antibodies.
Figure 5 shows there was a concentration-dependent decrease
of Ser5 phosphorylation of RNAP-II in MCF7 cells treated with
31b for 24 h. Moreover, two epithelial genes, encoding
carcinoembryonic antigen (CEA) and human epithelial cell
adhesion molecule (EpCAM), and one housekeeping gene,
encoding glyceraldehyde-3-phosphate dehydrogenase (GAPDH),
were selected for analysis of transcriptional inhibition using
quantitative real-time PCR (RQ RT-PCR) in the HT-29 colon
cancer cell line (Table 4). Copy numbers of GAPDH, CEA,
and EpCAM mRNAs decreased in a time-dependent manner
when incubated with either 31b at 50 µM or roscovitine at 70
µM. However, 31b was much more potent in this respect, which
is consistent with its direct inhibition of CDK9-cyclin T1.
Conclusions
The most potent pyrazole derivative from the series presented
here was 4-[(3,5-diamino-1H-pyrazol-4-yl)diazenyl]phenol 31b
(designated CAN508), which selectively inhibited CDK9 kinase
in vitro. The derivative proved to be a competitive inhibitor of
CDK2-cyclin E with respect to ATP. Cellular activity of 31b
blocks proliferation of several tested cancer cell lines. Despite
31b having high specificity toward CDK9-cyclin T1, its
inhibition does not seem to be the sole reason for the induction
of p53 and the observed antiproliferative potency. The well-
known CDK inhibitor roscovitine, which has similar affinity
for CDK9,11,32,38 causes accumulation of p53 in cells treated at
concentrations lower than 31b.32,48 A comparison of the CDK
selectivity profiles suggests that inhibition of other kinases
Kinase Selectivity. CDK selectivity was evaluated using a panel
of purified kinases (CDK1-cyclin B, CDK2-cyclin A, CDK4-cyclin
D1, CDK7-cyclin H, and CDK9-cyclin T1), as previously de-
scribed.38,49 Other protein kinase activities were measured in
duplicates at a single fixed concentration of 31b (10 µM): c-Abl,
CHK1, GSK3â, p70S6K, PKA and PKBR were assayed in 8 mM
MOPS, pH 7.0, 0.2 mM EDTA, MAPK1 and SAPK2R in 25 mM
Tris, pH 7.5, 0.02 mM EGTA, and CK2 in 20 mM HEPES, pH

Arylazopyrazole CDK9 Inhibitors
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 22 6507
7.6, 0.15 M NaCl, 0.l mM EDTA, 2 mM DTT, 0.1% Triton X-100,
all with 10 mM Mg-acetate, 10 µM ATP, [γ-33P]ATP (500 cpm/
pmol) and appropriate peptide substrates for 40 min. The reactions
were stopped by adding 5 µL of 3% aq H3PO4. Aliquots were then
spotted onto a P30 filtermat and washed 3× with 5% aq H3PO4
and air-dried prior to scintillation counting.
then incubated with specific antibodies overnight. After washing
three times in PBS-T, the membranes were incubated with a 1:2000
dilution of peroxidase-conjugated secondary antibodies. After
another three washes in PBS-T, peroxidase activity was detected
using ECL+ reagents (AP Biotech) according to the manufacturer’s
instructions.
Crystallization and Structure Determination. Human recom-
binant CDK2 was expressed, purified, and crystallized as previously
described.43,50 Data were collected at the Cyclacel home source
using a R-AXIS IV++ image plate (Rigaku). Data processing was
carried out using programs from the d*TREK software suite.51 The
structures were solved by molecular replacement using MOLREP52
and PDB entry 1PW2 as the search models for CDK2. ARP/
wARP53 was used for the addition of water molecules, and
REFMAC52 was used for structural refinement. A number of rounds
of refinement and model building with the program Quanta
(Accelrys) were carried out. Crystallographic data and statistics are
summarized in Supplemetary Table 4 (Supporting Information). The
structure has been deposited with the Protein Data Bank (PDB)
under the accession code 2CLX.
Molecular Model of the CDK9-31b Complex. A previously
published protocol was used to model the binary CDK9-31b
complex.45 Briefly, the atomic coordinates for CDK2-31b were
used as the starting points for modeling the CDK9-31b complex.
Therefore, the final model of the complex includes the atomic
coordinates of the inhbitor 31b. This has been confirmed
by computational docking studies using PRINCIANE (http://
www.biocristalografia.df.ibilce.unesp.br/princiane/). Fifteen residues
from the N-terminus and 45 from the C-terminus were removed
from the CDK9 model because there is no good template for these
fragments. The webserver Parmodel was used to generate all models
for the CDK9-31b complex.54 Several slightly different models
can be calculated by varying the initial structure. We generated
1000 models for the CDK9-31b complex, and the final models
were selected based on stereochemical quality. The optimization
of the complex was carried out by the use of the variable target
function method, followed by 500 steps of conjugate gradient
minimization using Parmodel.54 All optimization processes were
performed on a Beowulf Cluster (BioComp, S. J. do Rio Preto,
Brazil). The overall stereochemical quality of the final model for
the complex of CDK9-31b was assessed by the program Par-
model.54 Analysis of the stereochemical quality of the final model
of the CDK9 indicates that 98.6% nonglycine residues lie in the
most favored and additional allowed regions of the Ramachandran
plot and only 1.4% lie in the generously allowed regions. Root-
mean-square (rms) differences from ideal geometries for bond
lengths and angles were calculated with Parmodel.54 Overall
deviations from ideal geometry are 0.025 Å for bond distances and
2.51° for bond angles.
Cell Maintenance and Growth Inhibition Assay. The ability
of the test substances to inhibit cancer cell growth was determined
in vitro with K562, MCF7, HOS, and G361 cell lines. The cells,
cultured in DMEM (supplemented with 10% fetal calf serum, 4
mM glutamine, 100 IU/mL penicillin, 100 µg/mL streptomycin)
in a humidified CO2 incubator at 37 °C, were redistributed into
96-well microtiter plates at appropriate densities for their respective
cell sizes and growth rates. After a 12 h preincubation, test
compounds in 6-fold dilutions were added. Treatment (in the 1-100
µM range) lasted for 72 h. At the end of this period, the cells were
fed for 1 h with Calcein AM, and the fluorescence of the live cells
was measured at 485 nm/538 nm (ex/em) with a Fluoroskan Ascent
reader (Labsystems). IC50 values, the drug concentrations lethal to
50% of the cancer cells, were determined from the dose-response
curves.
Immunoblotting. For direct immunoblotting, total cellular
protein lysates were prepared by harvesting treated cells in Laemmli
sample buffer. Proteins were separated on 10% SDS-polyacrylamide
gels and electroblotted onto nitrocellulose membrane. The blotted
membranes were stained with Ponceau-S in 1% aq AcOH to verify
equal protein loading, destained, and blocked in PBS and 0.1%
Tween 20 (PBS-T) with 5% low fat milk. The membranes were
BrdU Incorporation and Cell Cycle Analysis. Subconfluent
HT29 cells were treated with 31b or roscovitine at concentrations
corresponding to IC50 values (50/70 µM) for 3, 6, and 12 h. The
cultures were fed a pulse of 10 µM 5-bromo-2′-deoxyuridine (BrdU)
for 30 min at 37 °C before harvesting. The cells were trypsinized,
fixed with ice-cold 70% ethanol, incubated on ice for 30 min,
washed with PBS, and resuspended in 2 M aq HCl for 30 min at
room temperature to denature their DNA. Following neutralization
with 0.1 M Na2B4O7, the cells were harvested by centrifugation
and washed with PBS containing 0.5% Tween-20 and 1% BSA.
They were then stained with anti-BrdU FITC-labeled antibody (1:
50, Becton-Dickinson) for 30 min at room temperature in the dark.
The cells were then washed with PBS, incubated with propidium
iodide (0.1 mg/mL) and RNAse A (0.5 mg/mL) for 1 h at room
temperature in the dark, and finally analyzed by flow cytometry
using a 488 nm single beam laser (FACSCalibur, Becton Dickin-
son).
RNA Isolation and Reverse Transcription. Total RNA was
extracted from treated/control HT-29 cells (5 × 106) cultured in
6-well panels with TRI reagent (Molecular Research Center,
Cincinnati, OH) according to the manufacturer’s instructions. The
concentration and purity of isolated RNA was assessed by UV
spectroscopy. Three microgram portions of total RNA were used
for reverse transcription in a final volume of 30 µL. The RNA
solution was preincubated with 0.3 µg of random primers (Promega)
at 70 °C for 5 min and immediately placed on ice. Then, 6 µL of
5× RevertAid reverse transcriptase buffer (Fermentas), 3 µL of 10
mM deoxyribonucleotide triphosphates (dNTPs), and 0.75 µL of
40 U/µL RNAsin ribonuclease inhibitor (Promega) were added, and
the mixtures were incubated for 5 min at room temperature. In the
final step, 150 U of RevertAid Moloney murine leukaemia virus
reverse transcriptase (Fermentas) was added to each tube, the
samples were incubated at 42 °C for 60 min, and the reverse
transcriptase was then heat-inactivated at 95 °C for 5 min.
Real-Time Polymerase Chain Reaction (RQ RT-PCR). Two
epithelial genes, encoding CEA (CEACAM5) and human EpCAM
(TACSTD 1), and one housekeeping gene, encoding GAPDH, were
selected for analysis of transcriptional inhibition using quantitative
real-time PCR. Primers were selected using PrimerPremier3
software and NCBI reference sequences (accession numbers
NM_004363, NM_002354, and NM_002046, respectively). Specific
primers from two different exons and probes (Generi-Biotech, Czech
Republic) to span introns were designed to reduce amplification
of genomic DNA. The following primers and probes were used:
CEA3 5′-TAA GTG TTG ACC ACA GCG ACC C-3′, CEA4 5′-
GTT CCC ATC AAT CAG CCA AGA A-3′, CEA probe 5′-ATG
TCC TCT ATG GCC CAG ACG ACC C-3′-BHQ1-HEX; Ep-
CAM3 5′-AAA CAC AAA GCA AGA GAA AAA CCT-3′,
EpCAM4 5′-AAT TTT GGA TCC AGT TGA TAA CG-3′,
EpCAM probe 5′-TTG CGG ACT GCA CTT CAG AAG GA-3′-
BHQ1-HEX; GAPDH-F 5′-GAA GAT GGT GAT GGG ATT TC-
3′, GAPDH-R 5′-GAA GGT GAA GGT CGG AGT-3′, GAPDH
probe 5′-CAA GCT TCC CGT TCT CAG CC-3′-BHQ1-FAM. RQ
RT-PCR reactions were performed in 25 µL reaction volumes
consisting of 1 U of HotStart Taq polymerase, 3 mM MgCl2, 10×
PCR buffer (AB Gene), 200 µM dNTPs (Promega), 100 ng DNA,
and variable amounts of primer/probe: 300 nM CEA3, 600 nM
CEA4, and 200 nM CEA probe; 400 nM EpCAM3, 400 nM
EpCAM4, and 200 nM EpCAM probe; 300 nM GAPDH-F, 300
nM GAPDH-R, and 200 nM GAPDH probe. The optimized thermal
profiles for amplification involved 15 min Taq polymerase activa-
tion at 96 °C, followed by 50 cycles at 95 °C for 15 s and 65 °C
for 15 s (CEA), 50 cycles at 95 °C for 15 s and 59 °C for 15 s
(EpCAM), and 50 cycles at 95 °C for 15 s and 60 °C for 30 s
(GAPDH). Gene expression was quantified from calibration curves

6508 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 22
Krysˇtof et al.
(21) Tetsu, O.; McCormick, F. Proliferation of cancer cells despite CDK2
inhibition. Cancer Cell 2003, 3, 233-245.
(22) Ljungman, M.; Paulsen, M. T. The cyclin-dependent kinase inhibitor
roscovitine inhibits RNA synthesis and triggers nuclear accumulation
of p53 that is unmodified at Ser15 and Lys382. Mol. Pharmacol.
2001, 60, 785-789.
using appropriately diluted gene specific amplicons ranging between
30 and 1 000 000 copies per reaction. Each sample was analyzed
in duplicate. Positive and negative controls were also incorporated
into each experiment to evaluate reproducibility and possible sample
contamination during the PCR procedure.
(23) Chao, S.-H.; Price, D. H. Flavopiridol inactivates P-TEFb and blocks
most RNA polymerase II transcription in vivo. J. Biol. Chem. 2001,
276, 31793-31799.
(24) Demidenko, Z. N.; Blagosklonny, M. V. Flavopiridol induces p53
via initial inhibition of Mdm2 and p21 and, independently of p53,
sensitizes apoptosis-reluctant cells to tumor necrosis factor. Cancer
Res. 2004, 64, 3653-3660.
(25) Whittaker, S. R.; Walton, M. I.; Garrett, M. D.; Workman, P. The
cyclin-dependent kinase inhibitor CYC202 (R-roscovitine) inhibits
retinoblastoma protein phosphorylation, causes loss of cyclin D1,
and activates the mitogen-activated protein kinase pathway. Cancer
Res. 2004, 64, 262-272.
(26) MacCallum, D. E.; Melville, J.; Frame, S.; Watt, K.; Anderson, S.;
Gianella-Borradori, A.; Lane, D. P.; Green, S. R. Seliciclib (CYC202,
R-roscovitine) induces cell death in multiple myeloma cells by
inhibition of RNA polymerase II-dependent transcription and down-
regulation of Mcl-1. Cancer Res. 2005, 65, 5399-5407.
(27) Lacrima, K.; Valentini, A.; Lambertini, C.; Taborelli, M.; Rinaldi,
A.; Zucca, E.; Catapano, C.; Cavalli, F.; Gianella-Borradori, A.;
MacCallum, D. E.; Bertoni, F. In vitro activity of cyclin-dependent
kinase inhibitor CYC202 (Seliciclib, R-roscovitine) in mantle cell
lymphomas. Ann. Oncol. 2005, 16, 1169-1176.
(28) Shan, B.; Zhuo, Y.; Chin, D.; Morris, C. A.; Morris, G. F.; Lasky,
J. A. Cyclin-dependent kinase 9 is required for tumor necrosis factor-
R-stimulated matrix metalloproteinase-9 expression in human lung
adenocarcinoma Cells. J. Biol. Chem. 2005, 280, 1103-1111.
(29) Havlicek, L.; Fuksova, K.; Krystof, V.; Orsag, M.; Vojtesek, B.;
Strnad, M. 8-Azapurines as new inhibitors of cyclin-dependent
kinases. Bioorg. Med. Chem. 2005, 13, 5399-5407.
(30) Havlicek, L.; Hanus, J.; Vesely, J.; Leclerc, S.; Meijer, L.; Shaw,
G.; Strnad, M. Cytokinin-derived cyclin-dependent kinase inhibi-
tors: synthesis and cdc2 inhibitory activity of olomoucine and related
compounds. J. Med. Chem. 1997, 40, 408-412.
(31) Krystof, V.; Lenobel, R.; Havlicek, L.; Kuzma, M.; Strnad, M.
Synthesis and biological activity of olomoucine II. Bioorg. Med.
Chem. Lett. 2002, 12, 3283-3286.
(32) Krystof, V.; McNae, I. W.; Walkinshaw, M. D.; Fischer, P. M.;
Muller, P.; Vojtesek, B.; Orsa´g, M.; Havlicek, L.; Strnad, M.
Antiproliferative activity of olomoucine II, a novel 2,6,9-trisubstituted
purine cyclin-dependent kinase inhibitor. Cell. Mol. Life Sci. 2005,
62, 1763-1771.
(33) Moravcova, D.; Krystof, V.; Havlicek, L.; Moravec, J.; Lenobel, R.;
Strnad, M. Pyrazolo[4,3-d]pyrimidines as new generation of cyclin-
dependent kinase inhibitors. Bioorg. Med. Chem. Lett. 2003, 13,
2989-2992.
(34) Moravec, J.; Krystof, V.; Hanus, J.; Havlicek, L.; Moravcova, D.;
Fuksova, K.; Kuzma, M.; Lenobel, R.; Otyepka, M.; Strnad, M.
2,6,8,9-Tetrasubstituted Purines as New CDK1 Inhibitors. Bioorg.
Med. Chem. Lett. 2003, 13, 2993-2996.
(35) Pevarello, P.; Brasca, M. G.; Amici, R.; Orsini, P.; Traquandi, G.;
Corti, L.; Piutti, C.; Sansonna, P.; Villa, M.; Pierce, B. S.; Pulici,
M.; Giordano, P.; Martina, K.; Fritzen, E. L.; Nugent, R. A.; Casale,
E.; Cameron, A.; Ciomei, M.; Roletto, F.; Isacchi, A.; Fogliatto, G.;
Pesenti, E.; Pastori, W.; Marsiglio, A.; Leach, K. L.; Clare, P. M.;
Fiorentini, F.; Varasi, M.; Vulpetti, A.; Warpehoski, M. A. 3-Ami-
nopyrazole inhibitors of CDK2/Cyclin A as antitumor agents. 1. Lead
finding. J. Med. Chem. 2004, 47, 3367-3380.
(36) Pevarello, P.; Brasca, M. G.; Orsini, P.; Traquandi, G.; Longo, A.;
Nesi, M.; Orzi, F.; Piutti, C.; Sansonna, P.; Varasi, M.; Cameron,
A.; Vulpetti, A.; Roletto, F.; Alzani, R.; Ciomei, M.; Albanese, C.;
Pastori, W.; Marsiglio, A.; Pesenti, E.; Fiorentini, F.; Bischoff, J.
R.; Mercurio, C. 3-Aminopyrazole Inhibitors of CDK2/Cyclin A as
antitumor agents. 2. Lead optimization. J. Med. Chem. 2005, 48,
2944-2956.
(37) Knockaert, M.; Gray, N.; Damiens, E.; Chang, Y.-T.; Grellier, P.;
Grant, K.; Fergusson, D.; Mottram, J.; Soete, M.; Dubremetz, J.-F.;
Le Roch, K.; Doerig, C.; Schultz, P. G.; Meijer, L. Intracellular targets
of cyclin-dependent kinase inhibitors: identification by affinity
chromatography using immobilised inhibitors. Chem. Biol. 2000, 7,
411-422.
(38) McClue, S. J.; Blake, D.; Clarke, R.; Cowan, A.; Cummings, L.;
Fischer, P. M.; MacKenzie, M.; Melville, J.; Stewart, K.; Wang, S.;
Zhelev, N.; Zheleva, D.; Lane, D. P. In vitro and in vivo antitumor
properties of the cyclin dependent kinase inhibitor CYC202 (R-
roscovitine). Int. J. Cancer 2002, 102, 463-468.
Acknowledgment. The authors thank K. Faksova´, E. Hirn-
erova´, J. Hudcova´, and the screening group members at Cyclacel
for their contributions. B. Vojteˇsˇek is acknowledged for the gift
of antibodies and Sees-editing for English corrections. The work
was supported by GACR Grant 204/03/D231, MSMT Grant
6198959216, and SMOLBnet FAPESP 01/07532-0. W.F.A. is
a researcher for the Brazilian Council for Scientific and
Technological Development, CNPq.
Supporting Information Available: Data from the elemental
analyses, melting points, recrystallization solvents of newly prepared
compounds, references to previously known compounds, as well
as X-ray data are included. This material is available free of charge
via the Internet at http://pubs.acs.org.
References
(1) Morgan, D. O. Cyclin-dependent kinases: engines, clocks, and
microprocessors. Annu. ReV. Cell DeV. Biol. 1997, 13, 261-291.
(2) Sherr, C. J. Cancer cell cycles. Science 1996, 274, 1672-1677.
(3) Sielecki, T. M.; Boylan, J. F.; Benfield, P. A.; Trainor, G. L. Cyclin-
dependent kinase inhibitors: useful targets in cell cycle regulation.
J. Med. Chem. 2000, 43, 1-18.
(4) Dai, Y.; Grant, S. Cyclin-dependent kinase inhibitors. Curr. Opin.
Pharmacol. 2003, 3, 362-370.
(5) Fischer, P. M.; Lane, D. P. Inhibitors of cyclin-dependent kinases as
anti-cancer therapeutics. Curr. Med. Chem. 2000, 7, 1213-1245.
(6) Knockaert, M.; Greengard, P.; Meijer, L. Pharmacological inhibitors
of cyclin-dependent kinases. Trends Pharmacol. Sci. 2002, 23, 417-
425.
(7) Hirai, H.; Kawanishi, N.; Iwasawa, Y. Recent advances in the
development of selective small molecule inhibitors for cyclin-
dependent kinases. Curr. Top. Med. Chem. 2005, 5, 167-179.
(8) McInnes, C.; Fischer, P. M. Strategies for the design of potent and
selective kinase inhibitors. Curr. Pharm. Des. 2005, 11, 1845-1863.
(9) Fischer, P. M. The design of drug candidate molecules as selective
inhibitors of therapeutically relevant kinases. Curr. Med. Chem. 2004,
11, 1563-1583.
(10) Fischer, P. M.; Gianella-Borradori, A. Recent progress in the
discovery and development of CDK inhibitors. Expert Opin. InVest.
Drugs 2005, 14, 457-477.
(11) Benson, C.; Kaye, S.; Workman, P.; Garrett, M.; Walton, M.; de
Bono, J. Clinical anticancer drug development: targeting the cyclin-
dependent kinases. Br. J. Cancer 2005, 92, 7-12.
(12) Loyer, P.; Trembley, J. H.; Katona, R.; Kidd, V. J.; Lahti, J. M.
Role of CDK/cyclin complexes in transcription and RNA splicing.
Cell. Signalling 2005, 17, 1033-1051.
(13) Schang, L. M. Advances on cyclin-dependent kinases (CDKs) as
novel targets for antiviral drugs. Curr. Drug Targets: Infect. Disord.
2005, 5, 29-37.
(14) Fischer, P. M. Cyclin-dependent kinase inhibitors: discovery,
development and target rationale for different therapeutic applications.
Drugs Future 2005, 30, 911-929.
(15) Chao, S.-H.; Fujinaga, K.; Marion, J. E.; Taube, R.; Sausville, E.
A.; Senderowicz, A. M.; Peterlin, B. M.; Price, D. H. Flavopiridol
inhibits P-TEFb and blocks HIV-1 replication. J. Biol. Chem. 2000,
275, 28345-28348.
(16) Schang, L. M. Effects of pharmacological cyclin-dependent kinase
inhibitors on viral transcription and replication. Biochim. Biophys.
Acta 2004, 1697, 197-209.
(17) Agbottah, E.; de La Fuente, C.; Nekhai, S.; Barnett, A.; Gianella-
Borradori, A.; Pumfery, A.; Kashanchi, F. Antiviral Activity of
CYC202 in HIV-1-infected Cells. J. Biol. Chem. 2005, 280, 3029-
3042.
(18) Heredia, A.; Davis, C.; Bamba, D.; Le, N.; Gwarzo, M. Y.; Sadowska,
M.; Gallo, R. C.; Redfield, R. R. Indirubin-3′-monoxime, a derivative
of a Chinese antileukemia medicine, inhibits P-TEFb function and
HIV-1 replication. AIDS 2005, 19, 2087-2095.
(19) Fischer, P. M. The use of CDK inhibitors in oncology: a pharma-
ceutical perspective. Cell Cycle 2004, 3, 742-746.
(20) Ortega, S.; Prieto, I.; Odajima, J.; Martin, A.; Dubus, P.; Sotillo, R.;
Barbero, J. L.; Malumbres, M.; Barbacid, M. Cyclin-dependent kinase
2 is essential for meiosis but not for mitotic cell division in mice.
Nat. Genet. 2003, 35, 25-31.

Arylazopyrazole CDK9 Inhibitors
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 22 6509
(39) Bain, J.; McLauchlan, H.; Elliott, M.; Cohen, P. The specificities of
protein kinase inhibitors: an update. Biochem. J. 2003, 371, 199-
204.
(40) Mettey, Y.; Gompel, M.; Thomas, V.; Garnier, M.; Leost, M.;
Ceballos-Picot, I.; Noble, M.; Endicott, J.; Vierfond, J.-M.; Meijer,
L. Aloisines, a new family of CDK/GSK-3 inhibitors. SAR study,
crystal structure in complex with CDK2, enzyme selectivity, and
cellular effects. J. Med. Chem. 2003, 46, 222-236.
(41) Zandomeni, R.; Zandomeni, M. C.; Shugar, D.; Weinmann, R. Casein
kinase type II is involved in the inhibition by 5,6-dichloro-1-beta-
D-ribofuranosylbenzimidazole of specific RNA polymerase II tran-
scription. J. Biol. Chem. 1986, 261, 3414-3419.
(42) Meggio, F.; Shugar, D.; Pinna, L. A. Ribofuranosyl-benzimidazole
derivatives as inhibitors of casein kinase-2 and casein kinase-1. Eur.
J. Biochem. 1990, 187, 89-94.
(43) Wu, S. Y.; McNae, I.; Kontopidis, G.; McClue, S. J.; McInnes, C.;
Stewart, K. J.; Wang, S.; Zheleva, D. I.; Marriage, H.; Lane, D. P.;
Taylor, P.; Fischer, P. M.; Walkinshaw, M. D. Discovery of a novel
family of CDK inhibitors with the program LIDAEUS: Structural
basis for ligand-induced disordering of the activation loop. Structure
2003, 11, 399-410.
(44) Stepanic, V.; Baranovic, G.; Smrecki, V. Structure and vibrational
spectra of conjugated acids of trans- and cis-azobenzene. J. Mol.
Struct. 2001, 569, 89-109.
(45) De Azevedo, W. F., Jr.; Canduri, F.; Da Silveira, N. J. F. Structural
basis for inhibition of cyclin-dependent kinase 9 by flavopiridol.
Biochem. Biophys. Res. Commun. 2002, 293, 566-571.
(46) Vesely, J.; Havlicek, L.; Strnad, M.; Blow, J. J.; Donella-Deana, A.;
Pinna, L.; Letham, D. S.; Kato, J.; Detivaud, L.; Leclerc, S.; Meijer,
L. Inhibition of cyclin-dependent kinases by purine analogues. Eur.
J. Biochem. 1994, 224, 771-786.
(47) Sherr, C. J. Principles of tumor suppression. Cell 2004, 116, 235-
246.
(48) Kotala, V.; Uldrijan, S.; Horky, M.; Trbusek, M.; Strnad, M.;
Vojtesek, B. Potent induction of wild-type p53-dependent transcrip-
tion in tumour cells by a synthetic inhibitor of cyclin-dependent
kinases. Cell. Mol. Life Sci. 2001, 58, 1333-1339.
(49) Wang, S.; Meades, C.; Wood, G.; Osnowski, A.; Anderson, S.; Yuill,
R.; Thomas, M.; Mezna, M.; Jackson, W.; Midgley, C.; Griffiths,
G.; Fleming, I.; Green, S.; McNae, I.; Wu, S. Y.; McInnes, C.;
Zheleva, D.; Walkinshaw, M. D.; Fischer, P. M. 2-Anilino-4-(thiazol-
5-yl)pyrimidine CDK inhibitors: Synthesis, SAR analysis, X-ray
crystallography, and biological activity. J. Med. Chem. 2004, 47,
1662-1675.
(50) Kontopidis, G.; Andrews, M. J. I.; McInnes, C.; Cowan, A.; Powers,
H.; Innes, L.; Plater, A.; Griffiths, G.; Paterson, D.; Zheleva, D. I.;
Lane, D. P.; Green, S.; Walkinshaw, M. D.; Fischer, P. M. Insights
into cyclin groove recognition: complex crystal structures and
inhibitor design through ligand exchange. Structure 2003, 11, 1537-
1546.
(51) Pflugrath, J. The finer things in X-ray diffraction data collection.
Acta Crystallogr. 1999, D55, 1718-1725.
(52) Murshudov, G. N.; Vagin, A. A.; Dodson, E. J. Refinement of
macromolecular structures by the maximum-likelihood method. Acta
Crystallogr. 1997, D53, 240-255.
(53) Lamzin, V. S.; Wilson, K. S. Automated refinement for protein
crystallography. Methods Enzymol. 1997, 277, 269-305.
(54) Uchoa, H. B.; Jorge, G. E.; Freitas Da Silveira, N. J.; Camera, J. C.,
Jr.; Canduri, F.; De Azevedo, W. F., Jr. Parmodel: a web server for
automated comparative modeling of proteins. Biochem. Biophys. Res.
Commun. 2004, 325, 1481-1486.
JM0605740