Imidazole pyrimidine amides as potent, orally bioavailable cyclin-dependent kinase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2008.10.075

The development of a novel series of imidazole pyrimidine amides as cyclin-dependent kinase (CDK) inhibitors is described. The series was found to have much improved CDK2 inhibition and potent in vitro anti-proliferative effects against cancer cell lines. Control of overall lipophilicity was important to achieve good in vitro potency along with acceptable physiochemical properties and margins against inhibition of both CYP isoforms and the hERG potassium ion channel. A compound with an attractive overall balance of properties was profiled in vivo and possessed suitable physiochemical and pharmacokinetic profiles for oral dosing.

Full text of DOI:10.1016/j.bmcl.2008.10.075

Bioorganic & Medicinal Chemistry Letters 18 (2008) 6486–6489
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Imidazole pyrimidine amides as potent, orally bioavailable cyclin-dependent
kinase inhibitors
*
Clifford D. Jones , David M. Andrews, Andrew J. Barker, Kevin Blades, Kate F. Byth, M. Raymond V. Finlay,
Catherine Geh, Clive P. Green, Marie Johannsen, Mike Walker, Hazel M. Weir
Cancer and Infection Research, AstraZeneca Pharmaceuticals, Alderley Park, Macclesfield, Cheshire SK10 4TG, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
The development of a novel series of imidazole pyrimidine amides as cyclin-dependent kinase (CDK)
inhibitors is described. The series was found to have much improved CDK2 inhibition and potent
in vitro anti-proliferative effects against cancer cell lines. Control of overall lipophilicity was important
to achieve good in vitro potency along with acceptable physiochemical properties and margins against
inhibition of both CYP isoforms and the hERG potassium ion channel. A compound with an attractive
overall balance of properties was profiled in vivo and possessed suitable physiochemical and pharmaco-
kinetic profiles for oral dosing.
Received 20 August 2008
Revised 13 October 2008
Accepted 13 October 2008
Available online 22 October 2008
Keywords:
Kinase
Cyclin-dependent kinase
Kinase inhibitor
Imidazole amide
CDK
Ó 2008 Elsevier Ltd. All rights reserved.
Cell cycle
Cancer
The cyclin-dependent kinase (CDK) family is a group of serine–
threonine protein kinases with roles in the coordination of the
eukaryotic cell cycle and transcriptional regulation. The proper
regulation of CDKs involved in the cell cycle is crucial for the or-
dered execution of each phase of the cycle. CDKs are initially acti-
vated by formation of a complex with a cyclin partner protein,
followed by phosphorylation to yield the fully active complex.
The transient expression of the cyclin partner proteins at specific
stages of the cell cycle allows the normal process of cell division
to proceed under strict control.
progressed into clinical trials for the treatment of cancer (Fig. 1).
Recently, flavopiridol was demonstrated to have activity in pa-
tients with refractory chronic lymphocytic leukaemia (CLL).2
Our own efforts in this area began with the imidazo[1,2-a]pyr-
idine (5, A = CH) and imidazo[1,2-b]pyridazine (5, A = N) CDK
inhibitors which were described in earlier letters (Fig. 2).3–5
A characteristic of cancer is uncontrolled cell growth and prolif-
eration and most cancer cells show deregulation of CDKs. Extensive
profiling of tumour tissue has repeatedly identified components of
the CDK machinery that are altered in cancer. This commonly oc-
curs through amplification of cyclin effectors such as cyclins D
and E, the inactivation of endogenous inhibitors such as p16 and
p27, or genetic mutations to CDK substrates. Because of their crit-
ical role in the regulation of the cell cycle and the observed expres-
sion/activity pattern in most human cancers, considerable effort
has been focused on the development of small molecule CDK inhib-
itors as potential therapeutic agents.1
OH
O
Cl
HN
N
N
N
HO
O
N
OH
OH
N
N
1
3
2
N
NH2
N
O
F
O
S
O
O
S
N
H
S
N
N
N
O
O
As a result, a number of CDK inhibitors, such as flavopiridol 1,
roscovitine (CYC202) 2, BMS-387032 (SNS-032) 3 and R547 4 have
N
H
F
4
N
H
* Corresponding author. Tel.: +44 1625 513670.
E-mail address: cliff.jones@astrazeneca.com (C.D. Jones).
Figure 1. CDK inhibitors tested in clinical trials.
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.10.075

C. D. Jones et al. / Bioorg. Med. Chem. Lett. 18 (2008) 6486–6489
6487
Table 1
N
H
N
N
H
N
CDK inhibition profile for compounds 7a–e.
N
N
N
H
N
O
A
N
N
S
N
R
O
N
N
5
6
N
R
N
Figure 2. Previous AstraZeneca CDK inhibitor series.
Compound
R
CDK2 IC50
M)
LoVo IC50
M)a
hERG IC50
M)b
Solubilityc
(lM)
(
l
(
l
(
l
Improvements in physical properties and cellular potency were
achieved and led to the discovery of the imidazole sulphone
AZD5438 6, which was investigated further as an orally bioavail-
able anti-cancer agent.6
6
–SO2Me
–CONH2
–CONHMe
–CONMe2
–CONHcPr
<0.006
0.003
0.003
0.008
0.017
0.002
0.80
0.70
0.47
0.48
0.60
5.7
3.2
17
>30
18
7
>32
588d
140
540
>2300
70
1600
7a
7b
7c
7d
7e
The replacement of the sulphone in 6 with a piperazine ring re-
sulted in a new series of potent CDK inhibitors with further
improvements in physical properties, which were suitable for oral
dosing.7 We decided to investigate if there were other groups that
would provide suitable replacements for sulphone. Amide linkers
gave an attractive lipophilicity range and would also be able to
form similar protein–ligand interactions that were key for potency
and selectivity in the sulphonamide and sulphone series. Using the
routes shown in Scheme 1, an initial set of alkyl amides (7a–e)
were synthesised in the imidazole-5H-pyrimidine series (Table 1).
Excellent levels of enzyme potency against CDK2 were observed
for the simple primary, secondary and tertiary amides 7a–c, which
translated into good anti-proliferative effects in LoVo cancer cells.
Encouragingly, the levels of potency observed were similar to, or
exceeded, the clinical candidate AZD5438 6. Larger alkyl groups
generally gave lower levels of CDK2 potency 7d. The targeting of
enzyme solvent exposed regions with hydrophilic groups is com-
monly used in kinase inhibitors to improve physical properties
without impacting potency. Incorporation of a hydrophilic amide
substituent 7e retained good enzyme activity but with reduced
anti-proliferative cellular potency compared to 7a–d, due to lower
cellular permeability (data not shown). Generally, low levels of
activity against the hERG potassium ion channel were observed
for the less lipophilic examples and overall offered a significant
improvement in hERG binding over AZD5438 6. As a number of
these initial examples also possessed excellent levels of solubility
with no significant CYP inhibition,9 we were encouraged to explore
the series further.
–CONHC2H4OH
a
IC50 for inhibition of BrdU incorporation to LoVo cells following 48 h exposure
to test compound.
hERG patch-clamp assay.8
Aqueous equilibrium solubility measured over 24 h.
Mesylate salt.
b
c
d
yield. The required 4-bromobenzamides were readily obtained by
reacting the corresponding acid and amine together using standard
amide coupling conditions. This route was generally applicable, ex-
cept for the primary amides (7a and 8a), which gave exclusively
bis-arylation of the aminopyrimidine in the palladium catalysed
coupling step. To avoid these problems a two-step process was
used, first coupling 4-bromobenzonitrile, then hydrolysing the ni-
trile to the primary amide under basic conditions to give 7a and 8a.
As the methylamide 7b had an attractive balance of potency
and physical properties, we decided to maintain this amide substi-
tuent whilst exploring substitution around the imidazole ring.
Variation of the imidazole substitution to obtain compounds
11–15 was synthetically more challenging and used the routes
shown in Scheme 2. The substituent R4 was introduced by reduc-
tive amination of 5-methylisoxazole-4-amine, followed by amide
bond formation to incorporate the substituent R3. Hydrogenolysis
of the isoxazole followed by base-assisted cyclisation gave suitably
substituted imidazole methyl ketones. Treatment with DMF–DMA
gave the aminopropanones which on cyclisation with guanidine
carbonate gave the corresponding 2-aminopyrimidine imidazoles.
These were coupled with 4-bromobenzamides using similar
The routes used to synthesise imidazole pyrimidine amides are
shown in Scheme 1. Palladium catalysed coupling of the previously
reported 4-imidazoyl-2-aminopyrimidines (R1 = F or H)7,10 with
suitably substituted 4-bromobenzamides under Buchwald–Har-
twig conditions gave the desired products in good to excellent
R3
O
R4
O
R4
NH2.HCl
N
(b)
(a)
N
N
N
H
N
H2N
R3
R1
N
R1
(a)
N
N
O
N
N
O
Br
N
(c)
N
O
R2
N
R
N
N
O
N
R2
H2N
7 (R1 = H, R2 = H)
8 (R1 = F, R2 = H)
9 (R1 = H, R2 = F)
10 (R1 = F, R2 = F)
R
O
N
N
(e)
(d)
11-15
R4
R3
R4
(b)
(c)
N
N
N
H
N
N
R3
N
R1
Scheme 2. Synthesis of substituted imidazole aminopyrimidines 11–15. Reagents
and conditions: (a) i—R4 cPentyl: cyclopentanone, NaCNBH3, MeOH, 69%; R4 = iPr:
acetone, NaOAc, NaCNBH3, AcOH, MeOH, 59%; ii—R3 cPr: cyclopropylcarbonyl
N
=
N
=
N
chloride, NEt3, DCM, 40%; R3 = Me: Ac2O, NaOAc, AcOH, 91%; (b) i—H2, 4 atm, 10%
Pd/C, EtOH, 44–84%; ii—NaOH, 1,4-dioxane, 54–68%; (c) DMF–DMA, DMF, 75–96%;
(d) guanidine carbonate, BuOH, 81–83%; (e) Pd(OAc)2, Xantphos, Cs2CO3, 1,4-
dioxane, 33–71%.
Scheme 1. Synthesis of imidazole amides 7–10. Reagents and conditions:
(a) Pd(OAc)2, Xantphos, Cs2CO3, 1,4-dioxane, 75–83%; (b) 4-bromobenzonitrile,
Pd(OAc)2, Xantphos, Cs2CO3, 1,4-dioxane, 81%; (c) KOH, EtOH, water, 53%.

6488
C. D. Jones et al. / Bioorg. Med. Chem. Lett. 18 (2008) 6486–6489
Table 3
conditions to those shown in Scheme 1 to give the substituted
CDK inhibition profile for compounds 8–10 and 16.
imidazole amides (11–15).
The importance of an a
-branched substituent in the R4 position
N
H
N
R1
is demonstrated by the reduction in potency moving from isopro-
pyl (7b, Table 1) to ethyl substitution (11, Table 2). Larger groups
such as 12 did not show any improvements in potency whereas
N
N
O
R2
the
a-branched cyclopentyl 13 returned potency back to low nano-
R
N
molar levels. These observations can be attributed to
a-branched
substituents providing an appropriately sized lipophilic group to
give optimal hydrophobic contacts with the glycine-rich loop
around the ATP-ribose binding domain.
Compound R1 R2
R
CDK2 IC50
M)
LoVo IC50 hERG IC50 Solubility
M) M) M)
(l
(
l
(
l
(l
8a
8b
8c
9a
9b
9c
10a
10b
10c
16
F
F
F
H
H
H
F
F
F
H
H
H
F
F
F
F
F
F
NH2
<0.001
0.003
0.002
0.27
0.19
0.13
nd
0.14
0.36
0.10
0.07
0.17
2.3
20
12
11
10
16
3
20
19
6
36
14
11
33
71
130
6
Structural studies indicated that larger groups than methyl might
be tolerated at the imidazole R3 position. This potential alternative
route to the solvent exposed region was attractive as it would give
additional options to modify the physical properties of the series.
In addition, it offered the opportunity to block a potential site of
metabolism and modulate pharmacokinetic (PK) properties. How-
ever, we were surprised to find that substitution at the R3 position,
even with comparatively small groups (14 and 15), led to a relatively
large reduction in CDK2 potency. These data indicated that the com-
bination of (R4 = iPr and R3 = Me) gave the optimal balance of CDK
inhibition and physical properties in this series, so we retained this
R3/R4 combination to explore variation at other positions.
NHMe
NMe2
NH2
NHMe <0.001
NMe2
NH2
<0.001
0.006
<0.001
NHMe <0.001
NMe2
4
0.002
0.009
16
>2600
F
Me NMe2
>32
to the unsubstituted compound 8c. Again, the disadvantage of
increasing lipophilicity is observed in the lower levels of solubility
of 10a–c but with smaller effects on hERG activity.
The potency benefits of substituting the central pyrimidine core
with fluorine has been reported previously.7 This is believed to be
due to a stronger interaction with the hinge region of the protein
through acidification of the aniline N–H by the electronegative
fluorine atom. Consequently, the 5-fluoropyrimidine benzamides
(R1 = F) (Table 3) were attractive targets and they were synthesised
using the routes shown in Scheme 1.
Comparing each 5-H pyrimidine (7a–c, Table 1) with the equiv-
alent 5-F pyrimidine 8a–c shows a small but significant increase in
enzyme potency against CDK2. These potency increases, along with
improved cellular permeability through increased lipophilicity,
gave larger improvements in anti-proliferative activity with com-
pounds starting to approach a cellular potency of 100 nM. How-
ever, the increased lipophilicity of 5-F pyrimidines compared to
5-H pyrimidines also results in a decrease in aqueous solubility
and an increase in hERG activity.
An increase in potency was also observed by substitution with
fluorine ortho to the secondary amide 9b, a smaller potency in-
crease was seen with the tertiary amide 9c. This indicated that
an internal fluorine–hydrogen bond between the secondary amide
and adjacent fluorine atom could be stabilising the biologically ac-
tive conformation. Incorporating both of these beneficial fluorine
substitutions into the same molecule gave compounds 10a–c. Both
the primary amide 10a and the methylamide 10b possessed excel-
lent levels of enzyme and anti-proliferative activity. Again, the ter-
tiary amide 10c showed a smaller benefit and had a similar profile
Interestingly, methyl substitution adjacent to the amide 16 led to
a large increase in solubility, despite the increase in lipophilicity
compared to 9c. We hypothesised that this is due to a steric clash be-
tween the methyl group and the dimethylamide adversely affecting
packing in the solid state. This possible change in conformation
could also be responsible for the much lower activity observed
against the hERG ion channel (cf. 9c with 16). Similar beneficial ef-
fects on potency and hERG activity were observed with other
ortho-substituted tertiary amides (data not shown) but substitution
was generally detrimental for potency so they were not pursued
further.
A number of compounds had suitable characteristics for further
progression, but the methylamide 9b appeared to have the best
overall balance of enzyme and cellular potency, along with suitable
physical properties. Additional selectivity screening against a
range of other kinases showed that within the CDK kinase family
the methylamide 9b showed greater than 50 fold selectivity for
CDK2 over CDK1 and CDK4 (Table 4). Good levels of selectivity
against non-CDK family kinases were also observed when 9b was
tested against a broader panel of kinases, with IC50s of >10
lM
against the kinases: Csk, EGFR, FGFR-1, IGF-1R, JAK2, Zap70 and
PKA. No issues with plasma protein binding (PPB) were observed,
with high levels of free drug across a number of species.
With a promising in vitro profile the methyl amide 9b pro-
gressed into in vivo studies. Good pharmacokinetic (PK) parame-
ters were observed with moderate to low clearance in nude
mouse and rat (Table 5). Rat bioavailability was moderate but im-
Table 2
Table 4
CDK inhibition profile for compounds 11–15.
Additional data for the methylamide 9b.
N
H
N
N
H
N
N
R4
N
N
O
N
O
R3
NH
N
F
NH
N
Compound R3
R4
Et
CDK2 IC50
M)
LoVo IC50
M)
Solubility
M)
(l
(
l
(l
Parameters
CDK2 (IC50
CDK1 (IC50
CDK4 (IC50
LogD
,
,
,
l
l
l
M)
M)
M)
<0.001
0.06
0.05
11
12
13
14
15
Me
0.056
3.1
2.1
0.19
0.66
2.9
35
27
24
10
24
Me
Me
cPr
cPrCH2 0.042
cPentyl 0.002
iPr
0.081
0.062
3.0
CH2OMe iPr
PPB (% free) mouse/rat/dog/human
12/8/30/8

C. D. Jones et al. / Bioorg. Med. Chem. Lett. 18 (2008) 6486–6489
6489
Table 5
References and notes
Pharmacokinetic parameters of the methylamide 9b.
1. Sharma, P. S.; Sharma, R.; Tyagi, R. Curr. Cancer Drug Targets 2008, 8,
53.
2. Byrd, J. C.; Lin, T. S.; Dalton, J. T.; Wu, D.; Phelps, M. A.; Fischer, B.; Moran, M.;
Blum, K. A.; Rovin, B.; Brooker-McEldowney, M.; Broering, S.; Schaaf, L. J.;
Johnson, A. J.; Lucas, D. M.; Heerema, N. A.; Lozanski, G.; Young, D. C.; Suarez, J.-
R.; Colevas, A. D.; Grever, M. R. Blood 2007, 15, 399.
3. Anderson, M.; Beattie, J. F.; Breault, G. A.; Breed, J.; Byth, K. F.; Culshaw, J. D.;
Ellston, R. P. A.; Green, S.; Minshull, C. A.; Norman, R. A.; Pauptit, R. A.; Stanway,
J.; Thomas, A. P.; Jewsbury, P. J. Bioorg. Med. Chem. Lett. 2003, 13, 3021.
4. Byth, K. F.; Culshaw, J. D.; Green, S.; Oakes, S. E.; Thomas, A. P. Bioorg. Med.
Chem. Lett 2004, 14, 2245.
PK parameters
Dose ( mol/kg)
T1/2 (h)
Cl (mL/min/kg)
Vss (L/kg)
Mouse iva
Rat ivb
Rat oral
Rat oral
l
2.0
16.7
18.5
9.5
1.8
—
1.0
5.8
11.9
1.7
1.4
—
5
—
—
—
1.5
21
20
—
—
—
13.7
47
AUC (
l
M h)
Bioavailability (%)
a
Female nude.
Male Han-Wistar.
b
5. Byth, K. F.; Cooper, N.; Culshaw, J. D.; Heaton, D. W.; Oakes, S. E.; Minshull, C.
A.; Norman, R. A.; Pauptit, R. A.; Tucker, J. A.; Breed, J.; Pannifer, A.; Rowsell, S.;
Stanway, J. J.; Valentine, A. L.; Thomas, A. P. Bioorg. Med. Chem. Lett. 2004, 14,
2249.
6. Anderson, M.; Andrews, D. M.; Barker, A. J.; Brassington, C. A.; Breed, J.; Byth, K.
F.; Culshaw, J. D.; Finlay, M. R. V.; Fisher, E.; Gingell, H. H. J.; Green, C. P.;
Heaton, D. W.; Nash, I. A.; Newcombe, N. J.; Oakes, S. E.; Pauptit, R. A.; Roberts,
A.; Stanway, J. J.; Thomas, A. P.; Tucker, J. A.; Weir, H. M. Bioorg. Med. Chem. Lett.
2008, 18, 5487.
7. Finlay, M. R. V.; Acton, D. G.; Andrews, D. M.; Barker, A. J.; Dennis, M.; Fisher, E.;
Graham, M. A.; Green, C. P.; Heaton, D. W.; Karoutchi, G.; Loddick, S. A.;
Morgentin, R.; Roberts, A.; Tucker, J. A; Weir, H. M. Bioorg. Med. Chem. Lett.
2008, 18, 4442.
proved at higher doses indicating that the compound was a suit-
able probe compound for additional in vivo efficacy studies.
In conclusion, we have discovered a new series of imidazole
pyrimidine amides that show excellent inhibition of CDK2 and
anti-proliferative activity along with good physical properties and
acceptable pharmacokinetics that are suitable for dosing orally.
Acknowledgments
8. Schroeder, K.; Neagle, B.; Trezise, DJ.; Worley, J. J. Biomol. Screen. 2003, 8, 50.
9. Against Cyp isoforms: 3A4, 2D6, 2C9, 2C19 and 1A2.
10. Andrews, D.; Finlay, M. R.; Green, C.; Jones, C. PCT Int. Application WO2006-
064251.
We acknowledge the excellent technical expertise of the follow-
ing scientists: Sandra E. Oakes, Sarah Kearney, Joanne Wilson and
Louise Moss.