Cyclin-dependent kinase 4 inhibitors as a treatment for cancer. Part 2: Identification and optimisation of substituted 2,4-bis anilino pyrimidines

Article abstract of DOI:10.1016/S0960-894X(03)00203-8

Through chemical modification and X-ray crystallography we identified the 2,4-bis anilino pyrimidines as potent inhibitors of CDK4. Herein, we describe the optimisation of this series.

Full text of DOI:10.1016/S0960-894X(03)00203-8

Bioorganic & Medicinal Chemistry Letters 13 (2003) 2961–2966
Cyclin-Dependent Kinase 4 Inhibitors as a Treatment for Cancer.
Part 2: Identification and Optimisation of Substituted 2,4-Bis
Anilino Pyrimidines
Gloria A. Breault,* Rebecca P. A. Ellston, Stephen Green, S. Russell James,
Philip J. Jewsbury, Catherine J. Midgley, Richard A. Pauptit, Claire A. Minshull,
Julie A. Tucker and J. Elizabeth Pease*
AstraZeneca, Alderley Park, Macclesfield, Cheshire SK10 4TG, UK
Received 19 September 2002; accepted 20 December 2002
Abstract—Through chemical modification and X-ray crystallography we identified the 2,4-bis anilino pyrimidines as potent inhib-
itors of CDK4. Herein, we describe the optimisation of this series.
# 2003 Elsevier Ltd. All rights reserved.
The cyclin-dependent kinases (CDKs) are important in
controlling entry into and progression through the cell
cycle.1 They are a family of serine/threonine kinases
whose activity is regulated at several levels. The binding
of each CDK to a specific cyclin partner protein is
required for activity. The synthesis and degradation of
cyclins is tightly controlled such that their level fluc-
tuates duringthe cell cycle. It is these fluctuatinglevels
of the cyclin that cause first one, and then another
member of the CDK family to become active duringcell
cycle progression. Inhibitory proteins also regulate
CDK activity and these inhibitors need to be either
sequestered or destroyed to allow the CDK enzyme to
become fully activated. As cells respond to the presence
of mitogens, the level of cyclin D1 increases and even-
tually triggers the activation of CDK4 and CDK6.
These kinases phosphorylate the retinoblastoma tumour
suppressor protein pRb, thereby abrogating its inhibi-
tion of members of the E2F family of transcription fac-
tors. This then triggers a programme of gene expression
that results in entry into the S phase of the cell cycle.
enzyme that binds p16 has been mutated makingthe
enzyme resistant to inhibition.5 Together, these findings
imply that the deregulation of the pRb pathway, and
CDK4, is important in cancer progression. The inhibi-
tion of CDK4 may therefore be a valuable approach to
treat tumours especially those that have lost the natural
CDK4 inhibitor, p16.
A number of groups have identified CDK inhibitors.6À15
As described in the previous paper16 a series of 4,6-bis
anilino pyrimidines were identified from
a high-
throughput screening campaign17 as inhibitors of
CDK4. Duringfurther investigation of this series, sev-
eral X-ray structures were obtained with the structural
surrogate CDK2 in order to improve our understanding
of the compound’s bindingmodes From these CDK2
complex structures, it was evident that the pyrimidine
N1 acts as a hydrogen bond acceptor with the 6-aniline
16
NH servingas a hydrogen bond donor.
There are now many examples where tumours contain
multiple copies of the cyclin D1 gene or express abnor-
mally high levels of the protein.2 Similarly, many
tumours have been described that contain mutations,
3,4
deletions or silencingof the p16 or the pRb egne.
Finally, mutations of CDK4 itself have been described
within melanoma patients, where the region of the
In contrast, it appeared the pyrimidine N3 was not
involved in any specific interaction. This led us to
*Correspondingauthor. E-mail: j.elizabeth.pease@astrazeneca.com
0960-894X/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00203-8

2962
G. A. Breault et al. /Bioorg. Med. Chem. Lett. 13 (2003) 2961–2966
Scheme 2. Synthesis of 2-1.
Scheme 1. Synthesis of 2,4-bis anilino pyrimidines.
with the analogous 4,6-bis anilino pyrimidine 1B (IC50
33 mM) and, therefore, we decided to explore the series
further. Our aim was to confirm the relative binding
modes, and explore the possibility of overlappingSAR
between the two series, and then to utilise this informa-
tion to obtain improvements in potency.
speculate about the likely utility of a 2,4-bis anilino
pyrimidine (i.e., where the pyrimidine N3 is moved to
the equivalent of the 5 position). We considered the
possibility that the pyrimidine 5H of the 4,6 series may
contribute to the twistingof the aniline rings out of the
plane of the pyrimidine in the bioactive conformation.
If the twisted conformation were crucial to the activity
then the 2,4 series may prove to be less potent. Despite
these concerns, we initiated chemistry around the 2,4-bis
anilino pyrimidines.
The route used for the preparation of the 2,4-bis anilino
pyrimidines is shown in Scheme 1. The 2,4-dichloropyri-
midine was reacted with the appropriate aniline to
introduce ringB. In this step, butanol was pre-
dominately the solvent of choice although NMP was
occasionally employed for poorly nucleophilic anilines.
Conversion of intermediate 1-1 to the final product
could be achieved in a stepwise fashion or, more con-
vergently, by reaction with the functionalised aniline 2-1
(Scheme 2) in butanol with acid catalysis to yield the
product in one step.
With the synthetic route established we were able to
explore the SAR and turned our attention to investi-
gating the effects of substitution in ring B. From the
initial compounds prepared, results indicated that the
2,4 substituted pyrimidines were generally more potent
The first compound prepared was the 2,4-bis anilino
pyrimidine 1A (IC50 2 mM); this compared favourably
Table 1. Structures and enzyme activity for the 2,4-bis anilino pyrimidines compared to the 4,6 bis anilino pyrimidines
Compd
X
A
B
CDK4 IC50 (mM)a
CDK2 IC50 (mM)a
CDK4 IC50 (mM)a
CDK2 IC50 (mMa)
1
2
3
All H
2-F
2-Cl
2
1
1
34
7
6
33
27
8
>100
>100
95
4
5
6
7
8
9
10
2-Br
1
2
19
2
1
7
4
4
28
***
85
***
>100
>100
>100
2-CN
2-OCH3
3-Cl
4-Cl
4-OCH3
4-CH3
***
23
***
41
57
25
15
5
18
36
26
2
2
aValues are means of at least two determinations. The assay-to-assay variation was generally Æ2-fold based on the results of a standard compound.

G. A. Breault et al. /Bioorg. Med. Chem. Lett. 13 (2003) 2961–2966
2963
Table 2. Structures and enzyme activity for the disubstituted 2,4-bis anilino pyrimidines compared to the disubstituted 4,6-bis anilino pyrimidines
Compd
X
A
B
CDK4 IC50 (mM)a
CDK2 IC50 (mM)a
CDK4 IC50 (mM)a
CDK2 IC50 (mM)a
3
2-Cl
1
6
1
2
5
1
8
1
4
2
6
95
22
81
22
35
11
12
13
14
2-Cl, 5-Cl
2-F, 5-CH3
2-Cl, 5-CH3
2-F, 5-CF3
0.8
0.7
2
0.6
aValues are means of at least two determinations. The assay-to-assay variation was generally Æ2-fold based on the results of a standard compound.
than the 4,6-equivalent with improvements in IC50 ꢀ10-
fold (Table 1) and in selectivity (for CDK4 over CDK2)
generally being ꢀ10-fold. There appeared to be little
variation in potency over the range of substituents in the
2,4 compounds chosen 1A–10A, unlike their 4,6 counter-
parts, 1B–10B. Interestingly, some substituents did not
conform to the general activity trend. For example, the
isomeric 2-bromo inhibitors 4A and 4B have virtually
identical potencies. Also of note is compound 6A, sub-
stituted with methoxy at the 2-position, as it is markedly
less potent as well as havingno selectivity versus CDK2.
obtained. The bindingmode of the 2,4 series is the same
as found earlier for the 4,6 series,16 utilizinga single
pyrimidine nitrogen and the A-ring aniline NH to form
hydrogen bonds to the protein backbone. However, the
tilt of the B-ringwith respect to the pyrimidine was
found to vary substantially between different enzyme/
inhibitor complex structures both within and between
the two series. This may go some way to explaining why
the SAR of the two series is difficult to understand and
do not appear to correlate.
Havingestablished that N-alkylation improved potency
in the 4,6 series, we decided to look at the effects within
the 2,4 series since the X-ray structure of the CDK2–
11A complex suggested displacement of the water was
still possible. The alkylated compounds were obtained
by reaction of intermediate 1-1 with the appropriate
alkyl bromide in DMF with potassium carbonate as
base (Scheme 3). This was ꢁthen reacted with inter-
mediate 2-1, by heatingat 95 C in butanol overnight.
As disubstitution had proved a successful way to
improve potency in the 4,6 series, we decided to investi-
gate the effects in the 2,4 series (Table 2). In contrast to
the 4,6 series, the disubstituted compounds showed no
substantial improvement in potency over their mono-
substituted analogues. For example, the 2,5-dichloro
analogue 11A showed an IC50 similar to that for the
simple 2-chloro 3A. Note also that there has been an
apparent decline in selectivity with the disubstituted
compounds, CDK4 versus CDK2. The reason for this
reduction in selectivity remains unclear.
Table 3 shows the activity results for the alkylated
compounds. The improvement in potency was less
The X-ray structure18 for the complex of CDK2 with
the 2,4 disubstituted compound 11A (Fig. 1) was
Table 3. Structures and enzyme activity for the alkylated 2,4-bis
anilino pyrimidines
Compd
CDK4 IC50 (mM)a
CDK2 IC50 (mM)a
R
11
15
16
17
18
19
0.8
0.1
0.2
0.3
1
1
0.7
2
H
CH2CN
CH2CCH
CH2CH2CN
CH2Ph
3
>10
12
0.2
(CH2)3CF3
aValues are means of at least two determinations. The assay-to-assay
variation was generally Æ2-fold based on the results of a standard
compound.
Scheme 3. Synthesis of the alkylated 2,4-bis anilino pyrimidines.

2964
G. A. Breault et al. /Bioorg. Med. Chem. Lett. 13 (2003) 2961–2966
marked than we had expected (ꢀ10-fold at best) by
comparison to the 4,6 series.16 In addition, the different
N-alkyl groups showed only modest variation in
potency. However, in keepingwith the findings in the
4,6 series,16 the cyanomethyl 15 still delivered good
potency but now stood out less from the acetylene 16. A
drop in potency was still observed as the chain was
lengthened by one carbon unit 17, but, bulky sub-
stituents (18 and 19) were better tolerated in this series.
It should also be noted that the N-alkylation also gave
improved selectivity with respect to CDK2 compared to
the disubstituted compounds (Table 2), possibly due to
a reduced freedom to rotate around the pyrimidine–4N
bond.
Lookingat the X-ray structure of CDK2 with 11A (Fig.
1), it was evident that the part of the bindingsite local
to the N-alkyl substituent could also be reached by pla-
cinga group at the 5-position of the pyrimidine. For
this reason, we decided to investigate substitution at the
5-position. Compounds were prepared from the com-
mercially available 5-substituted uracils by conversion
to the correspondingdichloropyrimidines usingphos-
phorus oxychloride and phosphorus pentachloride at
reflux. The synthetic route detailed in Scheme 2 could
then be followed (Table 4).
Figure 2. Electron density at the ATP-bindingsite in the X-ray struc-
ture of CDK2 (2Fo-Fc; contoured at 1.5s, green density) bound to 23
(Fo-Fc; contoured at 2.5s, red density). Apart from the solubilising
group, which is modelled as the racemate, the ligand is very well
defined. The bridging water is not displaced in this structure and the
5-bromo substituent stacks against Phe80.
Table 4. Structures and enzyme activity for the 5-substituted 2,4-bis
anilino pyrimidines
Figure 1. (a) X-ray structure of CDK2 complexed with 11A showing
final Fo-Fc difference electron density calculated for a model from
which 11A was omitted (red, contoured at 2.5s) for the bound ligand
in the ATP-bindingpocket. Selected nearby protein residues are
shown. Hydrogen bonding interactions with the protein are indicated
and are as expected. RingB is well defined. The solubilisingsub-
stituent on ringA has been modelled as a racemic mixture and is only
partially ordered. The bridging water molecule between the 4-anilino
NH and Asp145 is shown in green density. Other solvent molecules
have been omitted for clarity. Note that Lys33, which normally forms
a salt bridge with Asp145, has electron density (not shown for clarity)
that indicates it has been covalently modified: it is carboxylated. (b)
Superposition (based on Ca positions of CDK residues 80–86) of 11A
(green) and 14B16 (yellow) suggesting that, as for 14B, N-alkylation to
displace the water molecule may also be a route to improvingpotency
Compd
CDK4 IC50 (mM)a
CDK2 IC50 (mM)a
X
1
2
34
8
>10
0.5
0.3
H
CH3
F
Cl
Br
20
21
22
23
0.8
0.3
0.1
0.1
aValues are means of at least two determinations. The assay-to-assay
variation was generally Æ2-fold based on the results of a standard
compound.
in 11A. Figures
Raster3D.27À29
1 and 2 were prepared usingBobscript and

G. A. Breault et al. /Bioorg. Med. Chem. Lett. 13 (2003) 2961–2966
2965
Initially, we explored multi-atom substituents at the
5-position attemptingto pick up the interactions that
N-alkylation provided. However we were unable to
improve activity. Lookingmore broadly, we found that
halogens, especially chloro 22 and bromo 23, improved
the potency against both CDK2 and CDK4. However,
this improvement was not accompanied by the good
selectivity that had been observed for the N-alkylated
compounds (Table 2). These results suggested that the
smaller 5-substituents were not occupyingthe same part
of the bindingsite as the N-alkyl group. The X-ray
structure of CDK2 in complex with 23 was determined,
and, as anticipated, the bridging water molecule
remained in place and the bromo substituent packed
against Phe80 (Fig. 2).
2. Barnes, D. M.; Gillet, C. E. Breast Cancer Res. Treat. 1998,
52, 1.
3. Roussel, M. F. Oncogene 1999, 18, 5311.
4. Nevins, J. R. Hum. Mol. Genet. 2001, 10, 699.
5. Tsao, H.; Benoit, E.; Sober, A. J.; Thiele, C.; Haluska, F. G.
Cancer Res. 1998, 58, 109.
6. Kimball, S. D.; Webster, K. R. In Annual Reports in Med-
icinal Chemistry; Doherty, A. M., Ed.; Academic: San Diego,
2001; Vol. 36 p 139, and references therein.
7. Ikuta, M.; Kamata, K.; Fukasawa, K.; Honma, T.;
Machida, T.; Hirai, H.; Suzuki-Takahashi, I.; Hayama, T.;
Nishimura, S. J. Biol. Chem. 2001, 276, 27548.
8. Carini, D. J.; Kaltenbach, R. F.; Liu, J.; Benfield, P. A.;
Boylan, J.; Boisclair, M.; Brizuela, L.; Burton, C. R.; Cox, S.;
Grafstrom, R.; Harrison, B. A.; Harrison, K.; Akamike, E.;
Markwalder, J. A.; Nakano, Y.; Seitz, S. P.; Sharp, D. M.;
Trainor, G. L.; Sielecki, T. M. Bioorg. Med. Chem. Lett. 2001,
11, 2209.
Increasingthe bulk of the 5-substituent lead to
increased activity, possibly because the bioactive ringB
orientation becomes a more favourable small molecule
conformation (1 cf. 20). Increasingthe 5-substituent
polarisability also leads to increased activity, possibly
due to improved interactions with the p cloud of the
Phe80 aromatic ring( 21–23).
9. Honma, T.; Hayashi, K.; Aoyama, T.; Hashimoto, N.;
Machida, T.; Fukasawa, K.; Iwama, T.; Ikeura, C.; Suzuki-
Takahashi, I. J. Med. Chem. 2001, 44, 4615.
10. Fry, D. W.; Bedford, D. C.; Harvey, P. H.; Fritsch, A.;
Keller, P. R.; Wu, Z.; Dobrusin, E.; Leopold, W. R.; Fattaey,
A.; Garrett, M. D. J. Biol. Chem. 2001, 276, 16617.
11. Soni, R.; O’Reilly, T.; Furet, P.; Muller, L.; Stephan, C.;
Zumstein-Mecker, S.; Fretz, H.; Fabbro, D.; Chaudhuri, B. J.
Natl. Cancer Inst. 2001, 93, 436.
This led us to consider whether combining5-substitu-
tion and N-alkylation might provide compounds that
are both potent and selective. These compounds were
prepared accordingto Scheme 3 startingwith the
appropriately substituted dichloropyrimidine. Our work
in this area led us to compound 24 that had an IC50 for
CDK4 of 0.01 mM. The combination of 5- substitution
and N-alkylation gave not only good potency versus
CDK4 but also good selectivity (24, CDK2 IC50 0.2
mM).
12. Honma, T.; Yoshizumi, T.; Hashimoto, N.; Hayashi, K.;
Kawanishi, N.; Fukasawa, K.; Takaki, T.; Ikeura, C.; Ikuta,
M.; Suzuki-Takahashi, I.; Takashi, H.; Nishimura, S.; Mor-
ishima, H. J. Med. Chem. 2001, 44, 4628.
13. Soni, R.; Muller, L.; Furet, P.; Schoepfer, J.; Stephan, C.;
Zumstein-Mecker, S.; Fretz, H.; Chaudhuri, B. Biochem. Bio-
phys. Res. Commun. 2000, 275, 877.
14. Jeong, H.-W.; Kim, M.-R.; Son, K.-H.; Young Han, M.;
Ha, J.-H.; Garnier, M.; Meijer, L.; Kwon, B.-M. Bioorg. Med.
Chem. Lett. 2000, 10, 1819.
15. Ryu, C.-K.; Kang, H.-Y.; Lee, S. K.; Nam, K. A.; Hong,
C. Y.; Ko, W.-G.; Lee, B.-H. Bioorg. Med. Chem. Lett. 2000,
10, 461.
16. Beattie, J. F.; Breault G. A.; Ellston, R. P. A.; Green
S.; Jewsbury, P. J.; Midgely C. J.; Naven, R. T.; Pauptit,
R.A.; Tucker, J. A.; Pease, J. E. Bioorg. Med. Chem. Lett.
In press.
17. Breault, G. A.; Pease, J. E. PCT Int. Application WO
0012485 A1 20000309.
18. Protein and crystals were obtained accordingto estab-
lished procedures19,20 Diffraction data were collected on
beamline 14.3 at ESRF, Grenoble, at 100 K. Data processing,
data reduction and structure solution by molecular replace-
In conclusion, we have demonstrated that usingour
understandingof the bindingmode, we were able to
convert our initial 4,6-bis anilino pyrimidine series to
the 2,4-bis anilino pyrimidines. The 2,4-bis anilino pyri-
midines are also potent and selective inhibitors of
CDK4. However potency and SAR are not mirrored in
the two series which we believe is due to conformational
variability in particular of the B ring, and the scope for
different substitution in the pyrimidine core.
21
ment were carried out usingprograms from the CCP4 suite.
Compounds 11A and 23 were modelled as racemic mixtures of
the chiral alcohol usingInsightII, 22 and modelled into electron
23
density usingQUANTA.
24
The protein complex model was
and final structures25,26 have been
refined usingCNX,
deposited in the Protein Data Bank with deposition codes
1h01 and 1h08 together with structure factors and detailed
experimental conditions.
19. Lawrie, A. M.; Noble, M. E.; Tunnah, P.; Brown, N. R.;
Johnson, L. N.; Endicott, J. A. Nat. Struct. Biol. 1997, 4, 796.
20. Legraverend, M.; Tunnah, P.; Noble, M.; Ducrot, P.;
Ludwig, O.; Grierson, D. S.; Leost, M.; Meijer, L.; Endicott,
J. J. Med. Chem. 2000, 43, 1282.
Acknowledgements
We are grateful for the excellent support we have
received from Michael Block and would also like to
thank Sandra Oakes for repeat enzyme measurements.
21. CCP4 Acta Crystallogr. 1994, D50, 760.
22. InsightII, Accelrys.
23. Quanta2000, Accelrys.
24. CNX version 2000.1, Accelrys.
References and Notes
25. Crystallographic statistics for 11A are: space group
P212121, unit cell 53.1, 70.7, 72.5 A, resolution 1.8 A, 24527
1. Sherr, C. J. Science 1996, 274, 1672.

2966
G. A. Breault et al. /Bioorg. Med. Chem. Lett. 13 (2003) 2961–2966
unique reflections from 58,865 observations give 94% com-
pleteness with Rmerge=3.3% and mean I/s(I) of 14.The final
model containing2247 protein, 244 water, 12 glycerol and 37
inhibitor atoms has an R-factor of 19% (Rfree using5% of
data is 22%). Mean temperature factor for protein is 17.5 and
for ligand is 17.3 A2.
ness with Rmerge=5.7% and mean I/s(I) of 18. The final
model containing2247 protein, 181 water, six glycerol and 36
inhibitor atoms has an R-factor of 21% (Rfree using5% of
data is 24%). Mean temperature factor for protein is 32.7 and
for ligand is 40.8 A2.
27. Esnouf, R. J. Mol. Graph. 1997, 15, 132.
26. Crystallographic statistics for 23 are: space group P212121,
unit cell 53.2, 71.9, 72.0 A, resolution 1.8 A, 26,090 unique
reflections from 106,969 observations give 99.5% complete-
28. Meritt, E. A.; Murphy, M. E. P. Acta Crystallogr. 1994,
D50, 869.
29. Bacon, D.; Anderson, W. F. J. Mol. Graph. 1998, 6, 219.