The synthesis and SAR of 2-amino-pyrrolo[2,3-d]pyrimidines: A new class of Aurora-A kinase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2006.08.080

A new class of Aurora-A inhibitors have been identified based on the 2-amino-pyrrolo[2,3-d]pyrimidine scaffold. Here, we describe the synthesis and SAR of this novel series. We report compounds which exhibit nanomolar activity in the Aurora-A biochemical

Full text of DOI:10.1016/j.bmcl.2006.08.080

Bioorganic & Medicinal Chemistry Letters 16 (2006) 5778–5783
The synthesis and SAR of 2-amino-pyrrolo[2,3-d]pyrimidines:
A new class of Aurora-A kinase inhibitors
Kevin J. Moriarty,* Holly K. Koblish, Thomas Garrabrant, Jahanvi Maisuria,
Ehab Khalil, Farah Ali, Ioanna P. Petrounia, Carl S. Crysler, Anna C. Maroney,
Dana L. Johnson and Robert A. Galemmo, Jr.
Johnson and Johnson, PRDUS, LLC, Spring House, PA 19477, USA
Received 28 June 2006; revised 16 August 2006; accepted 17 August 2006
Available online 1 September 2006
Abstract—A new class of Aurora-A inhibitors have been identified based on the 2-amino-pyrrolo[2,3-d]pyrimidine scaffold. Here, we
describe the synthesis and SAR of this novel series. We report compounds which exhibit nanomolar activity in the Aurora-A
biochemical assay and are able to inhibit tumor cell proliferation. This study culminates in compound 30, an inhibitor with potent
activity against Aurora A (IC50 = 0.008 lM), anti-proliferative activity against several tumor cell lines and induces polyploidy in
H460 cells.
Ó 2006 Elsevier Ltd. All rights reserved.
Aurora kinases have recently gained prominence as a
new oncology drug target because of the critical role
they play in the regulation of the cell cycle, especially
in the later stages from the G2/M checkpoint through
the mitotic checkpoint and late mitosis. Three Aurora
kinase paralogs are expressed in mammals and at least
two of the Aurora kinases (Aurora-A and -B) are com-
monly overexpressed in human tumors including breast,
lung, colon, ovarian, and pancreatic cancers. Overex-
pression of Aurora-A leads to centrosome amplification
and aneuploidy, and has also been shown to compro-
mise spindle checkpoint function, allowing anaphase
to occur despite continued activation of the spindle
checkpoint.1,2 Furthermore, Aurora-A has been shown
to function as an oncogene.3,4 Overexpression of Auro-
ra-B has also been reported to produce multi-nuclearity
and induce aggressive metastasis, suggesting that the
overexpression of Aurora-B has multiple functions in
cancer development.5,6 Inhibition of Aurora kinase
activity has been shown to induce endoreduplication
and eventual cell death in cancer cells.
These findings suggest that overexpression of the Auro-
ra kinases may play a critical role leading to chromo-
somal instability and inhibiting their kinase activity
may have a therapeutic benefit in the treatment of can-
cer. Indeed, VX-680 (1), a potent inhibitor of Aurora-
A, -B, and -C kinases, has been shown to suppress
tumor growth in vivo and has now progressed into
Phase 1 clinical trials.7
NH
N
N
HN
N
H
N
O
N
S
N
1
In an effort to discover novel Aurora kinase inhibi-
tors, our compound collection was screened against
Aurora-A kinase. The pyrrolopyrimidine 2 was identi-
fied as a starting point for further optimization. This
compound showed an excellent level of Aurora-A
enzyme inhibition and a selectivity profile warranting
further exploration. Here, we wish to report the syn-
thesis and in vitro SAR of this new class of Aurora-
A inhibitors.
Keywords: Aurora-A; 2-Amino-pyrrolo[2,3-d]pyrimidine; Anti-prolif-
erative; Polyploidy.
*
Corresponding author. Tel.: +1 011
kevinjmoriarty@comcast.net
1 610 584 1791; e-mail:
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.08.080

K. J. Moriarty et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5778–5783
5779
dichloropyrimidine (3) was treated with ammonia or 2
equiv of cyclohexylamine in THF at room temperature
for 10 min. This was stirred for 12 h at room tempera-
ture or 50 °C, in the case of the cyclohexylamine reac-
tion, to produce 4 and 5 in >95% yields. This was
followed by a palladium catalyzed cross-coupling reac-
tion with tributyl-(2-ethoxy-vinyl)-stannane and 4 or 5,
furnishing the 6 and 7 in 65% and 60% yields, respective-
ly. Heating to 120 °C in AcOH for 2 h converted 6 and 7
to the pyrrolopyrimidines 8 and 9 in yields >85%. The
use of AcOH was critical for the high conversion. The
use of HCl–CH3OH for the cyclization led to lower
yields (ꢀ45%) and the isolation of several by-products.
Cdk1
Cdk7
Cdk4
Aurora-A
IRK
c-Src
VEGF
PDGFr
TrkB
IC50 = 1.60μM
IC50 = 1.50μM
IC50 = 0.031μM
IC50 = 0.047μM
IC50 = 1.00μM
IC50 = 0.71μM
IC50 = 0.10μM
IC50 = 0.29μM
IC50 = 0.16μM
N
N
HN
N
N
N
2
The synthesis of the pyrrolopyrimidines is described in
Schemes 1–4. Using a modified procedure of a route
outlined by Cheung and co-workers8 5-bromo-2,4-
Br
Lam–Chan9 or Buchwald10 coupling conditions were
applied to attach the aryl groups to the N-7 atom of
8. Yields for the coupling reaction ranged from 45% to
75% (Scheme 2). The Lam–Chan protocol was applica-
ble for the introduction of meta- and para-substituted
aryl rings at the N-7 position, however it could not be
extended to ortho substituted aryl rings or heterocyclic
systems. In those cases the transformation was carried
out using the Buchwald procedure. The pyridine/
Cu(OAc)2 combination was found to be the optimal
conditions for the Lam–Chan reaction while the use
of Cs2CO3/N,N-dimethylcyclohexane-1,2-diamine at
140 °C in DMF was found to be the optimal conditions
for the Buchwald conditions.
Br
N
N
c
a or b
Cl
N
NH
R
Cl
N
3
Cl
4 R = H
5 R = cyclohexyl
OEt
N
N
R
N
d
Cl
N
Cl
6 R = H
7 R = cyclohexyl
N
N
R
H
8 R = H
9 R = cyclohexyl
Scheme 1. Reagents and conditions: (a) NH3, THF, rt, 12 h; (b) 2
equiv cyclohexylamine, THF, 50 °C, 12 h; (c) Pd(PPh3)4, toluene,
tributyl-(2-ethoxy-vinyl)-stannane, 110 °C, 24 h; (d) AcOH, 120 °C,
2 h.
The 2-anilino-groups could be attached to 10 via two
methodologies (Scheme 3). When the aniline was suffi-
ciently nucleophilic and tolerated acidic conditions at
elevated temperatures the group could be introduced
thermally, otherwise, attachment was accomplished via
a palladium catalyzed cross-coupling reaction. The use
of the (t-Bu3P)2Pd/K3PO4 combination was critical for
the coupling reaction. Use of other palladium reagents
or bases yielded only trace amounts of the desired prod-
uct or no reaction.11 Microwave irradiation enhanced
the reaction rate shortening the reaction time from 24
to 48 h for the standard thermal reaction to 1 h. Yields
of 11 for the thermal and palladium reactions ranged
from 25–45% to 50–80%, respectively.
N
N
a or b
N
N
Ar(Het)
Cl
N
8
Cl
N
10
H
Scheme 2. Reactions and conditions: (a) Cu(OAc)2, ArB(OH)2,
0
˚
pyridine, 4 A mol sieves, 3d, CH2Cl2, rt; (b) CuI, Cs2CO3, N,N -
dimethylcyclohexane-1,2-diamine, Ar-X (X = I, Br) or Het-X (X = I,
Br), DMF, 140 °C, 24 h.
N
N
N
N
The second route involved heating 6 to 120 °C for 2 h in
the presence of anilines of sufficient nucleophilicity and
stability in AcOH. This resulted in addition of the ani-
line and cyclization to 12 in one pot and in yields of
45–65%. Attachment of aryl groups or heterocycles to
the N-7 atom could be done using the above-mentioned
conditions (Scheme 4).
a or b
Cl
N
HN
Ar1
N
Ar (Het)
Ar (Het)
10
11
Scheme 3. Reactions and conditions: (a) CH3OCH2CH2OH, cat HCl,
180 °C, 24 h; (b) (t-Bu3P)2Pd, K3PO4, DMF, 140 °C, 1 h, microwave,
Ar1NH2.
The compounds were tested for their ability to inhibit
the phosphorylation of serine 10 of histone-H3 by mur-
ine Aurora-A enzyme. The compounds were also count-
er screened against CDK4 to establish selectivity. Table
1 outlines the SAR observed around the 2-amino por-
tion of the pyrrolo[2,3-d]pyrimidine scaffold. Replace-
ment of the 4-(4-methylpiperazin-1-yl)phenyl group led
to an improvement in the overall selectivity versus
CDK4, most notably compound 15, which showed a
>500-fold improvement. Modeling studies indicated
that the 3-(phenyl)acetic acid of 15 was favorably
N
N
OEt
NH2
N
N
N
HN
N
HN
N
H
a
b or c
Ar(Het)
Cl
N
R
R
6
12
11
Scheme 4. Reagents and conditions: (a) RArNH2, AcOH, 120 °C, 2 h;
(b) Cu(OAc)2, ArB(OH)2, pyridine, 4 A mol sieves, 3d, CH2Cl2, rt; (c)
CuI, Cs2CO3, N,N0-dimethylcyclohexane- 1,2-diamine, Ar-X (X = I,
˚
Br) or Het-X (X = I, Br), DMF, 140 °C, 24 h.

5780
K. J. Moriarty et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5778–5783
Table 1. SAR for the 2-amino modifications
Table 2. SAR around the N-7 position
N
N
N
N
HN
R1
N
R
N
H3C
N
N
NH
a
Compound
R
IC50 (lM)
Compound R1
IC50 (lM)
a
Pi-Aurora CDK4
ELISA
Pi-Aurora CDK4
ELISA
ELISA
4-(4-Methylpiperazin-1-yl)phenyl 0.047
4-Morpholinophenyl 0.049
4-(2-Methoxyethyl)piperazin-1-yl 0.006
3-(Phenyl)acetic acid 0.0006
4-Phenylthiomorpholine dioxide 0.012
ELISA
2
17
18
19
20
21
Cyclohexyl
0.047
0.039
0.018
0.013
0.009
0.031
0.029
0.340
1.4
2
13
14
15
16
0.031
0.600
0.110
0.35
(Bicyclo[2.2.1]heptan-2-yl)
3-Methoxyphenyl
Pyridin-3-yl
3-(Methylsulfonyl)phenyl
2.0
2.1
1.00
2-(2,3-Dihydro-1H-indene) 0.018
a Values are means of three experiments.
a Values are means of three experiments.
positioned to form a salt bridge with the Arg137 of
Aurora-A while the 4-(4-methylpiperazin-1-yl) group
of 2 appeared to be making no observable H-bonding/
charge–charge interactions with the Aurora kinase
(Fig. 1). In CDK4 there is a glutamic acid situated in
the same relative position and introduction of the 3-
(phenyl)acetic acid moiety would potentially lead to an
unfavorable charge–charge interaction resulting in a de-
crease in potency against CDK4.
To optimize further and determine the optimal N-7
group we carried out a more extensive exploration of
this site. To get a clearer insight into the contributions
the aryl/heterocyclic groups have on enzymatic potency
the para-4-methylpiperazine ring was moved into the
meta position of the 2-anilino ring. Although both mod-
eling and a low resolution X-ray structure of 18 also
showed the para-4-methylpiperazine ring not making
any apparent interactions with the protein, it was be-
lieved that re-orienting the 4-methylpiperazine ring to
the meta position (26) would only further minimize
any possibilities of inhibitor–kinase interactions (Fig. 2).
Table 2 summarizes the SAR of the N-7 position of the
pyrrolopyrimidine ring. Replacement of the cyclohexyl
ring of 2 with a 3-pyridyl (19), 3-methylsulfonylphenyl
(20) or dihydroindene (21) resulted in a significant
improvement in the selectivity versus CDK4, even in
the presence of the 4-(4-methylpiperazine)phenyl group.
Introduction of a wide variety of a substituted aryl and
heterocyclic rings at the N-7 position was tolerated and
the results from the SAR studies are summarized in
Table 3. The nature and size of the substituents on the
aryl ring did have an effect on the potency. Introduction
of large groups (compare 26 to 27 and 28 to 29) caused a
loss in enzymatic potency while –SO2CH3 (23), –CN
(25), –OCH3 (26), and –CONH2 (28) functionality were
all tolerated.
Figure 1. Model of compounds 2 and 15 docked in Aurora-A. Blue,
compound 2; magenta, compound 15.
Figure 2. Model of compounds 18 and 26 docked in Aurora-A. Blue,
compound 18; magenta, compound 26.

K. J. Moriarty et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5778–5783
5781
Table 3. SAR of N-7 position
identified that exhibited sub-nanomolar potency in our
enzyme assay. As part of our protocol we tested 31
and 32 in human MV411 cells to assess cell penetration.
Overall we found the Caco-2 assay a poor predictor of
the cellular permeability characteristics of our inhibi-
tors. So to determine if any observed loss of cellular
activity against Aurora-A was potentially due to poor
cellular penetration we cross screened our inhibitors in
human MV411 cells. MV411s are an Flt3-driven cell line
and all of the most potent pyrrolopyrimidine analogs
exhibit excellent Flt3 enzyme activity with
IC50s < 10 nM, this allowed us to rapidly identify the
best compounds to further evaluate in other cell lines.
Compound 32 showed the best activity in both assays
and was selected for further profiling (Tables 5–8).
N
N
HN
N
R
N
N
Compound
R
Pi-Aurora ELISA
a
IC50 (lM)
22
23
24
25
26
27
28
29
30
Pyridin-3-yl
0.012
0.010
0.003
0.013
0.006
0.029
0.011
0.820
0.005
3-(Methylsulfonyl)phenyl
Phenyl
3-Benzonitrile
3-Methoxyphenyl
3-Isopropylphenyl
3-Benzamide
Compound 32 inhibited the phosphorylation of histone-
10
H3 Ser in several tumor cell lines as measured by the
Cellomics Arrayscan platform. Compound 32 also
exhibited anti-proliferative activity in a variety of hu-
man tumor cell lines while showing preferential activity
for cycling cells over non-cycling cells. This compound
was inactive in the hERG assay (>10 lM) and only
had activity against the A3 adenosine receptor (60% at
10 lM) in the 50 assay CEREP profile. Compound 32
was selective against a panel of kinases (Table 7) and
had good ADME properties in human and rat (Table
8). Further, Compound 32 induced accumulation of
cells with >4N DNA content and the polyploidy was
accompanied by an increasing sub-G1 fraction indica-
tive of apoptotic cell death (Fig. 3).
N-Cyclopentyl-3-benzamide
Pyridin-2-yl
a Values are means of three experiments.
Table 4. Proliferation assays
Compound
a
IC50 (lM)
A2870
A375
H460
23
24
26
30
NT
0.44
0.56
3.67
0.58
0.86
0.67
1.58
0.52
0.25
1.20
0.30
In summary, we have developed a series of 2-amino-pyr-
rolo[2,3-d]pyrimidine derivatives as novel Aurora-A
kinase inhibitors.12 An optimized analog 32 possessed
potent in vitro activity, good kinase selectivity, and
appropriate metabolic and pharmacokinetic properties.
The analog was able to induce polyploidy and apopto-
sis, two hallmarks of Aurora kinase inhibitors.
NT, not tested.
a Data presented here are means of three independent experiments.
Four compounds (23, 24, 26, and 30) were selected and
their anti-proliferative effects evaluated in three cell
lines. The results are summarized in Table 4. Com-
pounds 24 and 30 showed the best overall activity in
all three cell lines.
Experimental details: To measure the inhibition of auro-
ra kinase, Histone H3 (14-494, Upstate, Charlottesville,
VA) was exposed to compounds and 3 nM of purified
murine Aurora-A kinase in the presence of 100 mM
Hepes, pH 7.5, 1 mM DTT, 0.01% Tween 20, 10 lM
ATP, and 2 mM MgCl2. The reaction was terminated
after 35 min and activity was determined by measuring
Because of its excellent enzymatic activity and better cel-
lular potency we focused our optimization efforts on 30.
Holding the 2-pyridyl ring at N-7 constant, we concen-
trated our studies around the 2-amino position. From
the resulting SAR studies two analogs (31 and 32) were
Table 5. Modification of R1
Compound
R1
Pi-Aurora
ELISA IC50 (lM)
MV411 cell proliferation
IC50 (lM)
30
31
32
3-(4-Methylpiperazine)phenyl
3-(Phenyl)acetic acid
4-Phenylthiomorpholine dioxide
0.005
0.0006
0.0008
<0.013
0.67
<0.013

5782
K. J. Moriarty et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5778–5783
Table 6. Assay results for compound 32
Ser795 (#9301, Cell Signaling Technologies, Danvers,
MA) colorimetrically. Two compounds were used to
establish the assay VX6807 and ZM44743913 which
showed IC50s of 0.0006 and 0.039 lM, respectively.
N
N
O2S
N
N
N
H
N
Compounds were tested for their ability to affect the
proliferation of A2780 ovarian carcinoma, A375 mela-
noma, MV411 human leukemia cell and NCI-H460 lung
carcinoma cells grown in serum containing media. Cell
proliferation was determined by the differential lumines-
cence of cells before and after 72 h of exposure to com-
pound when measured by CellTiterGlo (Promega
Corporation, Madison, WI). Data were processed and
analyzed using GraphPad Prism (version 4).
32
Assays
IC50 (lM)
Cell proliferation
IC50 (lM)
Aurora-A
Flt3
Cell-based
0.0008
<0.001
0.082 (H460)
Confluent HMECs
A375
MES-SA
1.90
0.065
0.063
pHistoneH3
assay
0.299 (A375)
0.094 (A549)
A549
NCI-H460
HCT116
0.290
0.084
0.260
To measure the inhibition of Flt3, 20 lg/mL poly(Glu:-
Tyr) (4:1, Sigma P0275) was exposed to compounds and
purified human Flt3 [571–933] in the presence of
100 mM Hepes, pH 7.5, 113 lM ATP, 5 mM MgCl2,
1 mM DTT, and 0.01% Tween 20. The reaction was ter-
minated after 25 min, and the extent of poly(Glu:Tyr)
phosphorylation was quantitated by a fluorescent polar-
ization competition immunoassay (Ex: 485 nm, Em:
530 nm) in the presence of an anti-phosphotyrosine anti-
body and a phosphotyrosine tracer (Green Tyrosine
Kinase Assay Kit, PanVera P2837).
Table 7. Enzyme selectivity
Compound
IRK CDK1 Src PDGF c-Met Plk1 FAK
32
250 90
389 >1000 163
>1000 >1000
Selectivity ratio: IC50/IC50 (Aurora-A).
Table 8. ADME for compound 32
Molecular modeling studies were carried out using
Accelrys Insight II and Discovery Studio software. Visu-
alizations were done using Accelrys DS Viewer Pro 6.0
(http://www.accelrys.com) and PyMol v0.99 (http://
www.pymol.org). Initial molecular models were built
using the deposited Aurora-A structures 1MQ4 and
Caco-2a Caco-2a HLMb RLMb Clc t1/2 Rat
d
Compound
(A ! B) (B ! A)
16 15
%F
32
72
100
17 43 30
a app(x10-6).
b % Remaining at 10 min.
c (ml/min/kg) rat.
˚
1MUO. A low resolution crystal structure (3.1 A) of
d (po, min) rat.
murine Aurora-A with compound 18 was later obtained
that was sufficiently resolved to allow placement of the
inhibitor in the active site. This structure confirmed
our earlier models and was used in subsequent modeling
studies.
the phosphorylation of Ser10 of Histone H3 (9706, Cell
Signaling Technologies, Danvers, MA) colorimetrically.
To measure the inhibition of CDK4, retinoblastoma
(Rb) protein (12-439, Upstate, Charlottesville, VA)
was exposed to compounds and purified CDK4 in the
presence of 100 mM Hepes, pH 7.5, 75 mM NaCl,
2 mM DTT, 30 lM ATP, and 10 mM MgCl2. The reac-
tion was terminated after 35 min and activity was deter-
mined by measuring the phosphorylation of Rb on
Acknowledgments
The authors thank Jan Sechler, Stuart Emanuel, Angel
Fuentes-Pesquera, Christian Baumann, Heidi Ott,
Wendy Sun, and Rich Connors for counterscreen testing.
Control
0.5 M Compound 32
72 hours
0
50
100
150
200
250
0
50
100
150
200
250
Channels
4N
Channels
4N
2N
2N
8N
Figure 3. Flow cytometry. NCI-H460 cells were treated with compound 32 or DMSO control. The cells were harvested after 72 h and data were
obtained using a BD FACSCalibur and analyzed using ModFit LT 2.0.

K. J. Moriarty et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5778–5783
5783
A.; Demur, C.; Hercend, T.; Diu- Hercend, A.; Su, M.;
Golec, J. M. C.; Miller, K. M. Nat. Med. 2004, 10, 262.
8. Cheung, M.; Harris, P. A.; Lackey, K. E. Tetrahedron
Lett. 2001, 42, 999.
References and notes
1. Bischoff, J. R.; Plowman, G. D. Trends Cell Biol. 1999, 9,
454.
9. Lam, P. Y. S.; Clark, C. G.; Saubern, S.; Adams, J.; Averill,
K. M.; Chan, D. M. T.; Combs, A. Synlett 2000, 674.
10. Strieter, E. R.; Blackmond, D. G.; Buchwald, S. L. J. Am.
Chem. Soc. 2005, 127, 4120.
11. Lebedev, A. Y.; Khartulyari, A. S.; Voskoboynikov, A. Z.
J. Org. Chem. 2005, 70, 596.
12. During preparation of this manuscript a series of pyrrolo-
pyrimidine inhibitors of FAK kinase was disclosed. None of
our inhibitors exhibited any activity against the FAK kinase.
Choi, H.-S.; Wang, Z.; Richmond, W.; He, X.; Yang, K.;
Jiang, T.; Sim, T.; Karanewsky, D.; Gu, X.-j.; Zhou, V.; Liu,
Y.; Ohmori, O.; Caldwell, J.; Gray, N.; He, Y. Bioorg. Med.
Chem. Lett. 2006, 16, 2173.
13. Ditchfield, C.; Johnson, V. L.; Tight, A.; Ellston, R.;
Haworth, C.; Johnson, T.; Mortlock, A.; Keen, N.;
Taylor, S. S. J. Cell Biol. 2003, 161, 267.
2. Zhou, H.; Kuang, J.; Zhong, L.; Kuo, W. L.; Gray, J. W.;
Sahin, A.; Brinkley, B. R.; Sen, S. Nat. Genet. 1998, 20,
189.
3. Katayama, H.; Brinkley, W. R.; Sen, S. Cancer Metastasis
Rev. 2003, 22, 451.
4. Katayama, H.; Sasai, K.; Kawai, H.; Yuan, Z. M.;
Bondaruk, J.; Suzuki, F.; Fujii, S.; Arlinghaus, R. B.;
Czerniak, B. A.; Sen, S. Nat. Genet. 2004, 36, 55.
5. Ota, T.; Suto, S.; Katayama, H.; Han, Z. B.; Suzuki, F.;
Maeda, M.; Tanino, M.; Terada, Y.; Tatsuka, M. Cancer
Res. 2002, 62, 5168.
6. Tatsuka, M.; Katayama, H.; Ota, T.; Tanaka, T.;
Odashima, S.; Suzuki, F.; Terada, Y. Cancer Res. 1998,
58, 4811.
7. Harrington, E. A.; Bebbington, D.; Moore, J.; Rasmussen,
R. K.; Ajose-Adeogun, A. O.; Nakayama, T.; Graham, J.