Discovery and SAR of oxindole-pyridine-based protein kinase B/Akt inhibitors for treating cancers

Article abstract of DOI:10.1016/j.bmcl.2006.04.005

We describe a series of potent and selective oxindole-pyridine-based protein kinase B/Akt inhibitors. The most potent compound 11n in this series demonstrated an IC₅₀ of 0.17 nM against Akt1 and more than 100-fold selectivity over other Akt iso

Full text of DOI:10.1016/j.bmcl.2006.04.005

Bioorganic & Medicinal Chemistry Letters 16 (2006) 3424–3429
Discovery and SAR of oxindole–pyridine-based protein kinase
B/Akt inhibitors for treating cancers
Gui-Dong Zhu,a,* Viraj B. Gandhi,a Jianchun Gong,a Yan Luo,a Xuesong Liu,a Yan Shi,a
Ran Guan,a Shayna R. Magnone,a Vered Klinghofer,a Eric F. Johnson,a Jennifer Bouska,a
Alexander Shoemaker,a Anatol Oleksijew,a Ken Jarvis,a Chang Park,b Ron De Jong,c
Tilman Oltersdorf,c Qun Li,a Saul H. Rosenberga and Vincent L. Girandaa
aCancer Research, GPRD, Abbott Laboratories, Abbott Park, IL 60064, USA
bStructural Biology, GPRD, Abbott Laboratories, Abbott Park, IL 60064, USA
cIDUN Pharmaceuticals Inc., 9380 Judicial Drive, San Diego, CA 92121, USA
Received 10 March 2006; revised 31 March 2006; accepted 3 April 2006
Available online 27 April 2006
Abstract—We describe a series of potent and selective oxindole–pyridine-based protein kinase B/Akt inhibitors. The most potent
compound 11n in this series demonstrated an IC50 of 0.17 nM against Akt1 and more than 100-fold selectivity over other Akt iso-
zymes. The selectivity against other protein kinases was highly dependent on the C-3 substitutions at the oxindole scaffold, with
unsubstituted 9e or 3-furan-2-ylmethylene (11n) more selective and 3-(1H-pyrrol-2-yl)methylene (11f) or 3-(1H-imidazol-2-yl)meth-
ylene (11k) less selective. In a mouse xenograft model, 9d, 11f, and 11n inhibited tumor growth but with accompanying toxicity.
ꢀ 2006 Elsevier Ltd. All rights reserved.
Protein kinases are a large family of diverse but related
enzymes that regulate nearly all aspects of cell growth,
differentiation, and division. Dys-regulation of one or
more protein kinases has been associated with a wide
spectrum of human cancers. Partially encouraged by
the clinical success of Gleevec (inhibitor of BCR-ABL,
PDGFR, and c-Kit), as well as Iressa (EGFR inhibitor),
a search for small molecule inhibitors of protein kinases
as anti-cancer chemotherapeutics has received increas-
ing attention.1 Among the superfamily of protein
kinases, protein kinase B, also called Akt, is a pivotal
component of the phosphatidylinositol 30-kinase/Akt
signal transduction pathway that regulates many pro-
cesses crucial to carcinogenesis.2 Overexpression of
Akt as a result of, for example, inactivation of tumor
suppressor PTEN has been found in a variety of human
tumors.3 At the genomic level, Akt1 and Akt2 have been
shown to be amplified in many cancer types.4 Therefore,
Akt has long been considered an attractive target for the
treatment of cancers. There have been a number of small
molecule inhibitors that partially target Akt.1 While the
majority of these inhibitors are ATP-competitive, Linds-
ley et al. reported a series of selective allosteric and
non-ATP-competitive diphenylquinoxaline- and diphe-
nylpyridine-based inhibitors of Akt that target the
pleckstrin homology (PH) domain of the protein
kinase.5,6
R1 R8
NH
N
NH2
8
1
O
N
1
8
1: R = R =H
in vivo
1
8
2: R = OH, R = H
1
8
3: R = Cl, R = H
1
8
4: R = H, R = CH
3
Figure 1. Compound 2 was assigned as a major metabolite of Akt
inhibitor 1 in several species. Blocking the C-1 site of metabolism of
the isoquinoline provided compounds, for example, 3 or 4, with
respectable pharmacokinetic profile but with diminished Akt activity.
Keywords: Akt inhibitor; Protein kinase B; PKB; GSK3; FL5.12-Akt;
Anticancer.
*
Corresponding author. Tel.: +1 847 935 1305; fax: +1 847 935
5165; e-mail: gui-dong.zhu@abbott.com
0960-894X/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.04.005

G.-D. Zhu et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3424–3429
3425
We previously described a novel, potent (IC50 = 2.0 nM),
and selective Akt inhibitor 1 and showed significant effica-
cy in several mouse xenograft models7 (see Fig. 1). Major
shortcomings of this Akt inhibitor as a clinically useful
agent include short half-life in animals and poor oral bio-
availability. The C-1 position of the isoquinoline was
identified as a major site of metabolism, however, any
modification failed to provide potency against Akt.8
Herein, we have explored a number of alternative hetero-
cyclic scaffolds to replace the metabolically labile isoquin-
oline, and describe a new series of potent and selective
oxindole–pyridine based inhibitors.
through alkylation of 5-bromooxindole 5e with the
appropriate alkyl halides in the presence of base. Con-
densation of 5e with acetone and 5g with hydroxylamine
provided aryl bromides 5s and 5h. Fluorination of the
isatin 5g with DAST under standard conditions
(CH2Cl2, rt) afforded difluoro-analog 5t.
The synthesis of Akt inhibitors 11a–11p is outlined in
Scheme 2. Stille coupling of 5-bromooxindole 5e with
trimethylstannane 7 in the presence of Pd2(dba)3 and
(o-tol)3P furnished 10 in 47% yield. Treatment of 10 with
a variety of aldehydes in the presence of piperidine affor-
ded, after BOC-deprotection, compounds 11a–11p in
poor to modest yields. Alternatively, these compounds
(except for 11h) could be prepared on larger scale by ini-
tial condensation with an aldehyde, followed by Stille
reaction in improved overall yield. The stereochemistry
of the exo-indole double bond was deduced from related
Depicted in Scheme 1 is a general synthesis of com-
pounds 9a–9t. Stille reactions of either an arylbromide
5 or arylchloride 6 with trimethylstannane 78, under
the catalysis of tri-o-tolylphosphine (for bromide 5) or
2-dicyclo-hexylphosphino-20-(N,N-dimethyl-amino)bi-
phenyl(Cy-MAP, for chloride 6), provided compound 8.
Treatment of 8 with trifluoroacetic acid in methylene
chloride afforded Akt inhibitor 9. 5d, 5e, 5g, and 6 were
all commercially available, while 5a,9 5b,9 5f,10 5i,11 5j,11
5k,11 5l,12 5m,12 and 5n13 were prepared according to lit-
erature procedures. Aryl bromides 5o–5r were obtained
1
analogs in the literature14 and confirmed by H NMR
NOE studies. The ratio of the geometric isomers was
calculated from their H NMR spectra.
1
Table 1 highlights the heterocyclic scaffolds that we
investigated as potential replacements for the
BOC
NH
NH
Sn
O
N
NH
NHR
7
Ar-Br (5)
Ar-Cl (6)
R
O
i or ii
N
8: R=BOC
9: R=H
iii
Br
Br
Br
H
S
Br
N
O
O
O
Br
Br
O
O
N
N
N
N
S
N
H
H
5a
5b
5d
5i
5e
5f
BOC
CN
OH
BOC
O
N
N
N
Br
Br
Br
Br
N
N
N
Br
O
O
O
N
N
H
N
H
H
BOC
CN
5g
5h
5j
5k
H
N
O
O
S
Br
Br
S
O
Br
Br
N
O
O
O
O
N
BOC
N
H
N
H
N
H
O
5l
5m
5n
5s
5o
5t
5p
6
Ph
F
F
Cl
Br
N
Br
Br
Br
O
O
O
O
O
N
N
N
H
N
H
N
H
H
5q
5r
Scheme 1. Reagents and conditions. For Ar–Br: (i) Pd2(dba)3, P(o-tol)3, Et3N, DMF, 80 ꢁC, 6 h. For Ar-Cl: (ii) 2-dicyclohexylphosphino-20-(N,N-
dimethylamino)biphenyl(Cy-MAP), Pd2(dba)3, Et3N, DMF, 80 ꢁC, overnight; (iii) CF3CO2H.

3426
G.-D. Zhu et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3424–3429
Table 1. Enzyme and cellular assay results for compounds 1 and 9a–9t
H
N
BOC
O
NH
H
NH
a
7
Compound
R
Akt1 IC50
(nM)
MTT (F5.12-Akt)
a
N
O
O
IC50 (lM)
i
Br
N
1
2.0
5.4
0.42
0.39
ndb
8.7
5e
10
ii
N
N
RCHO
RCHO
iii
9a
9b
9c
9d
S
S
H
N
H
N
NH
NH2
488
O
7
O
N
O
O
iv
Br
R
N
N
R
N
41
12
11
H
N
Scheme 2. Reagents and conditions: (i) Pd2(dba)3, P(o-tol)3, Et3N,
DMF, 80 ꢁC, 6 h, 47%; (ii) a—piperidine, MeOH, 60 ꢁC, overnight;
b—CF3CO2OH, CH2Cl2; (iii) piperidine, MeOH, 60 ꢁC, overnight;
(iv) a—Pd2(dba)3, P(o-tol)3, Et3N, DMF, 80 ꢁC, 6 h; b—CF3CO2OH,
CH2Cl2.
4400
nd
O
O
9e
9f
3.2
99
0.94
40
N
H
isoquinoline pharmacophore in compound 1. Replace-
ment of the isoquinoline with the isosteric thieno[2,3-
c]pyridine, known to have good pharmacokinetic
properties, resulted in a compound (9a) with slightly
reduced potency in the enzyme assay (IC50 = 5.4 nM)
O
O
N
H
N
O
9g
9h
179
nd
30
and with comparable activity in
a MTT assay
N
H
(IC50 = 0.39 lM vs 0.35 lM for 1).15 Thieno[3,2-c]pyri-
dine isostere (9b), on the other hand, was much less ac-
tive, with an IC50 of 488 nM. Substitution with
benzofurazan (9c), while the oxygen atom could mimic
the hinge-binding nitrogen of the isoquinoline in 1, led
to more than an order of magnitude drop in potency
against Akt1. Replacement of the isoquinoline nitrogen
with a carbonyl group led to oxindole analogs 9d
and 9e. While 9d was essentially inactive, 9e showed a
respectable IC50 of 3.1 nM against Akt1. We postulated
that incorporation of a nitrogen atom into the phenyl
ring or introduction of another carbonyl at 3-position
of the oxindole would increase the basicity of the 2-oxo
group, therefore making the 2-carbonyl a better hydro-
gen bond acceptor. Unfortunately, both modifications
(9f, 9g) led to a reduction in potency. An oxime analog
(9h), however, showed a respectable IC50 against Akt
(IC50 = 6.5 nM).
OH
N
6.5
O
O
N
H
H
N
9i
9j
147
405
40
N
H
CN
N
nd
O
O
N
H
H
N
9k
20,700
nd
N
We also replaced the metabolically labile C-3 methy-
lene of the oxindole scaffold of 9e with nitrogen atom,
forming a potentially more stable cyclic urea structure.
Again, this modification resulted in compounds (9i, 9j)
with much diminished activity against Akt. An oxygen
or sulfur substitution for this methylene, however, was
found to provide compounds (9l, 9m) with single digit
nanomolar IC50 against Akt1. A sulfanimide analog
(9n) was nearly inactive (Akt1 IC50 = 15.6 lM).
Alkylation of the C-3 oxindole methylene with methyl,
ethyl, or benzyl groups led to compounds 9o–9r, with
monomethyl analog 9p the most potent (IC50 = 4 nM).
The geminal difluoro-analog 9t was 2-fold more
potent than compound 9e against Akt1, but,
unexpectedly, was much less cytotoxic in a MTT assay
(IC50 = 11.8 lM).
CN
O
O
O
9l
7.4
4.3
0.62
1.3
N
H
S
9m
N
H
H
N
O
O
9n
9o
15,600
nd
S
N
H
O
24.6
2.7
N
H

G.-D. Zhu et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3424–3429
3427
Table 1 (continued)
Table 2. Geometry ratios of the exo-oxindole double bond, enzyme
and cellular assay results for compounds 11a–11p
a
Compound
R
Akt1 IC50
(nM)
MTT (F5.12-Akt)
a
IC50 (lM)
a
Z/E ratio Akt1 IC50 MTT
Compound
R
(nM)
(F5.12-Akt)
a
IC50 (lM)
9p
O
4.0
1.5
11a
11b
11c
1:1
1:1
1:1
3.3
5.9
2.0
2.2
ndb
N
H
Cl
9q
9r
80
nd
O
O
352
N
H
Cl
Ph
11d
11e
11f
11g
1:3
10.4
10.3
1.5
4.8
1.3
0.2
nd
N
1020
nd
N
H
>98:1
>98:1
>98:1
N
9s
9t
52.7
1.5
1.88
N
H
O
O
N
H
660
F
N
H
F
11.8
N
H
11h
11i
2:3
1:1
8.2
1.2
1.20
0.45
N
a Values are means of two or more experiments. All compounds were
tested under 5 lM ATP.
HN
b Not determined.
11j
1:9
153
nd
SAR studies of 9e analogs with olefinic substitutions at
the C-3 position of the oxindole are summarized in Ta-
ble 2. In general, mono-substituted exomethylene ana-
logs displayed better binding affinity to Akt than the
bis-substituted analogs (e.g., 9s in Table 1,
IC50 = 52.7 nM). Relatively smaller groups tended to
give the best activity. The more sterically hindered
dichlorophenyl analog 11c resulted in a 50-fold drop
in Akt activity. Incorporation of a nitrogen at the
meta-(11d) or ortho-(11e) position of the phenyl ring
was also detrimental.
HN
N
11k
11l
>98:1
>98:1
3:1
7.2
40.3
0.9
0.25
3.2
N
H
N
H
11m
11n
11o
11p
0.83
0.4
S
1:8
0.17
Installation of a 2-pyrrole at the exo-methylene provided
compound 11f with a respectable IC50 of 1.5 nM against
Akt1. N-methylation on the pyrrole ring afforded com-
pound 11h with slightly reduced activity, while saturation
of the ring led to a much less active 11g (IC50 = 660 nM).
The 3-pyrrole analog 11i was equally potent as its C-2 reg-
ioisomer. An indole analog 11j was found to be two or-
ders of magnitude less active than 11i. Multiple alkyl
substitutions on the pyrrol-2-yl group of 11f were also
detrimental (11l). 3Z-Imidazol-2-yl (11k) showed a 5-fold
reduction in potency relative to 3Z-pyrrol-2-yl analog
11f. Unexpectedly, a thiophene analog 11m was 5-fold
more potent than its isostere phenyl analog 11b. As
shown in Table 2, 2-furyl analog 11n demonstrated supe-
rior potency over all other analogs investigated so far,
with an IC50 of 0.17 nM. Since the IC50 was lower than
the enzyme concentration (0.5 nM), the given number
was determined by interpolation (curve-fitting). Reduc-
tion in activity was observed with mono- or multiple alkyl
substitution of the furan ring (11o and 11p).
O
1:9
0.7
0.4
O
1:9
10
0.97
O
a Values are means of two or more experiments. All compounds were
tested under 5 lM ATP.
b Not determined.
Because of potential protein-induced cis/trans isomeriza-
tion of the ligand, it is hard to draw any conclusions
from aforementioned SAR. They can be the substituting
effects on the exo-double bond or just a reflection of a
Z/E ratio of these substituents in the ligand/protein
complexes. Several attempts to co-crystallize com-
pounds in this series with purified Akt, Akt-mutated
PKA and/or PKA protein failed to provide a single

3428
G.-D. Zhu et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3424–3429
good selectivity against most of the protein kinases we
tested and were more selective than the isoquinoline-
based benchmark inhibitor 1 versus, for example,
PKA (110- and 280-fold vs 1.6-fold) and CDK1 (790-
and 3700-fold vs 77-fold). On the other hand, the 2-pyr-
role analog 11f and 2-imidazole analog 11k were less
selective in general against most of the kinases we tested.
The much wider spectrum of kinase activity of 11f and
11k is probably responsible for the higher cytotoxicity
of both compounds (MTT IC50 = 0.2 lM for 11f and
0.25 lM for 11k, Table 4).
E 85
C 87
2.9 A
2.7 A
3.1 A
One of the major goals in exploring an alternative phar-
macophore for the isoquinoline in compound 1 was to
improve the pharmacokinetic properties. However, a
mouse PK screening of the oxindole series of Akt inhib-
itors revealed a similar profile as isoquinoline compound
1. There was no oral drug plasma levels observed with
these compounds in mice, with <10% oral bioavailabili-
ty observed for compound 11f in rat.
K 38
3.4 A
Due to the lack of oral bioavailability, compounds 9e,
11f, and 11n were administered subcutaneously in sepa-
rate MiaPaCa-2 mouse xenograft models at the maxi-
mum tolerable doses. As highlighted in Table 5, the
most potent Akt inhibitor 11n showed modest efficacy,
while the less potent inhibitor 11f demonstrated signifi-
cantly better efficacy in slowing tumor growth. Howev-
er, more toxicity, including lethargy, weight loss, and
skin irritation at the site of injection, were repeatedly ob-
served after administration of 11f. Compound 9e
showed similar efficacy as 11n in the MiaPaCa model
at 5-fold higher doses (75 mg/kg). Both 9e and 11n also
displayed a similar spectrum of toxicity as 11f but to a
significantly less extent.
Figure 2. X-ray structure of 11f in Chk-1 kinase. The pyrrol-2-yl group
in 11f orients in cis-configuration to the carbonyl of the oxindole
scaffold in the ligand/protein complex.16
crystal with sufficient quality for X-ray study. However,
an X-ray structure of 11f in a less related Chk-1 kinase
demonstrated an exclusive cis-configuration of the pyr-
role to the carbonyl of the oxindole in the ligand/protein
complex. The pyrrole nitrogen, the oxindole nitrogen,
and carbonyl formed three hydrogen bonds with the
backbone peptides in the hinge region (see Fig. 2).
In an effort to elucidate the superior efficacy of com-
pound 11f in the MiaPaCa models over 9e and 11n,
To further evaluate this series of Akt inhibitors and
understand the lack of correlation between the enzyme
activity and cytotoxicity, several representative com-
pounds were screened for kinase selectivity. As summa-
rized in Table 3, oxindoles 9e and 11n demonstrated
Table 4. Cellular activity of selected Akt inhibitors in comparison with
1 (IC50, lM)a
Cell
1
9e
11f
11k
11n
GSK3-P
FL5.12-Akt (MTT)
MiaPaCa-2 (MTT)
1.5
1.94
0.94
1.71
0.9
5.0
0.25
0.085
2.5
0.42
0.59
0.20
0.14
0.40
0.88
Table 3. Fold-selectivity of Akt inhibitors for Akt1 over selected
kinasesa
a Values are means of two or more experiments.
Kinase
1
9e
11f
11k
11n
Akt1
Akt2
Akt3
PKA
PKCc
PKCn
CDK1
ERK2
CK2
1
1
1
1
1
130
5
27
8
110
5
43
ndb
19
1
100
280
Table 5. Summary of the efficacies in MiaPaCa-2 mouse xenograft
models17 for selected Akt inhibitors in comparison to Gemzara
1.6
210
9200
77
110
570
2
40
2
1500
>50,000
3700
1700
>50,000
440
Compound Dose
(mg/kg/day)
Route Dose schedule
T/C at #day
>16,000
790
300
12,000
4
5700
0.4
nd
56
17
700
9100
>1000
220
1900
15
9e
11f
75
sc
sc
sc
ip
Bid; d1–21
Bid; d1–7
Bid; d1–17
60% at d28
13% at d15
69% at d24
>16,000
1200
25
15
SRC
11n
Gemzar
a The fold-selectivity for Akt1 over selected protein kinases is shown
for each selected Akt inhibitor. The listed kinases are those relatively
less selective for other series of Akt inhibitors and belong to AGC
(PKA, PKC), TK (SRC), and CMGC (CDK1, ERK2, CK2) family
of protein kinases.
120
Qd; d3, 6, 9, 12 34% at d21
a All Akt inhibitors were subcutaneously dosed twice daily as indicated
schedules after inoculation. Tumor size was measured at specified
days with digital calipers. The data were from different trials with a
similar efficacy for Gemzar as control. All animals were humanely
euthanized on indicated days due to development of toxicity.
b Not determined.

G.-D. Zhu et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3424–3429
3429
Duggan, M.; Lindsley, C. Bioorg. Med. Chem. Lett. 2005,
15, 905.
phosphorylation of the Akt downstream target GSK3
was measured in the presence of inhibitor. The results
are summarized in Table 4 together with their cytotoxic-
ity in MiaPaCa-2 cells. The higher cellular activity of 11f
for GSK3 phosphorylation (0.9 vs 1.94 lM and 2.5 lM
for 9e and 11n) and cell-killing in MiaPaCa correlated
well with their in vivo efficacy.
7. (a) Luo, Y.; Shomaker, A.; Liu, X.; Woods, K.; Thomas,
S.; de Jong, R.; Han, E.; Li, T.; Stoll, V.; Powlas, J.;
Oleksijew, A.; Mitten, M.; Shi, S.; Guan, R.; McGonigal,
T.; Klinghofer, V.; Johnson, E.; Leverson, J.; Bouska, J.;
Mamo, M.; Smith, R.; Gramling-Evans, E.; Zinker, B.;
Mika, A.; Nguyen, P.; Oltersdorf, T.; Rosenberg, S.; Li,
Q.; Giranda, V. Mol. Cancer Ther. 2005, 4, 977; (b) Li, Q.;
Woods, K.; Thomas, S.; Zhu, G. –D.; Packard, G.; Fisher,
J.; Li, T.; Gong, J.; Dinges, J.; Song, X.; Abrams, J.; Luo,
Y.; Johnson, E.; Shi, S.; Liu, X.; Klinghofer, V.; de Jong,
R.; Oltersdorf, T.; Stoll, V.; Jakob, C.; Rosenberg, S.;
Giranda, V. Bioorg. Med. Chem. Lett. 2006, 16, 2000.
8. Zhu, G.-D.; Gong, J.; Claibone, A.; Woods, K.; Gandhi,
V.; Thomas, S.; Luo, Y.; Liu, X.; Shi, S.; Guan, R.;
Magnone, S.; Klinghofer, V.; Johnson, E.; Bouska, J.;
Shoemaker, A.; Oleksijew, A.; Stoll, V.; de Jong, R.;
Oltersdorf, T.; Li, Q.; Rosenberg, S.; Giranda, V. Bioorg.
Med. Chem. Lett. 2006, in press.
In summary, we have developed a series of potent and
selective oxindole–pyridine-based Akt inhibitors. The
most potent compound 11n had an IC50 of 0.17 nM
against Akt1 and was over 100-fold selective against
other Akt isozymes. Compounds 9e and 11n were
among the most selective Akt inhibitors in this series
versus the closely related protein kinase A. Compounds
11f and 11k inhibited a much broader spectrum of pro-
tein kinases and showed higher cytotoxicity. In mouse
MiaPaCa xenograft models, compounds 9e, 11n, and
11f showed marginal to modest efficacy in inhibiting tu-
mor growth but were accompanied by toxicity. The
broader spectrum of activity of 11f versus other kinases
is likely responsible for its higher cytotoxicity, as well as
an improved in vivo efficacy.
9. Wikel, J.; Denney, M.; Vasileff, R. J. Heterocycl. Chem.
1993, 30, 289.
10. Mazeas, D.; Guillaumet, G.; Viaud, M.-C. Heterocycles
1999, 50, 1065.
11. Takeda, K.; Ogura, H. Synth. Commun. 1982, 12, 213.
12. Makosza, M.; Sienkiewicz, K. J. Org. Chem 1998, 63,
4199.
13. Carson, J. US Patent 3,177,221, 1965.
Acknowledgments
14. Sun, L.; Tran, N.; Tang, F.; App, H.; Hirth, P.;
McMahon, G.; Tang, C. J. Med. Chem. 1998, 41,
2588.
We thank Drs. Thomas Penning and Milan Bruncko for
proofreading this manuscript and valuable suggestions.
15. Zhu, G.-D.; Arendsen, D.; Gunawardana, I.; Boyd, S.;
Stewart, A.; Fry, D.; Cool, B.; Kifle, L.; Schaefer, V.;
Meuth, J.; Marsh, K.; Kempf-Grote, A.; Kilgannon, P.;
Gallatin, M.; Okasinski, G. J. Med. Chem. 2001, 44, 3469.
16. Compound 11f was soaked into the native Chk-1 kinase
crystals that were obtained in a solution containing 20 mg/
mL of the protein and PKI peptide. The structure of the
ligand/protein complex was determined and refined to a
References and notes
1. For reviews, see: (a) Barnett, S.; Bilodeau, M.; Lindsley,
C. Curr. Topics Med. Chem. 2005, 5, 109; (b) Li, Q.;
Zhu, G.-D. Curr. Topics Med. Chem. 2002, 2, 939.
2. For a review, see: Nicholson, K. M.; Anderson, N. G.
Cell. Signal. 2002, 14, 381.
3. For reviews, see: (a) Vivanco, I.; Sawyers, C. Nat. Rev.
2002, 2, 489; (b) Gills, J.; Dennis, P. Expert Opin. Investig.
Drugs 2004, 13, 787.
4. Staal, S. Proc. Natl. Acad. Sci. U.S.A. 1987, 84, 5034.
5. (a) Lindsley, C.; Zhao, Z.; Leister; William, H.; Robinson,
R.; Barnett, S.; Defeo-Jones, D.; Jones, R.; Hartman, G.;
Huff, J.; Huber, H.; Duggan, M. Bioorg. Med. Chem. Lett.
2005, 15, 761; (b) Barnett, S.; Defeo-Jones, D.; Fu, S.;
Hancock, P.; Haskell, K.; Jones, R.; Kahana, A.; Kral, A.;
Leander, K.; Lee, L.; Malinowski, J.; McAvoy, E.; Nahas,
D.; Robinson, R.; Huber, H. Biochem. J. 2005, 385(Pt. 2),
399.
˚
resolution of 3.5 A (Rwork = 22.1% and Rfree = 31.9%).
Crystallographic data described in this paper have been
deposited with PDB (ID: 2GHG).
17. Animal studies were conducted following the guidelines of
the internal Institutional Animal Care and Use Commit-
tee. Immunocompromised male scid mice (C.B-17-
Prkdcscid) were randomly assigned to treatment groups
and therapy was initiated the day after inoculation. Ten
animals were assigned to each group, including controls.
MiaPaCa-2 cells were obtained from the American Type
Culture Collection (Manassas, VA). 2 · 106 MiaPaCa-2
cells in 50% Matrigel (BD Biosciences, Bedford, MA) were
inoculated subcutaneously into the flank. Tumor size was
evaluated by twice weekly measurements with digital
calipers. Tumor volume was estimated using the formula:
V = L · W2/2.
6. Zhao, Z.; Leister, W.; Robinson, R.; Barnett, S.; Defeo-
Jones, D.; Jones, R.; Hartman, G.; Huff, J.; Huber, H.;