Azole-based inhibitors of AKT/PKB for the treatment of cancer

Article abstract of DOI:10.1016/j.bmcl.2010.01.067

Through a combination of screening and structure-based rational design, we have discovered a series of N¹-(5-(heterocyclyl)-thiazol-2-yl)-3-(4-trifluoromethylphenyl)-1,2-propanediamines that were developed into potent ATP competitive inhibitors of AKT. Studies of linker strand-binding adenine isosteres identified SAR trends in potency and selectivity that were consistent with binding interactions observed in structures of the inhibitors bound to AKT1 and to the counter-screening target PKA. One compound was shown to have acceptable pharmacokinetic properties and to be a potent inhibitor of AKT signaling and of in vivo xenograft tumor growth in a preclinical model of glioblastoma.

Full text of DOI:10.1016/j.bmcl.2010.01.067

Bioorganic & Medicinal Chemistry Letters 20 (2010) 1559–1564
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Azole-based inhibitors of AKT/PKB for the treatment of cancer
a
a
a
a
Qingping Zeng a, John G. Allen a, , Matthew P. Bourbeau , Xianghong Wang , Guomin Yao , Seifu Tadesse ,
James T. Rider a, Chester C. Yuan a, Fang-Tsao Hong a, Matthew R. Lee a, Shiwen Zhang b, Julie A. Lofgren b,
Daniel J. Freeman b, Suijin Yang b, Chun Li c, Elizabeth Tominey c, Xin Huang d, Douglas Hoffman e,
Harvey K. Yamane f, Christopher Fotsch a, Celia Dominguez a, Randall Hungate a, Xiaoling Zhang b
*
a Chemistry Research and Discovery, Amgen Inc., One Amgen Center Drive, Thousand Oaks, CA 91320, United States
b Oncology Research, Amgen Inc., One Amgen Center Drive, Thousand Oaks, CA 91320, United States
c Pharmacokinetics and Drug Metabolism, Amgen Inc., One Amgen Center Drive, Thousand Oaks, CA 91320, United States
d Chemistry Research and Discovery, Amgen, Inc., 360 Binney St., Cambridge, MA 02142, United States
e Small Molecule Process and Product Development, Amgen Inc., One Amgen Center Drive, Thousand Oaks, CA 91320, United States
f Protein Science, Amgen Inc., One Amgen Center Drive, Thousand Oaks, CA 91320, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Through a combination of screening and structure-based rational design, we have discovered a series of
N1-(5-(heterocyclyl)-thiazol-2-yl)-3-(4-trifluoromethylphenyl)-1,2-propanediamines that were devel-
oped into potent ATP competitive inhibitors of AKT. Studies of linker strand-binding adenine isosteres
identified SAR trends in potency and selectivity that were consistent with binding interactions observed
in structures of the inhibitors bound to AKT1 and to the counter-screening target PKA. One compound
was shown to have acceptable pharmacokinetic properties and to be a potent inhibitor of AKT signaling
and of in vivo xenograft tumor growth in a preclinical model of glioblastoma.
Received 19 December 2009
Revised 9 January 2010
Accepted 13 January 2010
Available online 21 January 2010
Keywords:
PKB
AKT
Ó 2010 Elsevier Ltd. All rights reserved.
Cancer
SAR studies
A serine/threonine kinase with three isoforms, AKT (protein ki-
nase B, PKB) is a key enzyme in the phosphatidylinositol-3-kinase
(PI3K) pathway, and is involved in signaling by multiple growth
factors. Gain-of-function mutations of PI3K1,2 and AKT,3 loss-of-
function mutations of the tumor suppressor phosphatase and ten-
sin homolog (PTEN),4 and amplification of human epidermal
growth factor receptor 2 (HER2)5 have been shown to confer con-
stitutive AKT activity. Genetic events leading to aberrant AKT activ-
ity represent one of the most common molecular mechanisms
associated with human cancers.6 These data taken together sup-
port the hypothesis that AKT inhibitors may be useful for the treat-
ment of cancer.
Glu228-Ala230 (the linker strand connecting the C- and N-terminal
domains of AKT1) were explored. Potency was evaluated by bio-
chemical inhibition of AKT1 and inhibition of phosphorylation of
proline-rich AKT substrate 40 (PRAS40) in U-87 MG cells.9,10 Selec-
tivity screening against protein kinase A (PKA), a close structural
homolog of AKT, was expected to lead to broader kinase selectivity.
In parallel, screening for selectivity against cyclin-dependent
kinase 2 (CDK2), a kinase essential for cell proliferation, was
expected to result in a reduction of off-target effects in in vivo
tumor xenograft studies.
Azole ring analogs were prepared by the convergent routes out-
lined in Scheme 1. The 4-trifluoromethylphenyl sidechain was
Several research groups7 have developed AKT inhibitors includ-
ing GSK2141795 entering phase I clinical trials, and allosteric
inhibitor MK-2206 that has completed phase I trials.8 In the pre-
ceding Letter9 we discussed the discovery and optimization of a
series of thiadiazole-based AKT inhibitors (i.e., 1, Fig. 1). Herein
we describe SAR studies of the azole ring that lead to the identifi-
cation of the thiazole series (2). By varying the substituent at the 5
position of the thiazole in 2 (Ar, linker binder), interactions with
N
N
N
Ar
NH
NH
S
S
N
NH2
NH2
F3C
F3C
2
1
* Corresponding author. Tel.: +1 805 313 5257.
E-mail address: johallen@amgen.com (J.G. Allen).
Figure 1. Thiadiazole-based AKT inhibitor 1 and thiazole-based AKT inhibitor of
type 2.
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.01.067

1560
Q. Zeng et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1559–1564
X
Y
X
Y
c, d
Boc
X = CH
Y = N
NH
Br
N
X
Y
S
S
X
Y
a
b
Br
NHBoc
Br
CO2H
N
S
X = N
Boc
S
NH2
NHBoc
OR
e, d
Y = CH
X = CH, Y = N; 4
N
O
3
X = N, Y = CH; 5
S
X = N
Y = CH
F3C
F3C
O
O
CF3
X = CH, Y = N; 9
X = CH, Y = N; 7
6
X = N, Y = CH; 10
X = N, Y = CH; 8
N
O
OH
N
N
O
N O
NH
CHO
f, g
I
NHBoc b, d
h, i
j
Cl
N
N
NH2
N
N
N
11
12
13
BocHN
14
F3C
CF3
15
N
O
k
16
NHBoc
N
O
N
NH2
l, d
N
F3C
CF3
18
17
O
N
O
NHBoc
HO
N
OH
N
Br
N
m
NH2
21
NH2
N
CF3
N
N
F3C
n, d
19
20
22
N
N
NH
N
O
O
CO2H
p
NH2 b, d
o
O
N
NH2
N
NH2
N
H
N
N
23
25
F3C
24
26
Scheme 1. Reagents and conditions: (a) diphenylphosphorylazide, Et3N, t-BuOH, 80 °C, 66%; (b) tert-butyl (4S)-4-(4-(trifluoromethyl)benzyl)-1,2,3-oxathiazolidine-3-
carboxylate 2,2-dioxide (6), Cs2CO3, DMF, 50 °C, 20–92%; (c) isoquinolin-6-ylboronic acid, tetrakis(triphenylphosphine)palladium, Na2CO3, dioxane, water, 90 °C, 74%; (d) TFA,
CH2Cl2, rt, 63–93%; (e) isoquinolin-6-ylboronic acid, bis(di-t-butylphenylphosphine)palladium dichloride, KOAc, acetonitrile, water, 100 °C, 4%; (f) hydroxylamine HCl, EtOH,
water, 50% NaOH, 0 °C, 84%; (g) NCS, DMF, 0–50 °C, 91%; (h) tributyl(ethynyl)stannane, Et3N, THF, 0 °C to rt, 47%; (i) iodine, THF, 80 °C, 78%; (j) t-butylcarbamate, X-Phos,
Pd2dba3, Cs2CO3, dioxane, 120 °C, 41%; (k) (S)-tert-butyl 1-(4-(trifluoromethyl)phenyl)but-3-yn-2-ylcarbamate (16), bis(triphenylphosphine)palladium dichloride,
copper(I)iodide, Et3N, 65 °C, 51%; (l) Pd–C, H2, MeOH, rt, 89%; (m) (i) CuCN, NMP, 150–170 °C; (ii) hydroxylamine HCl, MeOH, Na2CO3, rt to 50 °C, 9%; (n) (R)-4-(tert-
butoxycarbonyl)-5-(4-(trifluoromethyl)phenyl)pentanoic acid (21), CDI, DMF, 100 °C, 31%; (o) CDI, hydrazine, rt, 69%; (p) (i) BrCN, imidazole, CH2Cl2, reflux; (ii) hydrazine,
THF, reflux, 96%.
introduced as an amino acid derived building block by alkylation
bis(di-t-butylphenylphosphine)palladium catalyst was particularly
useful in the preparation of carbonyl-containing linker binders
(36–43).11 Stille couplings were used in cases where arylboronates
were difficult to form (28 and 29). Compounds containing 3-amino-
indazoles (32 and 35) and a 3-aminobenzisoxazole (33) could be
prepared from the corresponding common late stage fluorobenzoni-
trile intermediates (48 and 49) by treatment with methylhydrazine,
hydrazine or N-hydroxyacetamide in the presence of potassium
tert-butoxide.
We initially explored the SAR around the core azole ring while
keeping the isoquinoline linker binder element of the potent thia-
diazole analog 19 constant (Table 1). As discussed in the previous
Letter,9 the thiadiazole series of compounds bound in a U-shaped
conformation and formed a hydrogen bond with the sidechain of
the catalytic Lys179 residue, and the 2-amino group formed an io-
nic interaction with Asp292 (AKT1 numbering as shown in Figure
2A). A similar Lys179 contact was made by thiazole 9, but did
not appear to be possible for isomeric thiazole 10, explaining the
observed loss of potency. Compounds 15, 18, and 22 also appeared
with 1,2,3-oxathiazolidine-3-carboxylate 2,2-dioxide 6,11 Sono-
gashira coupling with alkyne 1612 or condensation with
c-amino-
acid 21.13 Compounds 9 and 10 were prepared from intermediates
4 and 5 by sequential sidechain alkylation with 6, Suzuki cou-
pling14 with isoquinoline boronic acid and finally deprotection
with TFA. Isoxazole analog 15 was prepared from 14 and 6. Isoxaz-
ole 18, containing a 5-position methylene, was prepared from 13
by Sonogashira coupling with 16 followed by hydrogenation. The
1,2,4-oxadiazole ring of analog 22, was assembled by condensing
21 with hydroxyacetimidamide 20. The 1,3,4-oxadiazole ring in
26 was prepared by reacting isoquinoline-6-carbohydrazide 24
with di(1H-imidazol-1-yl)methanimine prepared in situ.
Thiazole compounds containing alternative linker binding ele-
ments were prepared as shown in Scheme 2. Suzuki couplings of
7 with aryl boronic acids or aryl pinacolboronates, prepared by
lithium halogen exchange and quenching with triisopropylboro-
nate or by Miyaura coupling of aryl bromides or chlorides with
pinacol diborane,15 provided a range of arylated thiazoles 2. The

Q. Zeng et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1559–1564
1561
N
N
HN
N
Boc
N
N
Br
N
Ar
N
N
NH
NH
30
N
S
S
a, b
H2N
27
39
34*
9
31
36
NHBoc
NH2
F3C
F3C
HN
HN
O
HN
HN
O
HN
O
7
2, Ar =
O
S
N
O
O
38
40
37
O
H2N
N
HN
N
O
O
H
43
41
42
N
N
N
N
c
d, b
2, Ar =
N
Ar
Bu3Sn
NHBoc
NHBoc
N
S
S
44
45
28
29
N
O
N
H2N
O
B
e
f
NH
S
32: 2, Ar =
33: 2, Ar =
35: 2, Ar =
48
R1
49
O
N
a, b
N
NH2
7
+
R2
F3C
1 = CN, R2 = F; 48
NH2
R1
R2
1 = CN, R2 = F; 46
HN
N
g
49
R
R
R1 = F, R2 = CN; 47
R1 = F, R2 = CN; 49
NH2
Scheme 2. Reagents and conditions: (a) arylboronic acid or arylpinacolboronate, bis(di-t-butylphenylphosphine)palladium dichloride or bis(triphenylphosphine)dichloride
or tetrakis(triphenylphosphine)palladium, Na2CO3 or KOAc, dioxane or acetonitrile or dimethoxyethane, water or water–ethanol, 85–150 °C, 8–89%; (b) TFA, or TFA–CH2Cl2,
rt, 38–89%; (c) arylbromide, tetrakis(triphenylphosphine)palladium, CsF–CuI or LiCl, DMF, 80–100 °C, 46–73%; (d) tert-butyl (4S)-4-(4-(trifluoromethyl)benzyl)-1,2,3-
oxathiazolidine-3-carboxylate 2,2-dioxide, Cs2CO3, DMF, 50 °C, 33–96%; (e) methylhydrazine, 100 °C, 69%; (f) N-hydroxyacetamide, KOtBu, DMF, 20%; (g) hydrazine, 100 °C,
38–62%. *Compound 34 was prepared with the corresponding 4-chlorophenyl side chain instead of the 4-trifluoromethylphenyl side chain.
Table 1
AKT, PKA and CDK2 enzyme inhibition and inhibition of PRAS40 phosphorylation (Thr246) in U-87 MG cells for azole core analogsa
Compd
AKT1 IC50 (nM)
PKA IC50 (nM)
CDK2 IC50 (nM)
pPRAS40 IC50 (lM)
19
9
10
15
18
22
26
3.2 0.9
4.9 3.6
404 31
17
40 18
1600 230
85 12
7.6 15
5.9 4.2
187 23
40
50 8.7
980 74
133 9.1
53 23
12.4 5.7
1070 120
93
117 30
4630
0.25 0.11
0.43 0.27
NT
NT
3.8
NT
NT
348 100
a
Data is given as the mean standard deviation (SD) where three or more measurements were made. NT indicates not tested.
to be able to engage Lys179. However, the isoxazole core of 15 di-
rected a hydrogen atom toward the sidechain phenyl ring and may
have destabilized the U-shaped binding conformation of 15
compared to 1 or 9 (molecular mechanical modeling not shown).
Isoxazole 18 had a methylene at the sidechain rather than a 5-ami-
no N-linkage, and loss of interaction with Asp292 further de-
creased potency. The reduced potency of 1,2,4-oxadiazole analog
22 compared to 18 (ꢀ40-fold) may have been caused by the polar
N-4 atom of the isoxazole forming an unfavorable interaction with
the lipophilic phenyl ring, serving to destabilize the U-shaped
binding conformation. Perhaps for a similar reason oxadiazole 26
was less potent than 1.
In addition to the need for a hydrogen bond acceptor to inter-
act with the backbone NH of Ala230 in AKT1, an additional
hydrogen bond was required between the Glu228 backbone car-
bonyl and an appropriately positioned inhibitor hydrogen. The
potent isoquinoline containing compound 9 provided an aryl
CH:O interaction via isoquinoline H-1 and met these require-
ments. Additional analogs were prepared to examine AKT1 po-
tency and selectivity (Table 2). The 3-aminoisoquinoline analog
27 was designed to form a hydrogen bond with the Ala230 back-
bone carbonyl, however, this added interaction did not appear to
improve potency as 27 was 2–3-fold less potent than 9. One
potential explanation for this loss of potency is the desolvation

1562
Q. Zeng et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1559–1564
Figure 2. (A) Compound 36 modeled into AKT1 (PDB 3CQU6)17 and (B) co-crystallized with PKA (PDB 3L9L).
Table 2
AKT, PKA and CDK2 enzyme inhibition and inhibition of PRAS40 phosphorylation in U-87 MG cells for thiazole-based linker strand binder analogsa
Compd
AKT1IC50 (nM)
PKA IC50 (nM)
CDK2 IC50 (nM)
pPRAS40 IC50 (lM)
9
27
28
29
30
31
32
33
34b
35
36
37
38c
39
40
41
42
43
4.9 3.6
11.6 0.7
35 18
5.9 4.2
8.6 0.8
126 40
489 88
334 88
230 25
119 1.0
417 96
39.9 8.5
86.2 7.4
167 50
326 118
143 71
179
12.4 5.7
15.4 5.4
136 22
>2000
>1000
NT
>2000
>2000
NT
55.3 7.8
185 96
138 25
285 31
276
>25,000
>1000
>10,000
>9000
0.43 0.27
0.89 0.47
1.3 0.14
NT
NT
NT
NT
NT
0.73 0.35
NT
0.30 0.08
0.50 0.35
0.92 0.58
0.63, 0.64c
NT
NT
NT
6.0 1.0
141 4.8
613 91
895 226
2406 111
453 63
8.5 0.9
77 5.2
18.3 6.4
8.0 2.6
27 13
15.4 5.6
3315 284
>1000
>3000
95 16
2404 1014
485
>3000
356 39
a
b
c
Data is given as the mean standard deviation (SD) where three or more measurements were made. NT indicates not tested.
4-Chlorophenyl sidechain instead of 4-trifluoromethylphenyl sidechain.
n = 2.
required for the second hydrogen of the anilinic amine. Similarly,
quinazoline 28 placed its polar N1 atom in the same hydrophobic
environment and was sixfold less potent than 9. Compound 29,
containing the cinnoline linker binder, was nearly 30-fold less po-
tent than 9 possibly due to an electrostatic repulsion between the
Glu228 backbone carbonyl and the lone pair on the N1 of the
cinnoline.
Benzazoles were also tested. The 6 substituted indazoles 30 and
31 had IC50 values >600 nM, and may have formed a less than opti-
mal hydrogen bond with Glu228. Although binding with geometry
similar to isoquinoline, this loss of potency suggested the indazole
H-3 was a less efficient aryl CH:O interaction donor than isoquin-
oline H-1. The incorporation of an amino group at the indazole 3
position was presumably too bulky and caused a further decline
in potency in the case of the 6-yl substituted indazoles (32, IC50
Although trends toward more potent compounds were discov-
ered through azole ring and linker binder group modifications,
selectivity and cellular activity remained a challenge. None of the
azole core analogs in Table 1 nor the linker binder analogs 27–35
showed significant selectivity over PKA, and only 34 showed mod-
est selectivity (PKA/AKT1 = 4.7). The isoquinoline, indazole and
quinazoline analogs 9, 18, 27, 28 and 34, had significant and unfa-
vorable cell shifts with cellular IC50 values that were 37–95-fold
higher than their IC50 values in the enzyme assay.
Apart from aromatic linker strand binders, we also looked at
non-aromatic compounds. It was anticipated that lactam analogs
such as oxindole 36 could complement the hydrogen-bonding pat-
tern of the linker strand. Indeed, 36 was reasonably potent (AKT1
IC50 = 18 nM) and, interestingly, it was 9–10-fold selective over
PKA and CDK2. Benzoxazolone 37, benzothiazolone 38 and 3-
methylbenzoimidazolone 39 were similarly potent and selective.
Dihydroquinolinone 40 was less potent (AKT1 IC50 = 3315 nM)
possibly due to increased repulsion from the Ala230 carbonyl,
and in 41 the required U-shaped binding conformation was prob-
ably disrupted by an unfavorable interaction between the phthala-
zinone carbonyl and the sidechain phenyl ring. Monocyclic analog
>2 lM). Benzisoxazole 33, similar to 29, did not satisfy Glu228.
In contrast, the 5-yl substituted indazole analog 34 was >70-fold
more potent than the corresponding 6-yl substituted indazole 30.
The 3-amino group of indazole 35 was designed to interact with
the Ala230 backbone carbonyl, but the potency was ꢀ9-fold less
than the unsubstituted analog 34 (cf. 27 vs 9).

Q. Zeng et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1559–1564
1563
42 did not have the required conformational constraints and its
IC50 was >3 M. The isoindolin-1-one analog 43 had an
than 9 and achieved higher exposures following oral dosing at
5 mg/kg. Benzoxazole 37 had poor bioavailability and low oral
exposure. It was subsequently confirmed that compounds 36 and
37 were cleared via hepatic oxidation of the thiazole ring.19 Based
on rat PK data, compounds 9 and 36 were selected for in vivo phar-
macodynamic (PD) studies in mice, measuring inhibition of hepa-
tocyte growth factor (HGF) induced PRAS40 phosphorylation in
the liver.20 Six hours after a 30 mg/kg dose, 9 achieved a plasma
concentration 1.7-fold above its cellular IC50, and showed minimal
inhibition of PRAS40 phosphorylation in vivo. In contrast, plasma
concentration of 36 6 h post dose was about 10-fold over its cellu-
lar IC50, and significant inhibition of PRAS40 phosphorylation was
observed (43%).
We next tested compound 36 in an in vivo tumor xenograft
model. Accordingly, 36 was administered for 18 days to nude mice
bearing U-87 MG tumors (15, 30, 60 mg/kg qd po).21 Significant
inhibition of tumor growth (93%) was observed at the 30 mg/kg
dose relative to vehicle treated control. Although measured at dif-
ferent time-points, plasma exposures appeared to be similar in the
efficacy study and the single dose PD experiment. Tumor growth
and body weight data are shown in Figure 3. Tumor growth inhibi-
tion was dose proportional and treatment was tolerated with 2–
10% loss in body weight over the course of the study compared
to a gain of 2% for vehicle. In a screening panel of 43 kinases, com-
pound 36 had IC50 values <500 nM on 15 of these targets, including
kinases that are involved in tumor growth regulation (e.g., IC50 val-
l
IC50 = 94.6 nM, ꢀ5-fold less active than 36. In addition to improved
selectivity, the lactam analogs 36–39, and 43 showed reduced cell
shifts compared to the indazole and isoquinoline analogs (16–63-
fold). For example, although 36 was less potent in the enzyme as-
say compared to 1 (18.3 nM vs 3.2 nM), it had a smaller enzyme to
cell shift and cellular potency of the two compounds was similar
(36, IC50 = 0.30
viability was confirmed for 36 (IC50 = 0.63
l
M; 1, IC50 = 0.25
l
M). Inhibition of U-87 MG cell
l
M).16
Compound 36 was modeled into the active site of AKT117 using
a previously described hierarchical method,18 and was also co-
crystallized in PKA ( Fig. 2). Similar to the previously described
compounds,9 36 bound in a U-shaped conformation, making the
expected key interactions with the linker residue backbone amide
(AKT1, Ala230; PKA, Val123) and linker-2 glutamate carbonyl
(AKT1, Glu228; PKA, Glu121). The improved selectivity of com-
pound 36 relative to 9 may be explained in part by the smaller size
of the linker-strand binder. As discussed in the previous Letter, the
oxindole of 36 may have formed better hydrophobic interactions
with the AKT1 floor residue (Met281) than with the PKA floor res-
idue (Leu173).9 Alternatively, this improved selectivity profile may
also be explained by different linker residues in AKT versus PKA.
Asparagine in AKT (Asn231) would likely allow greater flexibility
of the AKT linker strand compared to PKA which contains a proline
residue (Pro124) in this position. This flexibility may allow the
AKT1 linker strand to relax in order to avoid a repulsive electro-
static interaction between the inhibitor carbonyl and the Ala230
backbone carbonyl, whereas PKA could not. Consequently an unfa-
vorable electrostatic interaction in PKA (3.4 Å vs 3.7 Å) would re-
duce the potency of the lactam-based inhibitors on PKA, thereby
improving selectivity.
ues: KDR, 0.30 lM; c-KIT, 0.11 lM; CDK1, 0.25 lM). Thus at the
high plasma concentrations achieved during the tumor growth
inhibition study, inhibition of other kinases may have contributed
to the efficacy observed.
Flexible and convergent syntheses have generated a series of
thiazole-based inhibitors of AKT. We have optimized compounds
for potency on AKT and have demonstrated modest selectivity
against PKA and CDK2 kinases. Inhibition of the phosphorylation
of one AKT substrate, PRAS40, was demonstrated in U-87 MG cell
culture and in a mouse liver PD model using 36. Inhibition of
Compounds 9, 36 and 37 were advanced into in vivo experi-
ments (Table 3). In rat pharmacokinetic (PK) studies, isoquinoline
compound 9 was found to be cleared at a high rate and had mod-
erate bioavailability. Oxindole compound 36 had lower clearance
Table 3
In vivo rat pharmacokinetic parameters, mouse liver PD effect and U-87 MG xenograft growth inhibition
Compd
Rat pharmacokinetic parameters
Mouse liver PD at 6 h
U-87 MG xenograft
Xenograft plasma concentration
growth inhibitiond (%) 3 h post last dose
a
b
b
t1/2
(h)
CLa
(L/h/kg) (L/kg)
Vssa
% Fb Cmax
AUC(0–24h)
pPRAS40
Plasma concd (lM)
(ng/mL) (ngh/mL)
(Thr246, %)c
9
36
37
2.9
4.3
3.8
2.7
1.0
1.6
7.1
6.8
6.7
20
64
11
66
417
68
387
2600
348
4
43*
NT
1.5
3.5
NT
NT
93**
NT
NT
6.2
NT
lM
a
b
c
Dosing was 2 mg/kg iv in DMSO in n = 3 male Sprague Dawley rats.
Dosing was 5 mg/kg po in 2% HPMC, 1% Tween 80, pH 2 in n = 3 male Sprague Dawley rats.
Dosing was 30 mg/kg po in 0.1% PBS/BSA in n = 3 female mice, *p = 0.0079. NT indicates not tested.
d
Dosing was 18 d 30 mg/kg po in female CD1 nude mice (n = 10), **p <0.0001. NT indicates not tested.
Figure 3. Compound 36 inhibited U-87 MG xenograft growth in CD1 nude mice with daily dosing over 18 d. (A) Tumor growth; (B) body weight normalized to day 1.

1564
Q. Zeng et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1559–1564
12. (a) Hauske, J. R.; Dorff, P.; Julin, S.; Martinelli, G.; Bussolari, J. Tetrahedron Lett.
1992, 33, 3715; (b) Brown, D. G.; Velthuisen, E. J.; Commerford, J. R.; Brisbois, R.
G.; Hoye, T. R. J. Org. Chem. 1996, 61, 2540.
13. Smrcina, M.; Majer, P.; Majerova, E.; Guerassina, T. A.; Eissenstat, M. A.
Tetrahedron 1997, 53, 12867.
14. Guram, A. S.; King, A. O.; Allen, J. G.; Wang, X.; Schenkel, L. B.; Chan, J.; Bunel, E.
E.; Faul, M. M.; Larsen, R. D.; Martinelli, M. J.; Reider, P. J. Org. Lett. 2006, 8,
1787.
PRAS40 phosphorylation in vivo coincided with inhibition of xeno-
graft tumor growth, albeit at high plasma concentrations where
inhibition of off-target kinases involved in cell proliferation or reg-
ulation may have contributed to tumor growth inhibition. As AKT
remains an attractive anti-cancer target, AKT inhibitors such as
the ones presented here may be important for further
development.
15.
Ishiyama, T.; Murata, M.; Miyaura, N. J. Org. Chem. 1995, 60, 7508.
16. U-87 MG cells were seeded on a 96-well cell culture plate at 6000 cells/well in
5% FBS, and treated with 36 in threefold serial dilutions for three days. Cell
viability was measured by alamarBlue cell staining (Invitrogen, DAL1100).
17. Lippa, B.; Pan, G.; Corbett, M.; Li, C.; Kauffman, G. S.; Pandit, J.; Robinson, S.;
Wei, L.; Kozina, E.; Marr, E. S.; Borzillo, G.; Knauth, E.; Barbacci-Tobin, E. G.;
Vincent, P.; Troutman, M.; Baker, D.; Rajamohan, F.; Kakar, S.; Clark, T.; Morris,
J. Bioorg. Med. Chem. Lett. 2008, 18, 3359.
References and notes
1. Hennessy, B. T.; Smith, D. L.; Ram, P. T.; Lu, Y.; Mills, G. B. Nat. Rev. Drug
Discovery 2005, 4, 988.
2. Bader, A. G.; Khang, S.; Zhao, L.; Vogt, P. K. Nat. Rev. Cancer 2005, 5, 921.
3. Testa, J. R.; Bellacosa, A. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 10983.
4. (a) Cully, M.; You, H.; Levine, A. J.; Mak, T. W. Nat. Rev. Cancer 2006, 6, 184; (b)
Leslie, N. R.; Downes, C. P. Biochem. J. 2004, 382, 1.
5. Tokunaga, E.; Kimura, Y.; Oki, E.; Ueda, N.; Futatsugi, M.; Mashino, K.;
Yamamoto, M.; Ikebe, M.; Kakeji, Y.; Baba, H.; Maehara, Y. Int. J. Cancer 2006,
118, 284.
6. Altomare, D. A.; Testa, J. R. Oncogene 2005, 24, 7455.
7. Lindsley, C. W.; Barnett, S. F.; Layton, M. E.; Bilodeau, M. T. Curr. Cancer Drug
Target 2008, 8, 7.
8. (a) ClinicalTrials.gov identifier NCT00920257.; (b) Tolcher, A. W.; Yap, T. A.;
Fearen, I.; Taylor, A.; Carpenter, C.; Brunetto, A. T.; Beeram, M.; Papadopoulos,
K.; Yan, L.; de Bono, J. J. Clin. Oncol. 2009, 27, 15s.
9. Zeng, Q.; Bourbeau, M. P.; Wohlhieter, G. E.; Yao, G.; Monenschein, H.; Rider, J.
T.; Lee, M. R.; Zhang, S.; Lofgren, J. A.; Freeman, D. J.; Li, C.; Tominey, E.; Huang,
X.; Hoffman, D.; Yamane, H.; Tasker, A. S.; Dominguez, C.; Viswanadhan, V. N.;
Hungate, R.; Zhang, X. Bioorg. Med. Chem. Lett. 2010, 20, 1652.
10. U-87 MG cells in 5% FBS were incubated with inhibitors in threefold serial
dilutions for 1 h at 37 °C. The cells were lysed and PRAS40 phosphorylation was
quantified by ELISA assay. pPRAS40 was normalized to total PRAS40.
11. Posakony, J. J.; Grierson, J. R.; Tewsonin, T. J. J. Org. Chem. 2002, 67, 5164.
18. Lee, M. R.; Sun, Y. J. Chem. Theor. Comp. 2007, 3, 1106.
19. Subramanian, R.; Lee, M. R.; Allen, J. G.; Bourbeau, M. P.; Fotsch, C.; Hong, F.-T.;
Tadesse, S.; Yao, G.; Yuan, C. C.; Surapaneni, S.; Skiles, G. L.; Wang, X.;
Wohlhieter, G. E.; Zeng, Q.; Zhou, Y.; Zhu, S.; Li, C. Chem. Res. Toxicol., 2010,
doi:10.1021/tx900414g.
20. Mice (n = 3) were dosed with compounds 9 or 36 (30 mg/kg po) and after 6 h,
AKT signaling was stimulated in mouse liver by treatment with HGF via tail
vein injection. Five minutes post-HGF treatment, the mice were sacrificed and
livers were harvested for quantitation of phospho- and total PRAS40 by ELISA.
pPRAS40 was normalized to total PRAS40 to calculate inhibition. Factorial
ANOVA followed by Dunnett’s post hoc test was used to determine statistical
significance. Data for 36 was from one of two experiments.
21. 10-Week-old female CD1 nude mice (Charles River Laboratories, Wilmington,
MA) were injected sc in the flank with 5 Â 10
6 U-87 MG cells in 0.2 mL DMEM.
On day 10, mice were randomized into groups of n = 10 with an initial tumor
volume of ꢀ200 mm3. Diet was supplemented. Tumor measurements and body
weights were recorded twice per week. Tumor volume was calculated as
length  width  height. Tumor measurements were analyzed by RMANOVA
using StatView version 5.0.1 (SAS Institute, Cary, NC). Scheffé’s post hoc test
was used to determine statistical significance. Data for 36 was from one of
three experiments.