Imidazo[4,5-c]quinolines as inhibitors of the PI3K/PKB-pathway

Article abstract of DOI:10.1016/j.bmcl.2007.12.018

Imidazo[4,5-c]quinoline derivatives have been discovered and developed as potent and effective modulators of the phosphatidylinositol 3-kinase (PI3K)/protein kinase B (PKB) pathway to lead to clinical development candidates. The SAR data of representative examples of this compound class and their biological profiling in cellular and in vivo settings are presented and discussed.

Full text of DOI:10.1016/j.bmcl.2007.12.018

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 1027–1030
Imidazo[4,5-c]quinolines as inhibitors of the PI3K/PKB-pathway
*
´ ´
´
´
Frederic Stauffer, Sauveur-Michel Maira, Pascal Furet and Carlos Garcıa-Echeverrıa
Novartis Institute for BioMedical Research, Oncology Drug Discovery, Klybeckstrasse 141, Postfach, CH-4002 Basel, Switzerland
Received 24 October 2007; revised 9 December 2007; accepted 11 December 2007
Available online 15 December 2007
Abstract—Imidazo[4,5-c]quinoline derivatives have been discovered and developed as potent and effective modulators of the phos-
phatidylinositol 3-kinase (PI3K)/protein kinase B (PKB) pathway to lead to clinical development candidates. The SAR data of rep-
resentative examples of this compound class and their biological profiling in cellular and in vivo settings are presented and discussed.
Ó 2007 Elsevier Ltd. All rights reserved.
The PI3K/PKB-pathway is one of the major signaling
NC
R2
N
pathways activated in cancer via deletion, activating
mutation or amplification of one or several of its com-
ponents. The node of the pathway is PKB (also called
Akt), a serine/threonine protein kinase responsible for
the phosphorylation of downstream effectors involved
in important cellular processes such as metabolism,
growth, proliferation, survival, and also angiogenesis.
Activation of PKB requires three sequential events:
recruitment at the membrane by phosphatidylinositol-
3,4,5-trisphosphate (PIP3), the product of the lipid ki-
nase PI3K which is negatively regulated by the tumor
suppressor phosphatase and tensin homolog deleted on
chromosome 10 (PTEN); phosphorylation of Thr-308
(PKBa numbering) by the 3-phosphoinositide-depen-
dent protein kinase-1 (PDK1); and phosphorylation of
Ser-473 (PKBa numbering) by mTORC2, which is a
complex formed by mTor and rictor.1 Given the role
of the PI3K/PKB-pathway in the biology of human can-
cers, some of the components of this signaling cascade
have become the focus of intensive drug discovery activ-
ities, mostly still at preclinical stage.2
R1
R9
N
O
N
R8
R7
3
R
N
N
N
R4
R6
N
a
c
1
O
N
N
N
HN
N
N
R
N
N
N
H2N
N
N
N
N
HN
N
O
Hinge-NH
b
Figure 1. (a) Generic structure of the imidazo[4,5-c]quinoline scaffold;
(b) different potential binding modes mimicking the canonical H-bond
interactions of adenine with the hinge region; (c) NVP-BEZ235 (1).
kinase, in order to achieve the desired activity and selec-
tivity profile. This scaffold has been used and optimized
for the identification of new modulators of the PI3K/
PKB-pathway, leading to the discovery and develop-
ment of NVP-BEZ235 (1) (Fig. 1c). This compound is
a dual PI3K/mTor inhibitor, which is currently undergo-
ing Phase I clinical trials in cancer patients.
The imidazo[4,5-c]quinoline scaffold (Fig. 1a) can be
considered a privileged ATP-site-directed kinase lead
structure for drug discovery activities. This chemotype
can adopt different binding modes in the ATP binding
cleft, to mimic the H-bond interactions of the adenine
moiety of ATP (Fig. 1b). The structural elements of
the compound and its binding mode can be exploited
to reach regions of the active site unique to the targeted
We report herein our initial efforts to identify modula-
tors of the PI3K/PKB-pathway, and the in vitro and
in vivo biological characterization of representative
examples of PDK1 and PI3K kinase inhibitors based
on the imidazo[4,5-c]quinoline scaffold.
Keywords: PDK1; PI3K; Oncology; PI3K/PKB-pathway.
*
Corresponding author. E-mail: frederic.stauffer@novartis.com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.12.018

1028
F. Stauffer et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1027–1030
receptor tyrosine kinase. The class IB PI3K consists of
R R
NC
R
p110c which is principally recruited by GPCR via its
regulatory subunit.4 The most relevant paralog for can-
cer therapy is probably the p110a protein, for which
activating mutations have been identified in tumor clin-
ical samples.5 The physiological role of the class I PI3K
isoforms has been discussed elsewhere.6
NC
+
X
N
R
NO2
Br
a
NH
b
NO2
Br
X = OH 8
X = Cl 9
NH2
R = H 10a
R = Me 10b
11
R
c
R
R R
NC
NC
R'
Compounds 2–7 were efficiently obtained starting from
6-bromo-4-hydroxy-3-nitroquinoline (8) (Scheme 1).
The position 4 of quinoline 8 is activated via chlorina-
tion using phosphoroxychloride to yield 9, followed by
an acid promoted SNAr reaction with aniline 10a or
10b. The nitro-intermediate 11 is subjected to Raney
nickel catalyzed hydrogenation to afford intermediate
12. Closure of the 5-membered ring of the imidazoquin-
oline system is achieved by heating 12 in neat methyl
orthoformate, acetate or propionate with reaction time
and temperature requirements increasing with the
homologation of the ortho ester. The resulting bromo-
intermediates are subjected to a Sonogashira coupling
reaction to afford compounds 2–6.
NH
N
d, e
N
N
NH2
Br
N
N
12
f,g, e
R= R'=
2
3
4
5
6
H
H
H
Me
NC
H
Me
Et
H
O
N
N
N
N
Me Me
7
Scheme 1. Reagents and conditions: (a) POCl3, 6 h at 125 °C; (b)
AcOH, 1 h at rt; (c) H2, cat. Ni/Ra, MeOH/THF, 12 h at rt; (d) methyl
ortho ester, 1–2 h at 110–130 °C; (e) 3-acetyleneylpyridine, cat.
t
Pd(PhCN)2Cl2, Bu3P, CuI, dioxane/NMP, 0.5–3 d at rt; (f) diphos-
gene, NEt3, CH2Cl2, 20 min at 0 °C; (g) MeI, NaOH(aq), cat. TBABr,
CH2Cl2, 18 h at rt.
A slightly modified protocol was used for the synthesis
of compound 7.7 Closure of the imidazolinone ring is
achieved with diphosgene in the presence of triethyl-
amine, followed by tetrabutylammonium bromide phase
transfer catalyzed methylation at position 3. As for com-
pounds 2–6, a Sonogashira coupling reaction following
a modified protocol using the bulky tri-tert-butylphos-
phine as ligand8 provides the target compound. Aniline
10b is obtained from 4-nitrophenylacetonitrile in two
steps (in a similar manner as described for steps g and
c, Scheme 1).
Table 1. Inhibition of PDK1 kinase activity in biochemical assays and
phosphorylation of T308-PKB in cellular settings, and anti-prolifer-
ative activity
Compounds
PDK1
a
IC50 (nM)
p-T308-PKB
b
IC50 (nM)
Proliferation
c
EC50 (nM)
2
3
4
5
6
7
34
110
94
77
416
275
400
446
143
13
327
150
138
139
45
Modification of the lead compound 2 by introducing a
methyl (compound 3) or ethyl (compound 4) group at
position 3 decreased the PDK1 kinase inhibitory activity
by up to one log unit (Table 1). Replacing the potential
metabolic weak spot and hydrogen bond acceptor meth-
ylcyano moiety with a dimethyl methylcyano group
(compound 5) had a small detrimental effect on PDK1
activity as shown in Table 1. Combining the dimethyl
methyl cyano substituent with a methyl group on the
imidazole ring gave compound 6. This derivative has a
reduced PDK1 inhibitory activity, but remains equipo-
tent against PI3K (IC50 = 56, 446, 35, and 117 nM for
p110a, p110b, p110d, and p110c, respectively) as com-
pared to compound 2. A good correlation was observed
between the inactivation of the pathway, as measured by
inhibition of phosphorylation of Thr308 of PKB (p-
T308-PKB), and the anti-proliferative activity of the
inhibitor (Table 1). The functional activity of compound
6 was further assessed in cellular settings.
245
>25,000
33
a Enzymatic activity with ATP concentration at 10 lM.
b In vitro cell capture ELISA in U87MG.
c In vitro antiproliferation on U87MG.
Starting from a previously identified imidazo[4,5-c]quin-
oline based hit, compound 2 (Scheme 1) was shown to
be a potent inhibitor of PDK1 (IC50 = 34 nM; Table
1). PDK1 is a serine/threonine kinase belonging to the
AGC kinase family. In addition to its catalytic domain,
this protein has a C-terminal pleckstrin homology (PH)
domain required for the binding of PIP3. PDK1 is un-
iquely responsible for a PI3K-dependent activating
phosphorylation of PKB, but other kinases of the
AGC family can also be activated by PDK1 in a
PI3K-independent manner.3 Compound 2 has a good
selectivity profile against other protein kinases, includ-
ing other members of AGC kinase family (e.g., PKA
and PKB, 24% and 0% inhibition at 10 lM, respec-
tively), but 2 is active against class I PI3K (IC50 = 64,
432, 98, and 67 nM for p110a, p110b, p110d, and
p110c, respectively). This family of lipid kinases is com-
posed of two subgroups, IA and IB. The class IA PI3K
subgroup consists of three catalytic subunits, p110a,
p110b, and p110d, all of which interact with regulatory
subunits allowing for their recruitment by an activated
Tumor cell lines, in which the PI3K/PKB-pathway is
known to be deregulated via PTEN deletion (U87MG
glioblastoma and PC3M prostate tumor cells), were ex-
posed to increasing concentrations of compound 6 and
cell extracts were monitored by Western blot (Fig. 2).
A dose-dependent decrease of p-PKB was observed in
both tumor cell lines, whereas the total PKB levels
stayed constant, demonstrating that the effect on p-

F. Stauffer et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1027–1030
1029
blocked the activation of the pathway (IC50 = 33 nM,
p-T308-PKB; Table 1). Analysis of the docking of com-
pound 2 in the PDK1 active site shows a short distance
PC3M
U87MG
Compound 6 in nM
˚
(3.4 A) between the carbon in position 2 of the imidazole
PKB
ring and the backbone carbonyl of Leu88 of the P-loop
(Fig. 4a). In agreement with the in vitro PDK1 kinase
activity profile, a methyl (compounds 3 and 6) or ethyl
(compound 4) substituent resulted in a slight steric de-
mand, but introduction of a carbonyl group at position
2 (compound 7) created an electrostatic mismatch. Such
a strong repulsive interaction is not present in the PI3K
active site because in this lipid kinase the structural mo-
tif corresponding to the P-loop of protein kinases has a
different spatial arrangement with no residue matching
the position of Leu88 in PDK1. As a consequence, the
carbonyl group of compound 7 is facing the protein–
water interface solvent with Trp812 and Met804—the
closest residues of the loop—being located above this
group (Fig. 4b).
p-T308-PKB
p-S473-PKB
GSK3α/β
p-S9-GSK3β
p-T32-FKHRL1
p-T389-p70S6K
Figure 2. Western blot analysis of U87MG and PC3M cell extract
treated with increasing concentrations of compound 6.
PKB could not be due to protein degradation. Inhibi-
tion of activation of PKB was associated with a reduc-
tion in the levels of phosphorylated direct or indirect
downstream effectors like Ser9-GSK3b, Thr32-
FKHRL1/FOXO3a, and Thr389-p70S6K, showing effi-
cient PI3K/PKB-pathway blockade upon in vitro com-
pound treatment. The in vivo efficacy of compound 6
was assessed using PC3M tumors grown subcutaneously
in athymic mice. The rate of tumor growth was reduced
and tumor volumes measured at day 15 corresponded to
a non-statistically significant treatment/control (T/C) ra-
tio of 50% at the highest dose tested, 75 mg/kg bid daily
po (Fig. 3).
The increase in the cellular potency of compound 7 as
compared to 6 also translates into improved in vivo effi-
cacy. At an equivalent dosing schedule of 50 mg/kg bid
daily, compound 7 achieved a statistically significant T/
C value of 31% versus the 80% value obtained with com-
pound 6. Almost tumor stasis (T/C of 9%) is achieved at
60 mg/kg with the same schedule (Fig. 5). Ex vivo anal-
Further modification of compound 6 by changing the
imidazo ring to N-methyl imidazolinone (compound 7)
resulted in a complete loss in PDK1 inhibitory activity.
In spite of this, compound 7 retained its activity against
PI3K (IC50 = 72, 2336, 201, and 382 nM for p110a,
p110b, p110d, and p110c, respectively) and effectively
2000
1800
1600
1400
1200
1000
800
600
400
200
0
8
10
11
12
13
14
15
16
9
Days post tumor transplantation
Figure 3. In vivo efficacy of compound 6. PC3M tumor bearing
animals were treated with compound formulated in NMP/PEG300, po
at a dose of either 50 (open circles) or 75 (dark triangles) mg/kg, twice
per day. The antitumor activity measured as T/C (mean difference of
treated group divided by the mean difference of vehicle treated control
group (dark circles), multiplied by 100) on day 15 was 85% and 50%,
respectively.
Figure 4. (a) Compound 2 docked in the X-ray co-crystal structure of
PDK1 with ATP (PDB code 1H1W) as compared to (b) the docking of
compound 7 in the X-ray co-crystal structure of PI3Kc with
staurosporine (PDB code 1E8Z).

1030
F. Stauffer et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1027–1030
ity profile against one or several of the components of
the pathway will maximize antitumor activity and
tolerability.
2000
1800
1600
1400
1200
1000
800
Compounds 6 and 7 have demonstrated their ability to
reduce tumor growth in an in vivo mouse model. Fur-
thermore, both compounds provide good starting points
to explore the possibility of developing dual PI3K/
PDK1 or selective PI3K inhibitors. Progress in these ser-
ies will be reported in due course.
600
*
*
400
200
0
Acknowledgments
14
16
18
20
22
24
26
28
Days post tumor transplantation
We thank Drs. D. Fabbro and K. Shoemaker for their
support and M. Hattenberger, M. Muller, S. Haller,
and E. Sager for their outstanding technical assistance.
Figure 5. Efficacy of compound 7. PC3M tumor bearing animals were
treated with compound formulated in NMP/PEG300, po at a dose of
either 50 (open circles) or 60 (dark triangles) mg/kg, twice per day. The
antitumor activity measured as T/C (mean difference of treated group
divided by the mean difference of vehicle treated control group (dark
circles), multiplied by 100) on day 27 was 31 and 9%, respectively.
*p < 0.05% versus control, one way ANOVA followed by post hoc
Dunnet’s test.
References and notes
1. Sarbassov, D. D.; Guertin, D. A.; Ali, S. M.; Sabatini, D.
M. Science 2005, 307, 1098.
2. (a) Kim, D.; Cheng, G. Z.; Lindsley, C. W.; Yang, H.;
Cheng, J. Q. Curr. Opin. Invest. Drugs 2005, 6, 1250; (b)
Hennessy, B. T.; Smith, D. L.; Ram, P. T.; Lu, Y.; Mills, G.
B. Nat. Rev. Drug Dis. 2005, 4, 988; (c) Stauffer, F.; Holzer,
P.; Garcia-Echeverria, C. Curr. Med. Chem.: Anti-Cancer
Agents 2005, 5, 449; (d) Morgensztern, D.; McLeod, H. L.
Anti-Cancer Drugs 2005, 16, 797; (e) Redaelli, C.; Granucci,
F.; De Gioia, L.; Cipolla, L. Mini-Rev. Med. Chem. 2006, 6,
1127; (f) Knight, Z. A.; Shokat, K. M. Biochem. Soc. Trans.
2007, 35, 245.
yses of tumor tissue after a single dose administration
(50 mg/kg po) showed a time-dependent correlation be-
tween compound concentration and PKB phosphoryla-
tion inhibition for compounds 6 and 7, but the latter
displaying a lower maximal concentration than the for-
mer (data not shown).
In conclusion, we have demonstrated the ability of a ser-
ies of imidazo[4,5-c]quinoline derivatives—compounds
2–7—to inhibit the kinase activity of components of
the PI3K/PKB-pathway. The effect of their inhibitory
activity on PDK1 and/or PI3K can be tracked to the cel-
lular inactivation of PKB and blockade of the pathway.
This biological effect correlates with a significant anti-
proliferative activity against tumor cell lines in which
the PI3K/PKB-pathway is known to be deregulated
via PTEN deletion. The results obtained with the dual
PDK1/PI3K inhibitors show that concurrent inhibition
of these enzymes is an effective strategy to block the
phosphorylation and activation of PKB. However, inhi-
bition of PI3K activity, as illustrated with compound 7,
is also sufficient to control PKB activation in cellular
and in vivo settings. The lack of a selective PDK1 inhib-
itor makes the evaluation of the PDK1 contribution to
the efficacy of the dual inhibitors difficult, but the simul-
taneous blockade of PDK1 and PI3K kinases by the
same compound may provide a therapeutic advantage
in terms of avoiding or delaying the appearance of drug
resistance. It is unclear at this moment if a specific activ-
3. Mora, A.; Komander, D.; van Aalten, D. M. F.; Alessi, D.
R. Semin. Cell Dev. Biol. 2004, 15, 161.
4. Hawkins, P. T.; Anderson, K. E.; Davidson, K.; Stephens,
L. R. Biochem. Soc. Trans. 2006, 34, 647.
5. Bader, A. G.; Kang, S.; Vogt, P. K. Proc. Natl. Acad. Sci.
U.S.A. 2006, 103, 1475.
6. Marone, R.; Cmiljanovic, V.; Giese, B.; Wymann, M. P.
Biochim. Biophys. Acta 2007. doi:10.1016/j.bbapap.
2007.10.003.
7. Analytical data for compound 7: 1H NMR ((CD3)2SO,
600 MHz) d 9.04 (s, 1 H, H-4), 8.61 (m, 1 H, H-2 pyridine),
8.57 (dm, 1 H, J = 4.9 Hz, H-6 pyridine), 8.05 (d, 1 H,
J = 8.7 Hz, H-6), 7.87 (AA0, 2 H, H-3,5 phenyl), 7.84 (dt, 1
H, J = 7.8 Hz, J = 1.8 Hz, H-4 pyridine), 7.72 (BB0, 2 H, H-
2,6-phenyl), 7.65 (dd, 1 H, J = 8.7 Hz, J = 1.7 Hz, H-7),
7.44 (dd, 1 H, J = 7.9 Hz, J = 4.9 Hz, H-5 pyridine), 7.04
(m, 1 H, H-9), 3.60 (s, 3 H, CH3), 1.79 (s, 6 H, (CH3)2). 13
C
NMR d 152.90, 151.42, 149.28, 143.73, 138.42, 134.57,
134.43, 130.86, 129.44, 128.40, 128.26, 127.03, 124.40,
124.11, 123.63, 123.54, 118.85, 118.84, 114.36, 91.87,
86.95, 36.76, 28.15, 27.75. HR-MS: Calcd for
C28H22N5O+: 444.18189. Found: 444.18186.
8. Hundertmark, T.; Littke, A. F.; Buchwald, S. L.; Fu, G. C.
Org. Lett. 2000, 2, 1729.