Novel azulene-based derivatives as potent multi-receptor tyrosine kinase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2010.08.025

A series of azulene-based derivatives were synthesized as potent inhibitors for receptor tyrosine kinases such as FMS-like tyrosine kinase 3 (FLT-3). Systematic side chain modification of prototype 1a was carried out through SAR studies. Analogue 22 was i

Full text of DOI:10.1016/j.bmcl.2010.08.025

Bioorganic & Medicinal Chemistry Letters 20 (2010) 6129–6132
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Novel azulene-based derivatives as potent multi-receptor tyrosine
kinase inhibitors
Chih-Hung Chen a, , On Lee a, Chung-Niang Yao a, Meng-Yun Chuang a, Yow-Lone Chang a, May-Hua Chang a,
⇑
Yen-Fang Wen a, Wan-Hsu Yang a, Ching-Huai Ko a, Nien-Tzu Chou a, Mai-Wei Lin a, Chin-Pen Lai a,
Chung-Yuan Sun a, Ling-mei Wang a, Yen-Chun Chen a,b, Tzong-Hsiung Hseu b, Chia-Ni Chang a,
Hui-Chun Hsu a, Hui-Chi Lin a, Yu-Li Chang a, Ying-Chu Shih a, Shuen-Hsiang Chou a, Yi-Ling Hsu a,
Hsiang-Wen Tseng a, Chih-Peng Liu a, Chia-Mu Tu a, Tsan-Lin Hu a, Yuan-Jang Tsai a, Ting-Shou Chen a,
Chih-Lung Lin a, Shu-Jiau Chiou a, Chung-Cheng Liu a, Chrong-Shiong Hwang a
a Biomedical Engineering Research Laboratories, Industrial Technology Research Institute, Hsinchu 30011, Taiwan
b Institute of Biotechnology, National Tsing Hua University, Hsinchu 30013, Taiwan
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of azulene-based derivatives were synthesized as potent inhibitors for receptor tyrosine kinases
such as FMS-like tyrosine kinase 3 (FLT-3). Systematic side chain modification of prototype 1a was car-
ried out through SAR studies. Analogue 22 was identified from this series and found to be one of the most
potent FLT-3 inhibitors, with good pharmaceutical properties, superior efficacy, and tolerability in a
tumor xenograft model.
Received 31 March 2010
Revised 4 August 2010
Accepted 5 August 2010
Available online 10 August 2010
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
Azulene
FMS-like tyrosine kinase 3
FLT-3
Acute myeloid leukemia
AML
Receptor tyrosine kinases (RTKs) are a large family of trans-
membrane receptors with diverse biological activity. At least 19
RTK subfamilies have been identified. One example is the plate-
let-derived growth factor receptor (PDGFR) subfamily. Members
adding a side chain to an azulene moiety or by altering oxindole
into azaoxindole, were investigated in this study.
The synthetic route for preparing azulene-oxindole-based
derivatives is shown in Scheme 1. Treatment of commercially
available tropolone with p-toluene sulfonyl chloride and triethyl
amine yielded p-(tosyloxy)tropone 4, whereas treatment of
p-(tosyloxy)tropone with dimethyl malonate in the presence of
sodium methoxide in toluene resulted in 5.9 Treatment of 5 with
tetrabutyl ammonium hydroxide in alkaline solution produced 6,
which was then treated with dibromoalkane in acetonitrile to yield
7a (n = 1) and 7b (n = 2). Compound 7a underwent [8+2] cycload-
of this family include PDGFRa, PDGFb, colony-stimulating factor-
1 receptor (CSF-1R), FMS-like tyrosine kinase 3 (FLT-3), and c-
KIT, and these are believed to promote angiogenesis and tumor cell
growth. The constitutive activation of FLT-3 and c-KIT by mutation
is directly associated with acute myeloid leukemia (AML) and gas-
trointestinal stromal tumors (GIST). FLT-3 is expressed on blast
cells in most patients with AML, and internal tandem duplication
(ITD) of the FLT-3 mutation has been found in up to 30% of all
AML patients. The important role played by FLT-3 in the survival
and proliferation of blast cells and its overexpression in most pa-
tients with AML make FLT-3 an attractive therapeutic target.1–8 Ra-
tional design efforts in our laboratory identified an azulene-
oxindole lead compound, designated 1a, as a submicromolar inhib-
itor of FLT-3 (Fig. 1). Modifications of this lead structure, either by
Y
X
F
O
O
N
N
H
H
2a
1a X = Y = H
FLT-3 IC50 = 0.47 µM
FLT-3 89% inhibition (1 µM)
⇑
Corresponding author.
E-mail address: ChihHungChen@itri.org.tw (C.-H. Chen).
Figure 1. Novel azulene-based compounds.
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.08.025

6130
C.-H. Chen et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6129–6132
COOMe
O
COOMe
OTs
OH
O
c
b
a
O
O
O
O
O
OH
5
TBA
3
4
6
COOMe
COOMe
COOMe
O
e
d
+
O
O
n
O
n
O
N
Br
O
9a n = 1
9b n = 2
n
7a n = 1
7b n = 2
8a n = 1
8b n = 2
Br
CH2OH
CHO
O
N
n
O
g
h
f
9a n = 1
9b n = 2
O
n
O
n
F
O
10a n = 1
10b n = 2
12a n = 1
11a n = 1
11b n = 2
N
N
H
N
12b n = 2
O
O
Scheme 1. Synthesis of compounds 12a and 12b. Reagents: (a) TsCl, Et3N/CH2Cl2; (b) CH2 (COOMe)2, MeONa/MeOH, toluene; (c) NaOH/H2O, tetrabutyl ammonium
hydroxide; (d) 1,2-dibromoethane (n = 1), 1,2-dibromopropane (n = 2)/CH3CN; (e) EtCHO, morpholine/toluene; (f) DIBAL/ether; (g) MnO2/CH2Cl2; (h) 5-fluoro-2-oxindole/
ethanol, piperidine.
dition with morpholine and propionyl aldehyde in toluene to pro-
duce the bromo compound 8a and morpholine-substituted com-
pound 9a. Reduction of the methyl ester 9a with DIBAL/heptane
in ether formed the primary alcohol 10a and subsequent oxidation
with MnO2/CH2Cl2 resulted in the aldehyde 11a. The subsequent
reaction of 11a with 5-fluoroxindole yielded the desired compound
12a in the single isomer (Z) form. Compound 12b could be pre-
pared by a procedure similar to that described for 12a.
The bromo compound 8b (n = 2) was treated with a selective
amine to produce 13a–b, and procedures b, c, and d were used to
produce the desired products 16a–b (Scheme 2). All these final
products could be obtained in the single isomer form or as a mix-
ture of the E/Z isomers. The latter could be separated by chroma-
tography and were assigned by 1H NMR experiments.10
1a was used as the lead compound in the series, and it showed
moderate inhibitory activity against the FLT-3 enzyme in the sub-
micromolar range (0.47 lM). The results of the SAR studies on 1a
are shown in Tables 1–3. We investigated the effects of different
substituents on the oxindole and azulene rings. As shown in Table
1, fluoro substitution at the 5-position of oxindole and incorpora-
tion of a fluoro or methyl group at the 3-position of the azulene
core moiety was essential for potency in the series (1h–1j).
Replacement of the methyl group of 1i (Z) with hydrogen resulted
in loss of potency (1b (Z) >1000 nM). Removal of the fluoro group
from 1i (Z) yielded 1g (Z), which had significantly lower inhibitory
activity. This trend was also observed in the case of 1h (E).
To improve the inhibitory activity against FLT-3, we explored
the effects of attaching an alkylaminoalkyl ether link to the seven
member ring (Table 2). Various terminal alkylamino groups were
investigated, including cyclic and noncyclic tertiary amine groups.
In most cases (X = 5-F, Y = C), a spacer length of 2–3 methylene
units was tolerated (except 12b (Z)) for FLT-3 inhibitory activity.
The compounds were assayed for their activity against a panel
of receptor tyrosine kinases, including FLT-3, c-KIT, and KDR. Given
that FLT-3 plays a pivotal role in AML, our emphasis was on opti-
mizing the potency of the compound against FLT-3. Compound
CO2CH3
CHO
CH2OH
c
b
a
8b n = 2
O
n
O
O
R
R
R
N
R
N
n
N
R
n
R
15a-b n = 2
14a-b n = 2
13a-b n = 2
R
O
N
R
n
13a, 14a, 15a, 16a: NRR = dimethylamine
13b, 14b, 15b, 16b: NRR = diethylamine
d
F
22: NRR = L-prolinol
O
23: NRR = 4-(N,N-dimethyl) piperidine
N
H
16a-b n = 2
22
23
Scheme 2. Synthesis of compounds 16a-b, 22, and 23. Reagents: (a) Amine/CH3CN; (b) DIBAL/ether; (c) MnO2/CH2Cl2; (d) 5-fluoro-2-oxindole/ethanol, piperidine.

C.-H. Chen et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6129–6132
6131
Table 1
Table 3
SAR at the X/R1 position
Variation of the R moiety in 4,5-difluoro-oxindole
OR
R1
F
X
F
O
O
N
H
N
H
R1
FLT-3 inhibition IC50 (nM)
Compound
R
FLT-3 inhibition IC50 (nM)
a
a
Compound
X
1a (E/Z 6:1)
1b (Z)
1c (E)
1d (Z)
1e (Z)
1f (E)
1g (Z)
1h (E)
1i (Z)
1j (Z)
H
F
H
H
H
H
H
CH3
CH3
CH3
CH3
F
470
24 (Z)
25 (Z)
5
N
N
>1000
>1000
>1000
>1000
>1000
>1000
60
128
23
>1000
3
Br
Br
CH3
H
H
F
14
N
26 (Z)
27 (Z)
12
17
O
N
F
F
a
See Ref. 11 for assay details.
1k (Z)
SU11248
SO2N(CH3)2
H
a
See Ref. 11 for assay details.
Table 4
In vitro cellular assay of compound 22
Table 2
SAR of X, Y, and R
In vitro assay
Cells/cellular kinase
MV4-11
22
SU11248
9
a
Cell proliferation IC50 (nM)
87
OR
Cellular kinase
phosphorylation IC50 (nM)
FLT-3-ITD
FLT-3-WT
cKIT
KDR
PDGFRb
67
70
44
186
28
ND
16
15
53
17
b
X
O
N
Y
a
H
See Ref. 12
See Ref. 13 and14 for assay details.
b
a
Compound
12a (Z)
Y
C
C
X
R
FLT-3 inhibition IC50 (nM)
N
5-F
5-F
19
Table 5
O
In vivo pharmacokinetic parameters of 22 (n = 3/dosing route)
N
Dose routea
Intravenous
Parameterb
12b (Z)
224
O
CL (mL/h/kg)
Vss (mL/kg)
T1/2 (h)
709.6
1232.4
1.8
16a (E)
16a (Z)
16b (E)
C
C
C
5-F
5-F
5-F
25
12
30
N
N
AUC (h ng/mL)
1413.4
N
Oral
Tmax (h)
2.0
566.3
3222.1
22.8
Cmax (ng/mL)
AUC (h ng/mL)
F (%)
16b (Z)
17 (Z)
C
5-F
H
39
N
N
1460
N
a
N
Dosing vehicle: 10% N-methyl-2-pyrrolidone, 20% cremo-
18 (Z)
19 (Z)
N
N
5-F
90
86
O
O
phor EL, and 70% phosphate-buffered saline. Values represent
the mean value of a study with three animals.
N
5-Br
b
See Ref. 15 for assay details.
20 (E)
21 (E)
C
C
6-F
7-F
742
177
N
(E). Similarly, incorporation of a fluoro group at the 7-position in
21 (E) resulted in a sevenfold loss of activity in comparison to
16a (E). We also studied the effect of the 7-azaoxindole moiety
on FLT-3 inhibitory activity. Compounds 18 (Z) and 19 (Z) dis-
played moderate activity (90 and 86 nM), whereas 17 (Z) showed
severe loss of potency when the halo group (F, Br) was removed
from the 5-position of the azaoxindole moiety. Furthermore, we
utilized the 4,5-difluoro substitutions of oxindole to constrain
our compound in the Z form only. Although these compounds
showed potent inhibitory activity against FLT-3, their solubility
was poor (data not shown). All of these results are summarized
in Table 3.
N
N
22 (E/Z 1.6:1)
C
C
5-F
5-F
98
18
OH
N
23 (Z)
N
a
See Ref. 11 for assay details.
The E/Z isomers 16b (E) and 16b (Z) showed almost equipotent
activity, whereas 16a (E) was twofold less potent than 16a (Z).
Next, we investigated the effects of C(6) and C(7) substituents on
the oxindole ring. Incorporation of a fluoro group at the 6-position
in 20 (E) led to a 29-fold loss in potency in comparison with 16a
Western blot analysis was used to test the target compounds for
their in vitro ability to inhibit cellular proliferation12 and inhibit
receptor tyrosine kinase autophosphorylation in cellular mod-
els.13,14 Compound 22 exhibited high cytotoxic potency against

6132
C.-H. Chen et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6129–6132
References and notes
1. Parcells, B. W.; Ikeda, A. K.; Simms, W. T.; Moore, T. B.; Sakamoto, K. M. Stem
Cells 2006, 24, 1174.
2. Furet, P.; Bold, G.; Meyer, T.; Roesel, J.; Guagnano, V. J. Med. Chem. 2006, 49,
4451.
3. Bhalla, K. N.; George, P.; Gutti, R.; Bali, P.; Tao, J.; Guo, F.; Sigua, C.; Li, Y.; Cohen,
P.; Atadja, P. A. J. Clin. Oncol. 2004, 22, 6541 (Post-meeting Edition).
4. Cools, J.; Mentens, N.; Furet, P.; Fabbro, D.; Clark, J. J.; Griffin, J. D.; Marynen, P.;
Gilliland, D. G. Cancer Res. 2004, 64, 6385.
5. Markovic, A.; MacKenzie, K. L.; Lock, R. B. Int. J. Biochem. Cell Biol. 2005, 37,
1168.
6. Kiyoi, H.; Shiotsu, Y.; Ozeki, K.; Yamaji, S.; Kosugi, H.; Umehara, H.; Shimizu,
M.; Arai, H.; Ishii, K.; Akinaga, S.; Naoe, T. Clin. Cancer Res. 2007, 13, 4575.
7. Patch, R. J.; Baumann, C. A.; Liu, J.; Gibbs, A. C.; Ott, H.; Lattanze, J.; Player, M. R.
Bioorg. Med. Chem. Lett. 2006, 16, 3282.
8. Mead, A. J.; Linch, D. C.; Hills, R. K.; Wheatley, K.; Burnett, A. K.; Gale, R. E. Blood
2007, 110, 1262.
9. Nozoe, T.; Asao, T.; Yasunami, M.; Wakui, H.; Suzuki, T.; Ando, M. J. Org. Chem.
1995, 60, 5919.
10. Lackey, K.; Cory, M.; Davis, R.; Frye, S. V.; Harris, P. A.; Hunter, R. N.; Jung, D. K.;
McDonald, O. B.; McNutt, R. W.; Peel, M. R.; Rutkowske, R. D.; Veal, J. M.; Wood,
E. R. Bioorg. Med. Chem. Lett. 2000, 10, 223.
11. In vitro kinase assays. FLT-3 kinase inhibition was determined by a time-
resolved fluorescence method. The resulting phosphorylated peptide was
measured in streptavidin-coated microtiter plates using the ULTRA Evolution
Multi-Detection Microplate Reader from Tecan. Phospho-tyrosine mouse mAb
(P-Try-100) and europium-labeled anti-mouse IgG were used for the assay. The
IC50 value of the compound was calculated by nonlinear regression with Sigma
Plot 10.0.
Figure 2. Compound 22 leads to tumor regression in MV4-11 xenograft nude mice.
This compound was administered at 50 mg/kg (po, b.i.d.) from day 16 to day 29
after subcutaneous implantation of MV 4–11 cells. No lethality or body loss was
observed in any group. The data represent the mean SEM (n = 4).
MV4-11 with IC50 values in the submicromolar range and also
showed high inhibitory activity against KDR, cKIT, PDGFRb, and
FLT-3 (Table 4). Taken together, the in vitro and cellular assay re-
sults demonstrated that 22 is a potent multi-receptor tyrosine ki-
nase inhibitor that inhibits kinase phosphorylation both in vitro
and in cells.
12. Cell proliferation assays. MV4-11 cells were plated in 96-well microtiter plates
(1 Â 104 cells per well), and cell viability was determined using MTT,
a
tetrazolium dye. The formazan crystals were dissolved in DMSO, and the
absorbance at 600 nm was recorded in an ELISA plate reader. The IC50 values
were calculated by nonlinear regression and defined as the concentration
required for 50% reduction in the absorbance of treated versus untreated
control cells.
Plasma pharmacokinetic studies were carried out with com-
pound 22 in rats.15 A single dose of the compound was adminis-
tered intravenously or orally. As shown in Table 5, compound 22
exhibited moderate clearance (0.7 L/h/kg), a short elimination half
life (1.8 h), and a medium volume of distribution (1.2 L/kg). The
oral bioavailability was acceptable (F = 22.8%). Compound 22 had
a notable antitumor activity in human leukemia MV4-11 xeno-
grafts in BALB/c nude mice.16 Vehicle or 22 was given at 50 mg/
kg (twice a day; b.i.ds.) for 14 days per oral (po), and tumor growth
inhibition (in %) by 22 treatment was 99 0.82% on day 30 (Fig. 2).
During the same time period, no progressive weight loss and
lethality were observed during the experimental period. These
observations indicated that compound 22 was well tolerated and
was highly efficacious in this human leukemia xenograft model.
In conclusion, a novel series of FLT-3 inhibitors were designed
and evaluated. Optimization led to the development of compound
22, which showed biochemical potency against its main target as
well as good potency in proliferation and cell-based mechanistic
assays. Compound 22 also had an acceptable pharmacokinetic pro-
file, with plasma exposure that was sufficient for obtaining good
antitumor activity in vivo. The azulene-based derivatives shown
here can potentially be developed as a new class of antileukemia
agents. Of particular interest, these compounds have been found
to inhibit not only FLT-3 and KDR but also TrkA and pim-1 kinases
(data not shown). The benefits of compound 22 in other cancers,
such as pancreatic tumors, will be evaluated in the future.
13. Chen, Y. C.; Chang, C. N.; Hsu, H. C.; Chiou, S. J.; Lee, L. T.; Hseu, T. H. Life Sci.
2009, 84, 915.
14. Western blot analysis. The cell lines used for measuring autophosphorylation in
this study included Kasumi-1(cKIT), HUVEC(KDR), RS4-11(FLT3), and MV4-
11(FLT3-ITD); these were obtained from the Bioresource Collection and
Research Center (Taiwan) and American Type Culture Collection (USA) and
were cultured as described in literature. Receptor stimulation was carried out
with each ligand (stem cell factor (SCF), vascular endothelial growth factor
(VEGF-162), and FLT-3 ligand (FL)). The probe primary anti-phosphotyrosine
antibody (PY99) and reprobe antibodies for c-KIT, KDR, and FLT-3 were
obtained from Santa Cruz Biotechnology and R and D Systems. The cells were
plated at 2–5 Â 105 cells per well in six-well plates. More details of the
experimental procedures are described in Ref. 13.
15. In vivo pharmacokinetics. The pharmacokinetic profile of 22 was investigated in
overnight-fasted male SD rats after administration of a single dose of the
compound intravenous or oral. The blood was collected and processed to
plasma by centrifugation and stored at À80 °C. Samples were analyzed by the
LC/MS/MS technique. The pharmacokinetics were analyzed by WinNonlin 5.2
(Pharsight) via noncompartmental analysis.
16. In vivo xenograft tumor model. The procedures for establishing tumor
xenografts and the dosing of 22 were carried out in accordance with the
guidelines of the ITRI Institutional Animal Care and Use Committee (IACUC).
Female BALB/c nude mice (CAnN.Cg-Foxn1NU/Crl, 6–8 weeks old) were
purchased from BioLASCO (Taiwan) and acclimated for 1 week in the Animal
Laboratory for Biomedical Research, ITRI. Female BALB/c nude mice were
implanted subcutaneously in the right flank with 1 Â 107 MV4-11 cells per
mice. Treatments were initiated when the tumor size was 200–300 mm3. Mice
were randomly assigned to cohorts (four mice per group for efficacy studies).
Compound 22 (50 mg/kg, b.i.d.) and vehicle (10% NMP + 20% Cremophor
EL + 70% saline) were given via oral gavage for 14 days. Caliper measurements
of the tumors were converted into mean tumor volume using the following
formula: 0.5  [length  (width)2].