Combretastatin A-4 derived 5-(1-methyl-4-phenyl-imidazol-5-yl)indoles with superior cytotoxic and anti-vascular effects on chemoresistant cancer cells and tumors

Article abstract of DOI:10.1016/j.ejmech.2016.04.045

5-(1-Methyl-4-phenyl-imidazol-5-yl)indoles 5 were prepared and tested as analogs of the natural vascular-disrupting agent combretastatin A-4 (CA-4). The 3-bromo-4,5-dimethoxyphenyl derivative 5c was far more active than CA-4 with low nanomolar IC₅₀

Full text of DOI:10.1016/j.ejmech.2016.04.045

European Journal of Medicinal Chemistry 118 (2016) 9e20
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Research paper
Combretastatin A-4 derived 5-(1-methyl-4-phenyl-imidazol-5-yl)
indoles with superior cytotoxic and anti-vascular effects on
chemoresistant cancer cells and tumors
Katharina Mahal a, Bernhard Biersack a, Sebastian Schruefer a, Marcus Resch b,
,
Ralf Ficner b, Rainer Schobert a *, Thomas Mueller c
a Organic Chemistry Laboratory, University Bayreuth, 95440 Bayreuth, Germany
b
€
€
Department of Molecular Structural Biology, Georg-August-University Gottingen, 37077 Gottingen, Germany
c Department of Internal Medicine IV, Oncology/Hematology, Martin-Luther-University Halle-Wittenberg, 06120 Halle-Saale, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
5-(1-Methyl-4-phenyl-imidazol-5-yl)indoles 5 were prepared and tested as analogs of the natural
vascular-disrupting agent combretastatin A-4 (CA-4). The 3-bromo-4,5-dimethoxyphenyl derivative 5c
was far more active than CA-4 with low nanomolar IC50 concentrations against multidrug-resistant KB-
V1/Vbl cervix and MCF-7/Topo mamma carcinoma cells, and also against CA-4-resistant HT-29 colon
carcinoma cells. While not interfering markedly with the polymerization of tubulin in vitro, indole 5c
completely disrupted the microtubule cytoskeleton of cancer cells at low concentrations. It also
destroyed real blood vessels, both in the chorioallantoic membrane (CAM) of fertilized chicken eggs and
within tumor xenografts in mice, without harming embryo or mouse, respectively. Indole 5c was less
toxic than CA-4 to endothelial cells, fibroblasts, and cardiomyocytes. In highly vascularized xenograft
tumors 5c induced distinct discolorations and histological features typical of vascular-disrupting agents,
such as disrupted vessel structures, hemorrhages, and extensive necrosis. In a first preliminary therapy
trial, indole 5c retarded the growth of resistant xenograft tumors in mice. © 2016 Elsevier Science Ltd. All
rights reserved.
Received 18 March 2016
Received in revised form
15 April 2016
Accepted 16 April 2016
Available online 19 April 2016
Keywords:
Imidazole
Indole
Anticancer drugs
Combretastatin A-4
Vascular-disrupting agents
Mouse tumor xenografts
© 2016 Elsevier Masson SAS. All rights reserved.
1. Introduction
biologically inactive E-isomer in protic solvents. So far, the search
for chemically more stable derivatives has led to compounds with
The natural tubulin binding agent combretastatin A-4 (CA-4,
Fig. 1) was isolated from the bark of the South African Cape Bush-
willow (Combretum caffrum). Its water soluble phosphate prodrug
Zybrestat™ (fosbretabulin) has entered advanced clinical trials that
revealed its tumor-selective angiotoxicity, even in multidrug-
resistant tumors [1,2]. However, due to its insufficient in vivo
cytotoxicity it had to be applied in combination with other drugs
such as carboplatin, paclitaxel or the anti-angiogenic agent bev-
acizumab [3,4]. The related combretastatin A-1 and its bisphos-
phate prodrug OXi4503 also proved efficacious against certain
tumor cells in in vivo models for which their redox-active catechol
moiety was believed to be responsible [5,6]. A drawback of com-
bretastatins is the tendency of their Z-alkene to isomerize to a
the alkene adopting a fixed Z-configuration, e.g., by being part of a
heterocycle [7]. Some 4,5-diaryloxazoles, -imidazoles, and -1,2,3-
triazoles of this type actually retained the cytotoxicity and
tubulin affinity of the parent combretastatins [8,9]. Among these
were orally applicable, water soluble derivatives with advanta-
geous pharmacokinetics and distinct in vivo activity such as the
imidazole 1 [8]. In another study 3-halo- and 3-aminostilbenes 2,
structurally reminiscent of combretastatin A-3, showed increased
affinity for tubulin and a more selective activity profile against
tumor cells with potential to overcome drug resistance [10]. Yet,
both structure variants have their shortcomings. Derivative 1 is
inferior to combretastatin A-4 with respect to cytotoxicity while
the analogous oxazoles are hampered with unsatisfactory phar-
macological properties [8]. The halo combretastatins 2 prepared by
Pettit et al. were quite active yet chemically unstable and required
formulations as phosphate prodrugs in order to reach sufficient
water solubility and uptake rates [10].
* Corresponding author.
E-mail address: rainer.schobert@uni-bayreuth.de (R. Schobert).
http://dx.doi.org/10.1016/j.ejmech.2016.04.045
0223-5234/© 2016 Elsevier Masson SAS. All rights reserved.

10
K. Mahal et al. / European Journal of Medicinal Chemistry 118 (2016) 9e20
N
N
R1
R2
MeO
OH
OMe
NH2
OMe
MeO
MeO
A
B
OMe
OMe
Combretastatin A-1: R1 = OH, R2 = OMe
Combretastatin A-3: R1 = H, R2 = OH
Combretastatin A-4: R1 = H, R2 = OMe
1
MeO
Hal = F, Cl, Br
OH
OMe
MeO
Hal
2
Fig. 1. Combretastatins A-1, A-3, and A-4 and known analogs 1 and 2: potent inhibitors
of tubulin polymerization. Designation of A- and B-ring in the natural drugs com-
bretastatin A.
Scheme 1. Reagents and conditions: (i) TosMIC reagent 4, MeNH2, AcOH, EtOH, reflux,
2 h, then 3, K2CO3, EtOH, reflux, 3 h; (ii) 3 M HCl/dioxane, CH2Cl2, r.t., 15 min.
CA-4 fragments were also covalently combined with other
tubulin modifier motifs such as indoles. For instance, Dalton et al.
panel of cancer cell lines was tested via MTT assays and compared
with that of CA-4 (Table 1). Though the lead compound CA-4 dis-
played the greatest cytotoxicity against highly proliferative 518A2
melanoma cells, it showed only moderate activity against the
multidrug-resistant cervix and breast carcinoma cell lines KB-V1/
Vbl and MCF-7/Topo. HT-29 colon carcinoma cells were resistant
to CA-4 by overexpression of the MRP-1 (multidrug-resistance
protein) type ABC (ATP-binding cassette) transporter which de-
toxifies cells from CA-4 [20]. In contrast, the indoles 5a-5h were
active against the resistant cell lines with nanomolar IC50 values.
The N-methyl indoles 5a-5d were particularly active at IC50 values
in the two-digit nanomolar range. Compounds 5a bearing the
original trimethoxy A-ring motif, and 5c featuring a 3-bromo-4,5-
dimethoxyphenyl A-ring, were slightly more active against the
resistant cancer cells than the chloro- and iodo-derivatives 5b and
5d. Within the triad of N-ethyl substituted indoles 5e-5g, only the
trimethoxyphenyl derivative 5e showed a comparable efficacy. Any
further substitution at the indole moiety resulted in an attenuated
growth inhibition. We hypothesize that the cell growth inhibiting
effect of the compounds 5a-i might be correlated to their ability to
bind to tubulin as reported for CA-4.
disclosed
the
antitumoral
2-(trimethoxybenzoylphenyl)
substituted indole I-387 [11]. Romagnoli et al. reported a 2-aroyl-3-
amino derivative which inhibited the growth of breast cancer cells
[12]. Simoni et al. prepared an (N-methylindol-5-yl)stilbene which
caused growth inhibition of triple-negative MDA-MB-231 breast
cancer cells [13]. Welsh et al. reported a triazole-bridged indole
analog of CA-4 with high tubulin affinity and anticancer activity
[14,15]. Pinney et al. developed and optimized 2-aryl-3-
aroylindoles such as OXi8006,
a highly cytotoxic vascular-
disrupting agent [16]. Zhang et al. reported 3-(trimethox-
yphenylseleninyl)-1H-indoles that bind to tubulin in a similar
orientation as CA-4 [17]. Based on the work of Wang et al. our group
developed a water-soluble hydrochloride of an N-methyl-(3-
chloroindol-5-yl)imidazole analog of combretastatin A-4 which
was far more cytotoxic against topotecan-resistant (BCRP-positive)
MCF-7 breast cancer cells than CA-4 [8,18]. We also prepared CA-4
analogous imidazoles with meta-halo-substituted A-rings that had
an improved anticancer activity [18,19]. Herein, we continue this
work and report on new CA-4 analogous imidazoles with 5-indole
residues and their superior anticancer and antivascular properties
in vitro and in vivo when compared with the lead compound CA-4
and with previously published indole 5a that had been tested only
on two cancer cell lines before [8].
The best indole compounds 5a and 5c were additionally tested
against HCT-116 colon carcinoma cells with either a functional tu-
mor suppressor protein p53 (wildtype, wt) or a knocked-out p53
(HCT-116 p53ꢀ/ꢀ). While the IC50 values of CA-4 and 5c were
similar against both cell lines, the trimethoxyphenyl derivative 5a
was more cytotoxic against p53-wildtype cells (Table 2).
2. Results and discussion
To assess the selectivity of the highly active derivatives 5a and
5c for cancer over non-malignant cells, we also determined their
IC50 values against primary fibroblasts (CHF) and cardiomyocytes
(CCM) that were explanted from chicken embryos (Table 3). All
values for CHF were in the micromolar range with compound 5c
being least cytotoxic. It also affected cardiomyocytes to a lesser
extent than CA-4 and derivative 5a. The distinct in vitro cytotoxicity
of the anticancer drug doxorubicin against cardiomyocytes is
correlated with its severe cardiotoxic side-effects in vivo [22]. We
therefore conclude that the bromo substituted indole 5c with the
best selectivity profile and the least toxic effects on chicken fibro-
blasts and cardiomyocytes will probably be the one with the lowest
general toxicity in subsequent in vivo experiments. It is worth
noting that the endothelial cell line included in our growth inhi-
bition tests (Table 1) was less affected by 5a and 5c than by CA-4
and that endothelial damage was also reported to be responsible
for cardiotoxic side-effects of some tubulin-binding agents [23].
However, Ea.hy926 is a hybrid cell line that does not fully reflect the
2.1. Chemistry
The 5-formylindole derivatives 3a-d and the TosMIC reagents
4a-d were prepared according to literature procedures [8,18]. The
aldehydes 3a-d were treated with MeNH2 to give the respective
imines, which were reacted with the TosMIC reagents 4a-d under
basic conditions to give the corresponding new N-methyl imidaz-
oles 5b-i (Scheme 1). The known compound 5a was prepared
analogously [8]. All compounds 5 were finally converted with 3 M
HCl in dioxane to the corresponding hydrochlorides, which were
purified by recrystallization. 5-Formylindoles that lacked a 3-chloro
substituent gave no stable hydrochlorides but underwent a rapid
degradation with dark red discoloration.
2.2. Biological evaluation
2.2.1. Cytotoxicity
The growth inhibitory activity of the compounds 5a-i against a

K. Mahal et al. / European Journal of Medicinal Chemistry 118 (2016) 9e20
11
Table 1
Inhibitory concentrations IC50 [nM] of combretastatin A-4 (CA-4) and the indoles 5a-i when applied to various human cancer cell lines and Ea.hy926 hybrid endothelial cells.
Cell line
518A2
HT-29
KB-V1/Vbl
MCF-7/Topo
Ea.hy926
b
a
b
a
b
a
b
a
b
a
b
a
b
a
b
a
b
a
b
c
c
c
CA-4
5a
1.8 0.1
>5000
300 200
26.0 2.9
154 33
11.0 2.0 c
24.6 2.8
65.4 7.3
18.2 0.5
174 13.7
32.4 2.0
51.9 6.3
28.2 1.3
56.7 4.8
35.7 2.9
739 121
33.0 8.0
216 76
82 4.1
26.2 5.2
15.6 2.4
33.7 1.7
31.4 5.8
31.3 4.7
23.4 5.9
35.1 1.2
39.7 4.8
75.8 8.9
43.4 7.7
246 37
116 22
482 34
178 37
903 100
223 44
>1000
20.2 1.2
26.9 7.2
18.8 1.4
53.3 2.3
30.2 1.7
91.3 9.9
81.2 10.1
208 38
5b
5c
5d
5e
5f
37.9 2.1
25.4 2.8
49.8 7.6
33.2 10
48.4 1.7
83.2 8.5
456 11
>1000
n. d.
29.9 4.5
n. d.
5g
5h
5i
211 51
168 30
292 43
223 44
>1000
489 39
767 88
793 89
Values are derived from doseeresponse curves obtained by measuring the percentage of viable cells relative to untreated controls after 24 h a or 72 h b exposure to the test
compounds using an MTT assay; human cancer cell lines: 518A2 melanoma, HT-29 colon, KB-V1/Vbl cervix, MCF-7/Topo breast carcinoma. Ea.hy926: endothelial hybrid cells. c
Values from earlier publications [18,21]. Values represent means of four independent experiments. Conspicuous values are highlighted in gray.
Table 2
correlates with their affinity for tubulin, we performed polymeri-
zation assays with purified porcine tubulin (Fig. 2a). CA-4 was used
as a control with known high tubulin affinity. 5 mM of CA-4 and
indole 5a were sufficient to impair the polymerization of tubulin
heterodimers to microtubules, though 5a was less active at equi-
Inhibitory concentrations IC50 [nM, 72 h]a of combretastatin A-4 (CA-4) and the
indoles 5a and 5c for the growth of human p53-wildtype and p53-negative HCT-116
colon carcinoma cells.
Cell line
CA-4
5a
9.5 1.5
5c
HCT-116 wt
HCT-116 p53ꢀ/ꢀ
a
2.6 0.2
3.4 0.4
27.4 1.5
21.3 2.3
molar concentrations. 5 mM of indole 5c slowed down the rate of
26.2 3.0
the polymerization, but could not suppress it. This divergent
tubulin affinity of compounds 5a and 5c is somewhat surprising,
given their comparable IC50 values in the cytotoxicity tests with HT-
29, KB-V1/Vbl and MCF-7/Topo cancer cells. The assumption of an
additional mode of cytotoxic action of either 5a or 5c would also be
supported by their distinctly different activities against the HCT-
116 (wt) carcinoma cells (Table 2).
Values are derived from doseeresponse curves obtained by measuring the
percentage of viable cells relative to untreated controls after 72 h. Values represent
means of three independent experiments.
Table 3
Inhibitory concentrations IC50 [m
M, 48 h]a of CA-4 and the indoles 5a and 5c on the
growth of primary chicken fibroblasts (CHF) and chicken cardiomyocytes (CCM) and
relative cancer cell selectivity (selectivity index).
By molecular docking studies we found that 5c can adopt a
favourable binding mode within the tubulin heterodimer (with a
theoretical Ebind ¼ ꢀ8.8 kcal/mol; Fig. 2b left) that differs from that
of the lead compound CA-4, and a slightly less favourable binding
mode (with a theoretical Ebind ¼ ꢀ8.3 kcal/mol; Fig. 2b right)
similar to that of CA-4.
Cell line
CA-4
5a
5c
CHF
CCM
selectivity index
5.48 0.24
0.874 0.08
1.81
11.61 1.5
1.87 0.4
2.99
23.02 1.8
3.97 1.3
3.19
Like CA-4, indole 5c can bind to the colchicine binding pocket
within the beta-tubulin subunit of the tubulin heterodimer, yet in a
different orientation (Fig. 2b, left). In this conformation, the indole
a
Values are derived from doseeresponse curves obtained by measuring the
percentage of viable cells relative to untreated controls after 48 h. Values represent
means of three independent experiments.
moiety is placed in the hydrophobic pocket made up by Leu-
b248,
Ala- 250, Ala- 316, Val- 318, Ala- 354, Ile- 378, where the A-ring
of colchicine and CA-4 is bound [25]. The p-methoxy group of the
b
b
b
b
b
sensitivity of primary endothelial cells.
phenyl ring (corresponding to the A-ring of CA-4) is locked by Van-
2.2.2. Cell cycle interference
der-Waals interactions with a methylene group of Lys-
b254 and the
The observed growth inhibition by the best two indoles 5a and
5c might be due to induction of apoptosis or inhibition of the cell
cycle progression as reported for many tubulin binding agents [24].
In cell cycle analyses with 518A2, HCT-116 and MCF-7/Topo cells
after 24 h incubation with the test compounds, we found an in-
crease in apoptotic cells, but also a significant accumulation of
treated cells in the G2/M phase (Table S1, Supplementary data). The
cell cycle progression of HT-29 colon carcinoma cells was arrested
by 5c mainly in mitosis as observed via concomitant immuno-
staining of phosphorylated histone H3 (phospho-H3) as a mitosis
marker (Table 4; Fig. S1, Supplementary data).
m-methoxy oxygen is engaged in a hydrogen bond with the side-
chain amine of Asn-a101. In addition, the non-methylated
Table 4
Percentage of HT-29 colon carcinoma cells in cell cycle phases after treatment with
DMSO (control), combretastatin A-4 (5 M) or indole 5c (50 and 100 nM) for 24 h.
m
Cell cycle phase (%)
Control
CA-4
100 nM 5c
50 nM 5c
G1
S
G2
mitosis
68.1 0.3
13.7 0.3
15.6 0.1
3.2 0.1
51.4 0.8
18.6 1.8
18.2 1.0
10.2 0.4
12.0 0.3
8.0 2.9
4.6 0.2
77.8 3.8
43.5 2.8
13.5 0.4
15.2 3.2
28.0 1.1
Values derived from flow cytometric cell cycle analyses with HT-29 cells double-
stained for their DNA content with propidium iodide and a phospho-histone H3
mitosis marker.
2.2.3. Tubulin affinity
To check the hypothesis that the compounds' cytotoxicity

12
K. Mahal et al. / European Journal of Medicinal Chemistry 118 (2016) 9e20
Fig. 2. Interference of indoles 5a and 5c with the polymerization of tubulin. a) Tubulin polymerization curves in the absence (control) or presence of combretastatin A-4 (CA-4), or
5a, or 5c recorded by the optical density (OD) at 340 nm wavelength. b) Docking studies: left: proposed binding mode of 5c (carbon colored light blue) in the binding pocket within
the beta-tubulin subunit of a bovine tubulin (pdb-code 1SA0). The energetically more favourable orientation with a theoretical Ebind ¼ ꢀ8.8 kcal/mol differs from that of CA-4. The
indole moiety is placed in the hydrophobic cleft made up by the side-chains of Leu-248, Ala-250, Ala-316, Val-318, Ala-354, and Ile-378, with carbon colored light yellow (b chain),
sulfur yellow, nitrogen blue, oxygen red. Right: proposed binding mode of 5c (carbon colored light blue) in the binding pocket within the beta-tubulin subunit. The energetically
slightly less favourable orientation with a theoretical Ebind ¼ ꢀ8.3 kcal/mol resembles that known of CA-4, with the phenyl moiety placed in the hydrophobic cleft made up by the
side-chains of Leu-248, Ala-250, Ala-316, Val-318, Ala-354, and Ile-378. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of
this article.)
nitrogen of the imidazole ring is able to form a hydrogen bond with
the backbone amide hydrogen of Val- 181 whereas the N-methyl
group of the imidazole ring favorably interacts with the methylene
of Lys- 352. The calculated tubulin binding energy of 5c
favorably interacts with the methylene of Lys-b352.
a
The existence of two energetically close binding orientations of
5c might even go some way towards explaining its weak interfer-
ence with tubulin in vitro and its reduced toxicity to non-malignant
cells by assuming a metastable conformation of tubulin ligated
with 5c with retarded polymerization dynamics.
Immunofluorescence microscopic analyses of the microtubule-
disrupting effects of 5a and 5c in 518A2 melanoma cells revealed
a similar efficacy of both derivatives (Fig. 3). Incubation with
100 nM of 5c for 24 h led to an extensive destruction of the
microtubular network while residual microtubular structures
remained after incubation with 5a. These differences in the effects
of 5a and 5c on cellular microtubules mirror their different IC50
(24 h) values against this cell line in the MTT tests (Table 1) and
might be due to different uptake rates.
b
(Ebind ¼ ꢀ8.8 kcal/mol) is comparable with that of colchicine
(Ebind ¼ ꢀ9.0 kcal/mol).
In the second, energetically slightly less favourable orientation
(Fig. 2b, right), the phenyl ring of 5c is placed in the hydrophobic
pocket, like the A-rings of colchicine and CA-4 [25]. The p-methoxy
group of the phenyl ring can link up to Ile-
interactions, while the p-methoxy oxygen is engaged in a hydrogen
bond with the sulfhydryl side-chain of Cys- 241. The m-methoxy
group of the phenyl ring can interact with a methylene group of
Lys- 254. The bromo substituent is placed in the hydrophobic
pocket made up by Ala- 316, Val- 318, Ala- 354, Ile- 378. The
chloro residue of the indole is oriented towards the amino side-
chain of Lys- 254 and the side-chain amine of Asn- 101. The N-
methyl group of the imidazole can connect via Van-der-Waals in-
teractions to the Val- 181 side-chain. In addition, the indole ring
b378 by Van-der-Waals
b
b
b
b
b
b
b
a
2.2.4. Antivascular effects
Since microtubule disruption is a major aspect of the anti-
vascular activity of CA-4 and its prodrug CA-4-phosphate [26,27],
b

K. Mahal et al. / European Journal of Medicinal Chemistry 118 (2016) 9e20
13
Fig. 3. Fluorescence microscopic visualization of microtubules (anti-alpha-tubulin) and nuclei (DAPI) in 518A2 melanoma cells after 24 h incubation with DMSO (control) or 100 nM
of 5a or 5c. 400-fold magnification, scale bar: 100 m.
m
we performed in vitro tube formation assays with Ea.hy926 endo-
thelial cells to qualitatively assess any vascular-disrupting effects of
the best performing indoles 5a and 5c. When applied at 250 nM
concentrations, both indoles disrupted preformed vessel-like
tubular networks of these cells to an extent comparable with the
effect of 100 nM CA-4 (Fig. 4). The compounds induced a retraction
of cells to the nodes, a change of their morphology from stretched
to rounded, and a loss of the cellecell connections that build up the
blood vessel-like mesh network.
fertilized chicken eggs (Fig. 6). Within a couple of days these
microtumors grew into the CAM and recruited new blood vessels in
their vicinity, either by secreting growth factors or via inflamma-
tion as a consequence of grafting. They can also interact with the
chicken CAM tissue in a way reminiscent of mouse xenograft
models [28]. 24 h after treatment with 5c a partial destruction of
surrounding blood vessels and a faint hemorrhage within the HT-
29 microtumor was observed. HE-staining of untreated CAM-
tumor cross sections (Fig. 6, right column) revealed some blood
vessels surrounded by HT-29 tumor cells at the interface between
the tumor pellet and the CAM or even blood vessels that had
already grown into the developing tumor graft. Intact blood vessels
were not found in CAM-tumor cross sections treated with 5c, and a
great number of erythrocytes corroborated the macroscopically
observed hemorrhage. In addition, a retraction of individual tumor
cells within the pellet and a loss of cellecell interactions as a
consequence of severe cytoskeletal damage or even induction of
apoptosis was observed. This remarkable vascular-disrupting ac-
tivity occurs at drug concentrations not compromising the em-
bryo's vitality.
This effect was also investigated in vivo by using the chorioal-
lantoic membrane (CAM) of fertilized chicken eggs as a model
system for rapidly developing blood vessels (Fig. 5).
Interestingly, 5 nmol of the trimethoxy derivative 5a were
strongly toxic to the chicken embryos. After 12 h incubation,
existing blood vessels had been almost completely disrupted and a
pronounced hemorrhage was observed on the CAM suggesting that
chicken embryos also suffered from internal hemorrhage. The
bromo derivative 5c was much better tolerated while retaining a
strong vascular-disrupting effect even when only 2.5 nmol were
applied. The lethality of chicken embryos was significantly reduced
(Table S2, Supplementary data). Though existing blood vessels were
destroyed and hemorrhage of blood vessel outgrowths on the CAM
could be observed, the blood vessel system was able to regenerate
within the following 24e48 h and embryos apparently developed
normally (data not shown).
2.2.5. Animal studies
We also investigated the vascular-disrupting activity of 5c in
highly vascularized xenograft tumors of the 1411HP germ cell tu-
mor cell line in mice which is an established animal model, pre-
viously used for testing other vascular-disrupting agents [18,21]. As
a second model we used the ovarian carcinoma cell line A2780cis
We also tested indole 5c on microtumor-xenografts of HT-29
colon carcinoma cell pellets in matrigel grafted on the CAM of

14
K. Mahal et al. / European Journal of Medicinal Chemistry 118 (2016) 9e20
Fig. 4. Effects of CA-4 and of the indoles 5a and 5c on the propensity of Ea.hy926 endothelial cells to form blood vessel-like networks when grown on matrigel layers. Cells grown
for 12 h and incubated with CA-4 (100 nM), 5a or 5c (250 nM) for additional 16 h were documented by light microscopy (100-fold magnification, scale bar: 200 m).
m
which also readily forms highly vascularized xenograft tumors.
Both cell lines are resistant to various anticancer drugs. A single
treatment of the xenograft bearing mice with 100 mg/kg body
weight of 5c induced a distinct discoloration in both tumor types
within 24 h due to intratumoral hemorrhage (Fig. 7A). Histological
examination of tumors of treated mice revealed typical features of
vascular-disrupting agents, such as disrupted vessel structures,
hemorrhage and extensive necrosis (Fig. 7BeE).
To prove the anticancer efficacy of 5c in resistant tumors we
finally undertook preliminary therapy trials in mice bearing
A2780cis tumors. Treatment with 5c resulted in a distinct tumor
growth retardation relative to controls (Fig. 8). Although this has to
be confirmed in a larger trial these findings recommend 5c as a
promising drug candidate for the treatment of resistant tumors.
compounds 5. The latter gave rise to a stronger mitotic arrest in
cancer cells and a more pronounced disruption of their microtu-
bular cytoskeleton, although they did not interfere with purified
tublin in vitro as much as CA-4. The vascular-disrupting effects of
the most active compound 5c in vitro and in vivo were comparable
to or better than those of the lead compound CA-4. It also showed a
better tolerance in non-malignant cells and animals and a distinct
tumor growth inhibiting effect in preliminary animal studies. The
fact that minute alterations of substituents at the periphery of the
indole ring, e.g., when going from R2 ¼ H in 5c to R2 ¼ Cl in 5h,
made such a big difference in activity, leaves ample room for
further drug optimisation.
4. Experimental section
3. Conclusions
4.1. Chemistry
The new imidazolylindoles 5 are conjugates of two pharmaco-
phores of vascular-disrupting agents, borrowed from the natural
combretastatins A and from known synthetic indoles. The intention
was to overcome the shortcomings of CA-4, such as insufficient
cytotoxicity, cancer selectivity, stability, and tolerance. We could
indeed identify some derivatives of general structure 5 that lived up
to this objective. All derivatives 5a-d featuring an N-methyl
substituted indole and a halo-dimethoxy or trimethoxy substituted
A-ring were highly active against all tested tumor cell lines,
including multidrug-resistant and CA-4-resistant ones. This is in
keeping with tubulin docking studies which pinpointed crucial
binding interactions originating from these residues. There are
subtle differences in the modes of action of CA-4 and active
4.1.1. General
IR spectra: FT-IR spectrophotometer equipped with an ATR unit.
1H NMR and 13C NMR spectra: Bruker DRX 500 and DRX 300
spectrometer; chemical shifts are given in parts per million using
the residual solvent peak as an internal standard; coupling con-
stants (J) are quoted in Hz. Mass spectra: Thermo Finnigan MAT
8500 spectrometer under EI (70 eV) conditions using a MAT SS 300
data system. Microanalyses indicated by the symbols of the ele-
ments were within 0.4% of the theoretical values.
4.1.2. 3-Chloro-1-methyl-5-[1-methyl-4-(3-chloro-4,5-
dimethoxyphenyl)-imidazol-5-yl]indole hydrochloride (5b)
N-Methyl-3-chloroindole-5-carboxaldehyde
3a
(81
mg,

K. Mahal et al. / European Journal of Medicinal Chemistry 118 (2016) 9e20
15
Fig. 5. Effects of the indoles 5a and 5c on developing blood vessels in the chorioallantoic membrane (CAM) of fertilized chicken eggs (6 d after fertilization). Dilutions of the
compounds (2.5 nmol or 5 nmol in PBS) were directly applied within rings of silicon foil (6 mm diameter) and alterations 12 h post application were documented with a ste-
reomicroscope (60-fold magnification).
0.42 mmol), 33% MeNH2/ethanol (260
m
L, 2.10 mmol) and acetic
3 h. The solvent was evaporated, the residue was suspended in
water and extracted with ethyl acetate. The organic phase was
dried over Na2SO4, concentrated in vacuum and purified by column
chromatography to give the free base of 5b (colorless oil; Rf ¼ 0.62,
acid (150 L) were refluxed in ethanol (15 mL) for 2 h, then TosMIC
m
reagent 4b (153 mg, 0.42 mmol) and K2CO3 (500 mg, 3.62 mmol)
were added and the reaction mixture was stirred under reflux for

16
K. Mahal et al. / European Journal of Medicinal Chemistry 118 (2016) 9e20
Fig. 6. Effects of indole 5c on developing blood vessels and HT-29 micro-xenografts 24 h post application on the chorioallantoic membrane (CAM) of fertilized chicken eggs. HT-29
tumor pellets grafted on the CAM for 48 h and then treated with the control DMSO (top row) or 5 nmol of 5c in PBS (application outside the silicon ring; bottom row) for 24 h,
documentation by stereomicroscopy (10-fold magnification, scale bar: 6 mm). Right column: HE-staining of HT-29 micro-xenograft cross sections, arrows indicate intact blood
vessels surrounded by HT-29 cells, documentation by light microscopy (320-fold magnification, scale bar: 100 mm). (For interpretation of the references to color in this figure legend,
the reader is referred to the web version of this article.)
ethyl acetate/methanol 95:5). The free base was dissolved in CH2Cl2
(5 mL) and treated with 3 M HCl/dioxane (1 mL). After stirring for
10 min the solvent was evaporated and the residue was recrystal-
lized from CH2Cl2/n-hexane. Yield: 98 mg (0.20 mmol, 48%);
colorless solid; mp 198 ꢁC; UV (MeOH) lmax (ε) 230 (30,800); nmax
(ATR)/cmꢀ1: 3392, 2941, 2845, 2602, 1622, 1565, 1542, 1493, 1464,
1456, 1318, 1239, 1115, 1046, 996, 971, 871, 852, 803; 1H NMR
4.1.4. 3-Chloro-1-methyl-5-[1-methyl-4-(3-iodo-4,5-
dimethoxyphenyl)-imidazol-5-yl]indole hydrochloride (5d)
Analogously to the synthesis of 5b, compound 5d (110 mg, 48%)
was obtained from 3a (81 mg, 0.42 mmol), 33% MeNH2/ethanol
(260
mL, 2.10 mmol), acetic acid (150 mL), and 4d (192 mg,
0.42 mmol); colorless solid of mp 160e165 ꢁC (ethanol/n-hexane);
nmax (ATR)/cmꢀ1: 3019, 2932, 2572, 1623, 1589, 1544, 1486, 1464,
1406, 1360, 1314, 1260, 1242, 1146, 1115, 1039, 995, 973, 872, 854,
(300 MHz, DMSO-d6): d 3.61 (6 H, s), 3.72 (3 H, s), 3.86 (3 H, s), 6.99
(1 H, d, 4J 2.1 Hz), 7.20 (1 H, d, 4J 2.1 Hz), 7.34 (1 H, dd, 3J 9.4 Hz, 4J
1.6 Hz), 7.7e7.8 (3 H, m), 9.32 (1 H, s); 13C NMR (75.5 MHz, DMSO-
807, 796, 751, 726, 670; 1H NMR (300 MHz, DMSO-d6):
d 3.57 (3 H,
s), 3.62 (3 H, s), 3.67 (3 H, s), 3.87 (3 H, s), 7.20 (1 H, d, 4J 2.0 Hz),
d6):
d 33.0, 33.9, 56.0, 60.3, 102.7, 110.8, 111.8, 117.0, 119.5, 120.4,
7.3e7.4 (2 H, m), 7.7e7.8 (3 H, m), 9.34 (1 H, s); 13C NMR (75.5 MHz,
124.1, 124.5, 125.1, 127.2, 127.9, 128.3, 130.9, 135.6, 136.0, 144.8,
153.4; m/z (EI) 417 (100) [Mþ], 415 (68) [Mþ], 402 (66), 400 (96).
Anal C21H19Cl2N3O2 (free base; C, H, N).
DMSO-d6): d 33.0, 33.9, 55.8, 59.9, 93.0, 102.7, 111.7, 112.1, 117.0,
120.4, 124.5, 125.1, 125.4, 127.6, 128.1, 128.2, 130.8, 135.5, 136.0,
148.5, 152.0; m/z (EI) 509 (33) [Mþ], 507 (100) [Mþ], 492 (41); Anal
C21H19ClIN3O2 (free base; C, H, N).
4.1.3. 3-Chloro-1-methyl-5-[1-methyl-4-(3-bromo-4,5-
4.1.5. 3-Chloro-1-ethyl-5-[1-methyl-4-(3,4,5-trimethoxyphenyl)-
imidazol-5-yl]indole hydrochloride (5e)
Analogously to the synthesis of 5b, compound 5e (104 mg, 54%)
was obtained from 3b (88 mg, 0.42 mmol), 33% MeNH2/ethanol
dimethoxyphenyl)-imidazol-5-yl]indole hydrochloride (5c)
Analogously to the synthesis of 5b, compound 5c (115 mg, 51%)
was obtained from 3a (81 mg, 0.42 mmol), 33% MeNH2/ethanol
(260
m
L, 2.10 mmol), acetic acid (150
m
L), and 4c (170 mg,
(260 mL, 2.10 mmol), acetic acid (150 mL), and 4a (150 mg,
0.42 mmol); colorless solid of mp 146 ꢁC (ethanol/n-hexane); UV
(MeOH) lmax (ε) 230 (34,040); nmax (ATR)/cmꢀ1: 3392, 2941, 2591,
1622, 1592, 1544, 1490, 1464, 1406, 1317, 1263, 1239, 1146, 1115,
0.42 mmol); colorless solid of mp 245 ꢁC (ethanol/n-hexane); nmax
(ATR)/cmꢀ1: 3984, 2948, 2530, 1617, 1597, 1553, 1519, 1494, 1462,
1445, 1417, 1405, 1337, 1321, 1286, 1261, 1220, 1196, 1161, 1120, 1030,
991, 875, 852, 828, 812, 798, 771, 747, 713; 1H NMR (300 MHz,
1042, 995, 973, 874, 853, 809; 1H NMR (300 MHz, DMSO-d6):
d 3.61
(6 H, s), 3.70 (3 H, s), 3.86 (3 H, s), 7.14 (1 H, d, 4J 2.1 Hz), 7.26 (1 H, d,
4J 2.1 Hz), 7.34 (1 H, dd, 3J 8.5 Hz, 4J 1.6 Hz), 7.7e7.8 (3 H, m), 9.36
DMSO-d6): d
1.36 (3 H, t, 3J 7.1 Hz), 3.47 (6 H, s), 3.61 (3 H, s), 3.65
(3 H, s), 4.28 (2 H, q, 3J 7.1 Hz), 6.81 (2 H, s), 7.35 (1 H, dd, 3J 8.5 Hz, 4J
(1 H, s); 13C NMR (75.5 MHz, DMSO-d6):
d 33.0, 33.9, 56.0, 60.2,
1.6 Hz), 7.73 (1 H, s), 7.8e7.9 (2 H, m), 9.41 (1 H, s); 13C NMR
102.7, 111.4, 111.8, 116.9, 117.0, 120.4, 122.3, 124.5, 124.7, 125.1, 127.7,
128.3, 130.9, 135.5, 136.0, 145.9, 153.2; m/z (EI) 461 (100) [Mþ], 459
(82) [Mþ], 446 (59), 444 (57); Anal C21H19BrCl2N3O2 (free base; C, H,
N).
(75.5 MHz, DMSO-d6): d 15.4, 33.9, 40.9, 55.6, 60.0, 102.9, 104.4,
111.7, 117.3, 120.7, 122.4, 124.6, 125.3, 126.7, 129.0, 130.3, 135.0, 137.7,
152.8; m/z (EI) 427 (35) [Mþ], 425 (100) [Mþ], 412 (28), 410 (92);
Anal C23H24ClN3O3 (free base; C, H, N).

K. Mahal et al. / European Journal of Medicinal Chemistry 118 (2016) 9e20
17
Fig. 7. Vascular-disrupting effects of indole 5c in nude mouse xenograft tumors. A: Discoloration of 1411HP and A2780cis xenograft tumors due to intratumoral hemorrhages 24 h
after treatment. B: Lateral section of the treated 1411HP tumor shown in A after HE staining featuring hemorrhages (H) and a large necrotic core area (N) surrounded by a cortical
layer of vital tumor cells (T). C: Lateral section of the treated A2780cis tumor shown in A after HE staining featuring hemorrhages (H) and a core area characterized by early signs of
massive necrotic cell death. D: Normal blood vessel structure of an untreated A2780cis tumor showing erythrocytes surrounded by an intact endothelial cell layer (black arrows). E:
Magnified disrupted blood vessel, framed in C, showing agglutinated erythrocytes surrounded by residual cell fragments of the disrupted endothelial cell layer (black arrows) and
dying tumor cells. (B and C: 100-fold magnification, D and E: 400-fold magnification, zoomed).
4.1.6. 3-Chloro-1-ethyl-5-[1-methyl-4-(3-bromo-4,5-
(75.5 MHz, DMSO-d6): d 15.4, 33.9, 40.9, 56.0, 60.2, 103.0, 111.4,
dimethoxyphenyl)-imidazol-5-yl]indole hydrochloride (5f)
Analogously to the synthesis of 5b, compound 5f (110 mg, 52%)
was obtained from 3b (88 mg, 0.42 mmol), 33% MeNH2/ethanol
111.8,116.8,117.0,120.5,122.3,124.5,124.6,125.3,126.8,127.6,130.9,
135.1, 135.4, 145.8, 153.2; m/z (EI) 475 (100) [Mþ], 473 (77) [Mþ],
460 (54), 458 (45); Anal C22H21BrClN3O2 (free base; C, H, N).
(260
mL, 2.10 mmol), acetic acid (150 mL), and 4c (172 mg,
0.42 mmol); colorless solid of mp 240 ꢁC (ethanol/n-hexane); nmax
(ATR)/cmꢀ1: 2946, 2936, 2607, 1623, 1595, 1546, 1490, 1464, 1448,
1407, 1347, 1316, 1262, 1219, 1146, 1115, 1042, 995, 874, 853, 795,
4.1.7. 3-Chloro-1-ethyl-5-[1-methyl-4-(3-iodo-4,5-
dimethoxyphenyl)-imidazol-5-yl]indole hydrochloride (5g)
Analogously to the synthesis of 5b, compound 5g (114 mg, 50%)
was obtained from 3b (88 mg, 0.42 mmol), 33% MeNH2/ethanol
753, 726, 676; 1H NMR (300 MHz, DMSO-d6):
d
1.37 (3 H, t, 3J
7.2 Hz), 3.59 (3 H, s), 3.63 (3 H, s), 3.69 (3 H, s), 4.29 (2 H, q, 3J 7.2 Hz),
7.16 (1 H, d, 4J 2.1 Hz), 7.25 (1 H, d, 4J 2.1 Hz), 7.33 (1 H, dd, 3J 8.5 Hz,
4J 1.6 Hz), 7.71 (1 H, s), 7.8e7.9 (2 H, m), 9.41 (1 H, s); 13C NMR
(260
mL, 2.10 mmol), acetic acid (150 mL), and 4d (192 mg,
0.42 mmol); colorless solid of mp 235 ꢁC (ethanol/n-hexane); nmax
(ATR)/cmꢀ1: 3003, 2968, 2932, 2608, 1622, 1590, 1544, 1486, 1461,

18
K. Mahal et al. / European Journal of Medicinal Chemistry 118 (2016) 9e20
1500
1400
1300
1200
1100
1000
900
800
700
600
500
400
300
200
100
0
DMSO-d6):
d 15.3, 34.4, 40.9, 56.4, 60.7, 101.8, 111.8, 112.3, 117.3,
NaCl
5c
118.8, 120.5, 122.9, 123.9, 124.5, 125.2, 125.8, 128.5, 131.0, 134.3,
136.1, 146.3, 153.7; m/z (EI) 511 (44) [Mþ], 509 (100) [Mþ], 507 (60)
[Mþ], 496 (22), 494 (52), 492 (27); Anal C22H20BrCl2N3O2 (free base;
C, H, N).
4.2. Cell lines and culture conditions
The human carcinoma cell lines HT-29 (colon), HCT-116 (colon),
MCF-7 (breast), and KB-V1 (cervix) were purchased from The
German Collection of Microorganisms and Cell Cultures (DSMZ,
Braunschweig). MCF-7 and KB-V1 cells were rendered multidrug-
resistant, indicated as MCF-7/Topo and KB-V1/Vbl, by repeated
application of topotecan or vinblastin, respectively. The HUVEC-
derived endothelial hybrid cell line Ea.hy926 was obtained from
The American Type Culture Collection (ATCC no. CRL-2922). The
human melanoma cell line 518A2 was a gift from the Department of
Radiotherapy and Radiobiology, University Hospital Vienna. It is not
available from cell banks, yet easily identified by its large size and
its flat, spread-out morphology. The cells were grown in DMEM or
RPMI (HT-29) medium, supplemented with 10% fetal bovine serum
(FBS), 1% Antibiotic-Antimycotic solution (both from Gibco) and
0
2
4
6
8
10
12
14
16
days after treatment
Fig. 8. Antitumor activity of 5c in resistant A2780cis mouse xenograft tumors. Shown
is the increase of mean tumor volumes of each group (n ¼ 4) standard deviation
normalized to day 0 (start of treatment). Mean tumor volumes at start of treatment
were 218 mm3 for the group treated i.p. with 5c (40 mg/kg body weight on days 0, 3, 7,
10 and 14) and 204 mm3 for the control group treated with saline. The tumors of the
control group showed heterogeneous growth rates whereas tumors of mice treated
with 5c grew more slowly and more evenly leading to a narrower growth rate dis-
tribution at early stages. The treatment and monitoring of the control group was
discontinued on day 9 when three mice had tumors exceeding 1500 mm3 volume. On
day 16, two tumors of the treatment group had exceeded 1500 mm3.
250 mg/mL gentamycin (SERVA). Primary chicken heart fibroblasts
(CHF) were explanted from 10 day-old chicken embryos and
separated from other cell types for several weeks [29]. The estab-
lished cell line from a single cell cluster of fibroblasts was finally
1408, 1346, 1314, 1260, 1219, 1146, 1115, 1039, 995, 870, 854, 795,
749, 725, 677; 1H NMR (300 MHz, DMSO-d6):
d
1.37 (3 H, t, 3J
7.2 Hz), 3.56 (3 H, s), 3.64 (3 H, s), 3.66 (3 H, s), 4.29 (2 H, q, 3J 7.2 Hz),
7.20 (1 H, d, 4J 2.0 Hz), 7.3e7.4 (2 H, m), 7.71 (1 H, s), 7.8e7.9 (2 H,
grown in DMEM (10% FBS, 1% AntieAnti, 250 mg/mL gentamycin)
and used before the 20th passage. Chicken cardiomyocytes (CCM)
were separated by trypsination of single pulsating cell clusters
[29e31], grown to confluence in DMEM (20% FBS, 1% AntieAnti,
m), 9.38 (1 H, s); 13C NMR (75.5 MHz, DMSO-d6):
d 15.4, 33.9, 40.9,
55.7, 59.9, 93.0, 102.9, 111.7, 112.1, 117.0, 120.5, 124.5, 125.2, 125.3,
126.7, 127.5, 128.1, 130.8, 135.1, 135.4, 148.5, 152.0; m/z (EI) 523 (33)
[Mþ], 521 (100) [Mþ], 506 (42); Anal C22H21ClIN3O2 (free base; C, H,
N).
250 mg/mL gentamycin) and used before the 3rd passage. All cells
were incubated at 37 ꢁC, 5% CO2, 95% humidified atmosphere. Only
mycoplasma-free cell lines were used.
4.3. Growth inhibition assays (MTT assay)
4.1.8. 2,3-Dichloro-1-methyl-5-[1-methyl-4-(3-bromo-4,5-
dimethoxyphenyl)-imidazol-5-yl]indole hydrochloride (5h)
Analogously to the synthesis of 5b, compound 5h (90 mg, 41%)
was obtained from 3c (96 mg, 0.42 mmol), 33% MeNH2/ethanol
Inhibition of cell proliferation by CA-4 or the new derivatives
was determined by using a standard MTT assay as previously
described [32]. Briefly, cells grown in 96-well cell culture plates
(5000 cells/well; Ea.hy926 and CHF: 10,000 cells/well, CCM:
20,000 cells/well) were treated with dilution series of CA-4, the
(260
mL, 2.10 mmol), acetic acid (150 mL), and 4c (172 mg,
0.42 mmol); colorless solid of mp 152e155 ꢁC (ethanol/n-hexane);
nmax (ATR)/cmꢀ1: 3003, 2937, 2837, 2590, 1623, 1595, 1547, 1490,
1464, 1405, 1361, 1318, 1263, 1238, 1152, 1116, 1042, 996, 976, 878,
derivatives 5a-i (100 mMe5 pM) or respective amounts of DMSO
(solvent controls) for 72 h. Cells were then incubated with MTT at a
final concentration of 0.5 mg/mL for 2 h. The formazan precipitate
was dissolved in 10% SDS, 0.6% acetic acid in DMSO and the
absorbance at wavelength 570 and 630 (background) nm was
measured with a microplate reader (Tecan). The percentage of
viable cells was calculated with respect to DMSO-treated controls
set to 100%. Half-maximal inhibitory concentrations (IC50) are
represented as the mean of at least three experiments standard
deviation (S.D.). The relative selectivity of the compounds was
represented as the relation of the logarithmic mean of IC50 values
against cancer cells (CA-4-resistant HT-29 were excluded from this
calculation for CA-4) and the mean (log) of IC50 values in non-
malignant fibroblasts.
853, 808, 758, 730, 680; 1H NMR (300 MHz, DMSO-d6):
d 3.61 (3 H,
s), 3.64 (3 H, s), 3.70 (3 H, s), 3.86 (3 H, m), 7.13 (1 H, d, 4J 2.0 Hz),
7.28 (1 H, d, 4J 2.0 Hz), 7.40 (1 H, dd, 3J 8.6 Hz, 4J 1.2 Hz), 7.73 (1 H, d,
4J 1.2 Hz), 7.84 (1 H, d, 3J 8.6 Hz), 9.37 (1 H, s); 13C NMR (75.5 MHz,
DMSO-d6):
d 30.9, 33.9, 56.1, 60.2, 101.0, 111.5, 112.0, 116.8, 118.2,
119.9, 122.4, 123.8, 124.4, 124.6, 125.6, 127.8, 130.5, 134.9, 135.6,
145.9, 153.3; m/z (EI) 497 (51) [Mþ], 495 (100) [Mþ], 493 (66) [Mþ],
488 (24), 486 (53), 484 (33); Anal C21H18BrCl2N3O2 (free base; C, H,
N).
4.1.9. 2,3-Dichloro-1-ethyl-5-[1-methyl-4-(3-bromo-4,5-
dimethoxyphenyl)-imidazol-5-yl]indole hydrochloride (5i)
Analogously to the synthesis of 5b, compound 5i (94 mg, 43%)
was obtained from 3d (102 mg, 0.42 mmol), 33% MeNH2/ethanol
4.4. Cell cycle analyses and immunostaining of mitotic cells
(260
mL, 2.10 mmol), acetic acid (150 mL), and 4c (172 mg,
0.42 mmol); colorless solid of mp 226e230 ꢁC (ethanol/n-hexane);
nmax (ATR)/cmꢀ1: 2980, 2562, 1623, 1609, 1593, 1546, 1509, 1490,
1468, 1452, 1404, 1345, 1319, 1280, 1238, 1218, 1150, 1114, 1042, 996,
For cell cycle analyses, HT-29 cells grown in 6-well cell culture
plates (200,000 cells/well) were incubated with DMSO (control) or
100 nM of the indoles 5a or 5c for 24 h. Due to their resistance to
852, 807, 758; 1H NMR (300 MHz, DMSO-d6):
d
1.29 (3 H, t, 3J
CA-4, a high concentration of 5 mM was used to get a typical effect
7.1 Hz), 3.58 (3 H, s), 3.63 (3 H, s), 3.70 (3 H, s), 4.39 (2 H, q, 3J 7.1 Hz),
7.1e7.2 (2 H, m), 7.39 (1 H, dd, 3J 8.6 Hz, 4J 1.3 Hz), 7.74 (1 H, d, 4J
1.3 Hz), 7.88 (1 H, d, 3J 8.6 Hz), 9.33 (1 H, s); 13C NMR (75.5 MHz,
on the cell cycle progression. Cells were then harvested by trypsi-
nation, resuspended in 4% formaldehyde (in PBS) and fixed at room
temperature for 20 min. For permeabilization, cells were incubated

K. Mahal et al. / European Journal of Medicinal Chemistry 118 (2016) 9e20
19
with 90% ice-cold methanol for 30 min on ice prior to immuno-
staining (37 ꢁC, 1 h) with a primary phospho-histone H3 antibody
as a marker for mitosis (anti-phospho-histone H3 (Ser10), rabbit
(25,000 cells/well) were incubated with 100 nM 5a or 5c for 24 h.
After fixation (4% formaldehyde in PBS, 20 min, rt), permeabiliza-
tion and blocking (0.1% Triton X-100, 1% BSA in PBS) of the cells, a
primary antibody for alpha-tubulin (anti-alpha-tubulin, mouse,
monoclonal antibody, Invitrogen/Life Technologies) was added in
1% BSA/PBS at 37 ꢁC for 1 h and microtubules were visualized with a
secondary goat anti-mouse-IgG-AlexaFluor-488 conjugate (Invi-
trogen/Life Technologies). Coverslips were then mounted into
polyclonal antibody, Millipore; 5 mg/mL in 1% BSA/PBS) and a sec-
ondary antibody-AlexaFluor-488 conjugate (goat anti-rabbit-IgG-
AlexaFluor-488 conjugate, Invitrogen). Cells were washed by
centrifugation and stained with propidium iodide staining solution
(50 mg/mL propidium iodide, 0.1% sodium citrate, 50 mg/mL RNase A
in PBS) for 30 min at 37 ꢁC. The fluorescence intensity of 10,000
single cells at an emission wavelength of 620 nm (propidium iodide
fluorescence) and 525 nm (AlexaFluor-488 fluorescence; excitation
with a 488 nm laser source) was recorded with a Beckman Coulter
Cytomics FC500 flow cytometer and analyzed for the distribution
(%) of single, viable (sub-G1 events < 3.5%) cells to G1, S, G2 and M
phase of the cell cycle by using the CXP software (Beckman
Coulter).
Mowiol 4-88-based mounting medium containing 1 mg/mL DAPI
(both from Carl Roth) for counterstaining the nuclei. Immunofluo-
rescence images were taken by using an Axio Imager.A1 fluores-
cence microscope (400-fold magnification, ZEISS).
4.8. Tube formation assay with Ea.hy926 endothelial cells
Ea.hy926 endothelial hybrid cells (100,000 cells/well) were
seeded into 24-well plates onto glass coverslips coated with 10 mL
4.5. Tubulin polymerization assay
of Matrigel™ basement membrane matrix (BD Biosciences; gela-
tion for 30 min in a humidified atmosphere at 37 ꢁC). The cells were
allowed to differentiate into blood vessel precursor-like networks
through growth factor stimulation within the next 12 h. Then, CA-4
(100 nM) or the indoles 5a or 5c (250 nM) were added to the wells
for a further 16 h and a vascular-disruptive effect on the established
cell network was documented by microscopy (Axiovert 135, ZEISS;
100-fold magnification). The viability of treated cells was addi-
Alpha- and beta-tubulin heterodimers were purified from
porcine brain by the method of Castoldi and Popov [33]. To start the
polymerization to microtubules, equal volumes of 10 mg/mL
tubulin stock solutions in BRB80 (Brinkley's buffer, 80 mM PIPES-
KOH, 1 mM MgCl2, 1 mM EGTA) and 2 ꢂ polymerization buffer
(20% glycerol, 3 mM GTP in BRB80) containing CA-4, 5a, 5c (10 mM
pre-dilution in BRB80) or equal amounts of DMSO (control) were
tionally assessed by 2 h incubation with 200 mL 0.5 mg/mL MTT and
mixed in pre-warmed, black 96-well plates (
m
Clear® black 96-well
subsequent determination of the formazan precipitate (cf. MTT
assay description) by absorbance measurements. Viability in
treated wells was calculated with respect to DMSO controls and
was determined to be >90% when pictures were taken.
plates, UV-star, Greiner bio-one). The polymerization was moni-
tored at 37 ꢁC by measuring the increase in the optical density at
340 nm every 20 s with a microplate reader (Tecan). Obtained
polymerization curves were then normalized with respect to the
starting absorbance values and are represented as the mean of two
independent measurements in duplicates (S.D. of each time point
<5%).
4.9. Chorioallantoic membrane (CAM) assay with fertilized chicken
eggs [39]
White leg hen eggs (SPF eggs, VALO Biomedia) were incubated
until day 5 after fertilization at 37 ꢁC and 50e60% humidity. A
window of 2e3 cm diameter was cut into the pole end of the
eggshell and rings of silicon foil (6 mm diameter) were placed on
the developing vessels within the CAM. Eggs were then sealed with
sterile tape and incubated for a further 24 h. Dilutions of DMSO
4.6. Molecular docking into the tubulin colchicine-binding site
Coordinate files of the ligand structures were generated using
the GlycoBioChem PRODRG2 Server (http://davapc1.bioch.dundee.
ac.uk/prodrg/submit.html) [34]. Molecular docking calculations
were carried out using the Autodock Vina software [35]. For this,
Gasteiger partial charges [36] were calculated on ligand atoms
using Autodock Tools. The X-ray structure of the crystallized
tubulin-colchicine complex (PDB accession code: 1SA0 [37]) was
downloaded from the Protein Data Bank (http://www.rcsb.org) and
later used for docking. As for the ligand, polar hydrogen atoms were
added to the protein and Gasteiger partial charges were calculated
using Autodock Tools. Water molecules, heteroatoms and ligands
were removed from the structure prior to docking calculations.
(control), 5a or 5c (500
silicon ring (20
mM in PBS) were directly pipetted inside the
m
L ¼ 10 nmol, 10 L ¼ 5 nmol) and alterations in the
m
blood vessel organization were documented 0 h, 12 h and 24 h post
application with a stereomicroscope (Traveller, 60-fold magnifica-
tion). For assessing effects on in vivo tumor growth with the CAM
assay model, pellets of 1 ꢂ 106 HT-29 cells in 10
mL 1:1-mixture of
matrigel and serum-free RPMI (gelled at 37 ꢁC for 30 min) were
directly placed on the CAM and incubated for 48 h until an inter-
action between the tumor pellets and the surrounding blood ves-
Residues Lys-
b
254, Lys-
b
352, Asn-
a
101, Val-
b
318 and Ile-b378 were
sels was observed. 100 mL of a 50 mM dilution in PBS (5 nmol) of
treated as flexible residues. Simulation boxes were centered on the
originally crystallized ligand colchicine. A 18 ꢂ 22 ꢂ 20 Å simula-
tion box was used in the docking calculations, using an exhaus-
tiveness option of 20. It should be noted that the standard error of
Autodock Vina is 2.85 kcal molꢀ1 [35], which means that binding
affinities cannot be quantitatively predicted in silico using docking
calculations. In addition, the estimated coordinate error of the
tubulin template structure, which was solved to a moderate reso-
lution of 3.58 Å, already amounts to ~0.5 Å. Figures were prepared
by means of the program PYMOL [38].
indole 5c were then pipetted on the CAM outside of a silicon ring
around the tumor pellet, incubated for 24 h and the changes were
documented with a stereomicroscope (Traveller, 10-fold magnifi-
cation). For histological examination, the micro-xenografts were
explanted, fixed in 4% formalin and embedded in paraffin. Hema-
toxylin/eosin (HE) staining of tissue slices was performed according
to standard protocols. Images shown in this paper are representa-
tive for at least three independent experiments with the same
outcome.
4.10. Animal studies
4.7. Immunofluorescence microscopy of the microtubule
cytoskeleton
The investigations of this study were approved by the Labora-
tory Animal Care Committee of Sachsen-Anhalt, Germany. Xeno-
graft tumors were generated in athymic nude mice (Harlan and
518A2 cells grown on glass coverslips in 24-well plates

20
K. Mahal et al. / European Journal of Medicinal Chemistry 118 (2016) 9e20
[8] L. Wang, K.W. Woods, Q. Li, K.J. Barr, R.W. McCroskey, S.M. Hannick, L. Gherke,
R.B. Credo, Y.H. Hui, K. Marsh, R. Warner, J.Y. Lee, N. Zielinski-Mozng, D. Frost,
S.H. Rosenberg, H.L. Sham, J. Med. Chem. 45 (2002) 1697e1711.
[9] N.R. Madadi, N.R. Penthala, K. Howk, A. Ketkar, R.L. Eoff, M.J. Borrelli,
P.A. Crooks, Eur. J. Med. Chem. 103 (2015) 123e132.
[10] G.R. Pettit, M.D. Minardi, H.J. Rosenberg, E. Hamel, M.C. Bibby, S.W. Martin,
M.K. Jung, R.K. Pettit, T.J. Cuthbertson, J.-C. Chapius, J. Nat. Prod. 68 (2005)
1450e1458.
[11] S. Ahn, C.B. Duke, C.M. Barrett, D.J. Hwang, C.-M. Li, D.D. Miller, J.T. Dalton,
Mol. Cancer Ther. 9 (2010) 2859e2868.
[12] R. Romagnoli, P.G. Baraldi, T. Sarkar, M.D. Carrion, C.L. Cara, O. Cruz-Lopez,
D. Preti, M.A. Tabrizi, M. Tolomeo, S. Grimaudo, A. Di Cristina, N. Zonta,
J. Balzarini, A. Brancale, H.-P. Hsieh, E. Hamel, J. Med. Chem. 51 (2008)
1464e1468.
Winkelmann, Borchen, Germany) using the germ cell tumor cell
line 1411HP and the ovarian carcinoma cell line A2780cis. 8 Million
cells of either cell line were resuspended in PBS and injected sub-
cutaneously into the flank of mice. Both cell lines form highly
vascularized xenograft tumors in nude mice and are resistant to
various anticancer drugs. Xenografts of 1411HP were previously
used as an established animal model for the test of vascular-
disrupting activity [18,21]. To test compound 5c for tumor
vascular disrupting activity in vivo, 100 mg/kg bodyweight of it
were administered by intraperitoneal injection. The resulting tu-
mor discoloration was documented after 24 h with a Canon IXUS
50. For histological examination, the tumors were explanted, fixed
in 5% formalin, and embedded in paraffin. Hematoxylin/eosin (HE)
staining of the tissue slices was performed according to standard
protocols. HE images were analyzed by microscopy (Axio Lab,
Zeiss).
[13] D. Simoni, F.P. Invidiata, M. Eleopra, P. Marchetti, R. Rondanin, R. Baruchello,
G. Grisolia, A. Tripathi, G.E. Kellogg, D. Durrant, R.M. Lee, Bioorg. Med. Chem.
17 (2009) 512e522.
[14] Q. Zhang, Y. Peng, X.I. Wang, S.M. Keenan, S. Arora, W.J. Welsh, J. Med. Chem.
50 (2007) 749e754.
[15] S. Arora, X.I. Wang, S.M. Keenan, C. Andaya, Q. Zhang, Y. Peng, W.J. Welsh,
Cancer Res. 69 (2009) 1910e1915.
[16] (a) M.T. MacDonough, T.E. Strecker, E. Hamel, J.J. Hall, D.J. Chaplin,
M.L. Trawick, K.G. Pinney, Bioorg. Med. Chem. 21 (2013) 6831e6843;
(b) T.E. Strecker, S.O. Odutola, R. Lopez, M.S. Cooper, J.K. Tidmore,
A.K. Charlton-Sevcik, L. Li, M.T. MacDonough, M.B. Hadimani, A. Ghatak, L. Liu,
D.J. Chaplin, R.P. Mason, K.G. Pinney, M.L. Trawick, Cancer Lett. 369 (2015)
229e241.
[17] Z. Wen, J. Xu, Z. Wang, H. Qi, Q. Xu, Z. Bai, Q. Zhang, K. Bao, Y. Wu, W. Zhang,
Eur. J. Med. Chem. 90 (2015) 184e194.
[18] R. Schobert, B. Biersack, A. Dietrich, K. Effenberger, S. Knauer, T. Mueller,
J. Med. Chem. 53 (2010) 6595e6602.
[19] B. Biersack, Y. Muthukumar, R. Schobert, F. Sasse, Bioorg. Med. Chem. Lett. 21
(2011) 6270e6273.
[20] R. Schobert, K. Effenberger-Neidnicht, B. Biersack, Int. J. Clin. Pharmacol. Ther.
49 (2011) 71e72.
[21] K. Mahal, B. Biersack, H. Caysa, R. Schobert, T. Mueller, Invest. New Drugs 33
(2015) 541e554.
The antitumor activity of 5c was analysed in nude mice bearing
xenograft tumors grown from resistant A2780cis cells. After
establishment of tumors the mice were divided in two groups
(n ¼ 4) with similar mean tumor volumes at the start of treatment
(treatment group: 218 mm3; control group: 204 mm3). The mice
received i.p applications of 40 mg/kg body weight on days 0, 3, 7, 10
and 14. The control group received normal saline. The tumor vol-
umes were calculated by caliper measurement using the formula
a2 ꢂ b ꢂ 0.5 with a being the short and b the long dimension.
Conflict of interest
The authors declare that there are no conflicts of interest.
Acknowledgment
[22] R. Spagnuolo, S. Recalcati, L. Tacchini, G. Cairo, Br. J. Pharmacol. 163 (2011)
299e312.
[23] I. Mikaelian, A. Buness, M.-C. de Vera-Mudry, C. Kanwal, D. Coluccio,
E. Rasmussen, H.W. Char, V. Carvajal, H. Hilton, J. Funk, J.C. Hoflack, M. Fielden,
F. Herting, M. Dunn, L. Suter-Dick, Toxicol. Sci. 117 (2010) 144e151.
[24] C. Kanthou, O. Greco, A. Stratford, I. Cook, R. Knight, O. Benzakour, G. Tozer,
Am. J. Pathol. 165 (2004) 1401e1411.
[25] T.L. Nguyen, C. McGrath, A.R. Hermone, J.C. Burnett, D.W. Zaharevitz,
B.W. Day, P. Wipf, E. Hamel, R. Gussio, J. Med. Chem. 48 (2005) 6107e6116.
[26] C. Kanthou, G.M. Tozer, Expert Opin. Ther. Targets 11 (2007) 1443e1457.
[27] C. Kanthou, G.M. Tozer, Int. J. Exp. Pathol. 90 (2009) 284e294.
[28] M. Klingenberg, J. Becker, S. Eberth, D. Kube, J. Wilting, BMC Cancer 14 (2014)
339.
We are indebted to Franziska Reipsch (Department of Internal
Medicine IV, Halle) for her excellent technical assistance.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.ejmech.2016.04.045.
[29] T. Lindl, J. Bauer, Zell- und Gewebekultur: Einführung in die Grundlagen sowie
€
ausgewahlte Methoden und Anwendungen, G. Fischer, Stuttgart, 1994.
[30] J.W. Sanger, B. Mittal, J.M. Sanger, Cell Motil. 4 (1984) 405e416.
[31] A. Du, J.M. Sanger, K.K. Linask, J.W. Sanger, Dev. Biol. 257 (2003) 382e394.
[32] C. Spoerlein, K. Mahal, H. Schmidt, R. Schobert, J. Inorg. Biochem. 127 (2013)
107e115.
[33] M. Castoldi, A.V. Popov, Protein Expr. Purif. 32 (2003) 83e88.
[34] A.W. Schüttelkopf, D.M.F. van Aalten, Acta Crystallogr. D. Biol. Crystallogr. 60
(2004) 1355e1363.
[35] O. Trott, A.J. Olson, J. Comput. Chem. (2010) 455e461.
[36] J. Gasteiger, M. Marsili, Tetrahedron 36 (1980) 3219e3228.
[37] R.B. Ravelli, B. Gigant, P.A. Curmi, I. Jourdain, S. Lachkar, A. Sobel, M. Knossow,
Nature 428 (2004) 198e202.
[38] W. DeLano, The PyMOL Molecular Graphics System, DeLano Sci. LLC, San
Carlos CA, 2003.
References
[1] G.R. Pettit, S.B. Singh, E. Hamel, C.M. Lin, D.S. Alberts, D. Garcia-Kendall,
Experientia 45 (1989) 209e211.
[2] C.J. Mooney, G. Nagaiah, P. Fu, J.K. Wasman, M.M. Cooney, P.S. Savvides,
J.A. Bokar, A. Dowlati, D. Wang, S.S. Agarwala, S.M. Flick, P.H. Hartman,
J.D. Ortiz, P.N. Lavertu, S.C. Remick, Thyroid 19 (2009) 233e240.
[3] G.C. Tron, T. Pirali, G. Sorba, F. Pagliai, S. Busacca, A.A. Genazzani, J. Med. Chem.
49 (2006) 3033e3044.
[4] J.W. Lipperts, Bioorg. Med. Chem. 15 (2007) 605e615.
[5] S.E. Holwell, P.A. Cooper, M.J. Thompson, G.R. Pettit, L.W. Lippert, S.W. Martin,
M.C. Bibby, Anticancer Res. 22 (2002) 3933e3940.
[6] L.K. Folkes, M. Christlieb, E. Madej, M.R.L. Stratford, P. Wardman, Chem. Res.
Toxicol. 20 (2007) 1885e1894.
[39] B. Nitzsche, C. Gloesenkamp, M. Schrader, M. Ocker, R. Preissner, M. Lein,
A. Zakrzewicz, B. Hoffmann, M. Hoepfner, Br. J. Cancer 103 (2010) 18e28.
[7] T. Brown, H. Holt, M. Lee, Heterocyclic Antitumor Antibiotics, Springer, Berlin,
Heidelberg, 2006.