Endowing indole-based tubulin inhibitors with an anchor for derivatization: Highly potent 3-substituted indolephenstatins and indoleisocombretastatins

Article abstract of DOI:10.1021/jm3015603

Colchicine site ligands with indole B rings are potent tubulin polymerization inhibitors. Structural modifications at the indole 3-position of 1-methyl-5-indolyl-based isocombretastatins (1,1-diarylethenes) and phenstatins endowed them with anchors for fu

Full text of DOI:10.1021/jm3015603

Subscriber access provided by BOSTON UNIV
Article
Endowing Indole-Based Tubulin Inhibitors with an Anchor for Derivatization:
Highly Potent 3-Substituted Indole Phenstatins and Isocombretastatins.
Raquel Álvarez, Mª Pilar Puebla, Jose Fernando Díaz, Ana C. Bento, Rósula García-Navas, Janis
de la Iglesia-Vicente, Faustino Mollinedo, Jose Manuel Andreu, Manuel Medarde, and Rafael Pelaez
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/jm3015603 • Publication Date (Web): 07 Mar 2013
Downloaded from http://pubs.acs.org on March 11, 2013
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 52
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Structures of combretastatin analogues undergoing clinical trials.
1
65x39mm (300 x 300 DPI)
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 52
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5. Effects of compounds 10, 11, 14, 22, 24, 29 and 32 on the microtubule network of HeLa cells.
Cells were incubated in the absence (Control) or in the presence of 10-8 M of compounds 10, 11, 14, 22, 24,
29 and 32 for 18 h and then fixed and processed to study the immunofluorescence of microtubules (red
fluorescence) and nuclei (blue fluorescence), as described in Experimental Section. Bar: 25 µm. The
photomicrographs shown are representative of at least three independent experiments performed.
1
20x127mm (150 x 150 DPI)
ACS Paragon Plus Environment

Page 3 of 52
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. Caspase-3 activation following treatment of HeLa cells with compounds 14, 24, 29 and 32. Cells
were treated with 50 nM of compounds 14, 24, 29 and 32 for the indicated times and analyzed by
immunoblotting with anti-activated caspase-3 (SDS-12% polyacrylamide gels) and anti-PARP antibodies
(SDS-8% polyacrylamide gels). Untreated control cells (C) were run in parallel. The migration positions of
activated caspase-3, as well as of full-length PARP and its cleavage product p85, are indicated. Data shown
are representative of three experiments performed.
1
88x40mm (150 x 150 DPI)
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 52
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
86x125mm (72 x 72 DPI)
ACS Paragon Plus Environment

Page 5 of 52
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 8. Molecular models of carboxylate 17 and amide 29 bound at the colchicine site (upper and middle
rows respectively, two different views) and superimpositions of the former with the electron diffraction
structure of paclitaxel-stabilized tubulin sheets 1JFF.pdb (lower left) and with the tubulin - colchicine
complex 1SA0.pdb (lower right). The carboxylate hydrogen-bonds to the sidechain of Asnβ350 and to a
water molecule and forms a salt bridge with the sidechain ammonium group of Lysβ352. The amide
hydrogen-bonds to the backbone carbonyl of Proβ358. The superimposition with 1JFF.pdb shows that the
ligand occupies the space taken by Lysβ352 in 1JFF.pdb (shown in light green) and displaces the sheets
underneath downwards. The superimposition with the colchicine complex (shown in blue) shows a high
resemblance between the binding modes of the two ligands.
1
84x123mm (72 x 72 DPI)
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 52
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table of Contents Graphic
231x51mm (72 x 72 DPI)
ACS Paragon Plus Environment

Page 7 of 52
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
Endowing IndoleꢀBased Tubulin Inhibitors with an
Anchor for Derivatization: Highly Potent 3ꢀSubstituted
Indole Phenstatins and Isocombretastatins.
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
†
†
‡
#
#
Raquel Álvarez, Pilar Puebla, J. Fernando Díaz, Ana C. Bento, Rósula García-Navas, Janis de la
#
#
‡
†
*,†
Iglesia-Vicente, Faustino Mollinedo, José Manuel Andreu, Manuel Medarde and Rafael Peláez
†
Laboratorio de Química Orgánica y Farmacéutica, CIETUS and IBSAL, Facultad de Farmacia,
‡
Universidad de Salamanca, Campus Miguel de Unamuno, Eꢀ37007 Salamanca, Spain, Centro de
#
Investigaciones Biológicas, CSIC, Madrid 28040, Spain, and Instituto de Biología Molecular y Celular
del Cáncer, Centro de Investigación del Cáncer, CSICꢀUniversidad de Salamanca, Campus Miguel de
Unamuno, Eꢀ37007 Salamanca, Spain.
RECEIVED DATE
*
To whom correspondence should be addressed. Telephone: 34 923 294528. Fax: 34 923 294515.
Eꢀmail adress: pelaez@usal.es
†
Laboratorio de Química Orgánica y Farmacéutica, CIETUS and IBSAL
‡
Centro de Investigaciones Biológicas
#
Instituto de Biología Molecular y Celular del Cáncer
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 52
ABSTRACT. Colchicine site ligands with indole B rings are potent tubulin polymerization inhibitors.
Structural modifications at the indole 3ꢀposition of 1ꢀmethylꢀ5ꢀindolyl based isocombretastatins (1,1ꢀ
diarylethenes) and phenstatins endowed them with anchors for further derivatization and resulted in
highly potent compounds. The substituted derivatives displayed potent cytotoxicity against several
human cancer cell lines due to tubulin inhibition, as shown by cell cycle analysis, confocal microscopy
and tubulin polymerization inhibitory activity studies and promoted cell killing mediated by caspaseꢀ3
activation. Binding at the colchicine site was confirmed by means of fluorescence measurements of
MTC displacement. Molecular modeling suggests that the tropoloneꢀbinding region of the colchicine
site of tubulin can adapt to hosting small polar substituents. Isocombretastatins accepted substitutions
better than phenstatins, and the highest potencies were achieved for the cyano and hydroxyiminomethyl
substituents, with TPI values in the submicromolar range and cytotoxicities in the subnanomolar range.
A 3,4,5ꢀtrimethoxyphenyl ring usually afforded more potent derivatives than a 2,3,4ꢀtrimethoxyphenyl
ring.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
INTRODUCTION
The eukaryotic microtubule system serves a variety of architectural functions in cells and provides a
dynamic scaffold for cell shape maintenance, flagellar and ciliary movement, and the intracellular
transport of vesicles and other cellular components. It also supplies the conveyor belt of the mitotic
spindle for chromosomes during mitosis, thus allowing them to become properly positioned at the
metaphase plate and after which the sister chromatids are pulled apart towards the cell poles in
1
anaphase. The important role of microtubule dynamics makes them important targets in anticancer
chemotherapy. Most drugs interfering with microtubule dynamics act directly on tubulin, such as the
2
stabilizing taxanes, and the destabilizing vinca alkaloids and colchicine site ligands.
2
ACS Paragon Plus Environment

Page 9 of 52
Journal of Medicinal Chemistry
Among the colchicine site ligands, combretastatins have become structural models for the design of
1
2
new analogues as a result of their cytotoxic effects on tumor cells and their ability to collapse the tumor
vasculature, in the shape of different salts of the phosphate prodrug of combretastatin Aꢀ4
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(
fosbretabulin), the diphosphate prodrug of combretastatin Aꢀ1 (OXi4503) and the combretastatin
3
analogue AVE 8062, which are currently undergoing phase I and II clinical trials (Fig. 1). However,
combretastatins show low aqueous solubility due to the hydrophobic nature of the colchicine site,
chemical instability brought about by their isomerization to the inactive but thermodynamically more
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
stable trans isomers, and insufficient cytotoxicity against cells in the tumor rim, resulting in tumor
5
regrowth. Several approaches have been undertaken to solve each of these disadvantages, with different
degrees of success.
Figure 1. Structures of combretastatin analogues undergoing clinical trials.
We have recently prepared a new family of macrocyclic combretastatin analogues that are unable to
become converted to the trans isomers but they display reduced potency, probably owing to steric
6
reasons. With the same aims, but following a different strategy, we have also recently described a new
family of combretastatin isomers, which we have called isocombretastatins, and structurally related
7
ketones and oximes, which are potent inhibitors of tubulin polymerization and cytotoxic compounds.
The most potent analogues of these families carry a 1ꢀmethylꢀ5ꢀindolyl B ring, but these compounds
suffer from low water solubility. Other groups have also found that Nꢀmethylꢀ5ꢀindolyl rings are good
scaffolds for colchicine site ligands in terms of potency but are associated with poor physicochemical
8
properties. In the same vein, we have also found that Nꢀmethylꢀ5ꢀindolylꢀsubstituted combretastatins
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 52
9
are poor cytotoxic compounds despite their potent inhibition of tubulin polymerization. Liou et al have
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
explored the possibility of incorporating acidic substituents at the indole nitrogen of 5ꢀaroylindoles in
1
0
order to improve their physical chemical properties, but these result in much lower potencies. The
most successful strategy applied to increase water solubility in combretastatins is the formation of
phosphate prodrugs of phenol groups already present in the parent compounds, but 2ꢀ or 3ꢀ
hydroxyindoles are not stable and therefore this approach cannot be easily applied to 5ꢀindolyl
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
11
analogues.
In this work, we set out to study the possibility of incorporating polar groups at the position three of
the indole of 5ꢀbenzoylindoles that might serve as an anchor for further derivatization in ways similar to
what is seen for phenolic groups in combretastatins. The phenolic group of combretastatin Aꢀ4 can be
replaced by a hydrogen atom or with amines with only small changes in potency, but this position has
12
proved to be important as a solubilizing group and as anchors for derivatization purposes. A similar
group is thus needed to extend the possibilities of the development of 5ꢀindolylꢀcontaining tubulin
1
3
polymerization inhibitors. Following the pharmacophore model of Nguyen for colchicine site ligands
7
and our docking studies of 5ꢀindolylisocombretastatins, phenstatins and related compounds it could be
suggested that the 3 position of the indole ring might accommodate hydrogen bond donors (such as
amino substituents), which might hydrogenꢀbond to the carbonyl oxygen atom of Thrβ177
(corresponding to pharmacophore point D1, such as the phenolic group of combretastatin Aꢀ4) but also
hydrogen bond acceptors to bind the amide nitrogen of Valβ179 (corresponding to pharmacophore point
A2, such as the carbonyl oxygen of the tropolone ring of colchicine), as has been discussed for CAꢀ4
1
4
and related 4ꢀarylcoumarins. To date few substitutions on the indole ring have been described, but
potent combretastatin, sulfonamide and phenstatin analogues carrying carbazole rings suggested to us
that an expanded tropolone site might be exploited for improved potency and/or pharmacokinetic
9
,15
modulation.
In order to explore new possibilities for the indoleꢀbased interfacial inhibitors, and continuing our
4
ACS Paragon Plus Environment

Page 11 of 52
Journal of Medicinal Chemistry
previous work studying the effect of different nitrogen substituents or other modifications at the
phenstatin bridge and the effect of a 2,3,4ꢀtrimethoxyphenyl as Aꢀring, here we prepared and assayed a
set of compounds characterized by the presence of substituents at the 3ꢀposition of the 5ꢀindolyl unit.
Considering that the introduction of the more sterically demanding indoles might affect the overall
disposition of the ligands in the colchicine site, the effect of a 2,3,4ꢀtrimethoxyphenyl ring was also
explored. Additionally, the replacement of the indole by benzimidazole and benzoxazole systems was
addressed. These modifications are very interesting in terms of interactions with the protein, because
new possibilities for binding can be assessed, as well as for activity modulation, solubilization
improvement and derivatization (Figure 2).
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
X, Y = MeO/H
Y
Z
Y
Z
R
MeO
MeO
MeO
MeO
Z = O, CH2, H/CH3
N
N
R= CHO, CN, CH OH, CHNOH,
2
X
X
COOH, CONH2, NH2 , CH2NH2
protein binding, solubilization and
derivatisation
Described indolephenstatins
New indolephenstatins
Figure 2. Structures of the new 3ꢀsubstituted indole analogues.
RESULTS AND DISCUSSION
Chemical synthesis
The synthesis of the phenstatinꢀlike scaffolds was planned and carried out according to our previously
described methodologies for the synthesis of Nꢀsubstituted indoleꢀphenstatins, and their derivatives
modified on the bridge, such as isocombretastatins (Scheme 1). The introduction of the 3ꢀsubstituents
was attempted at two different stages: on the monoarylic starting materials and on the preassembled
nonꢀsubstituted diarylic precursors. Electrophilic reactions on the 5ꢀsubstitutedꢀNꢀmethylꢀ1Hꢀindoles
proved to be problematic, easily evolving into complex red mixtures. Therefore, substitutions were
performed at the diarylic stage. Several attempts to introduce amino groups at the three positions of the
indole rings of phenstatins and isocombretastatins by means of a nitration ꢀ reduction sequence failed
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 52
1
6
due to chemical instability of the amines. Instead, VilsmeierꢀHaack formylation was selected and we
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
prepared the other derivatives from the aldehydes.
Starting from commercial 5ꢀbromoꢀ1Hꢀindole, Nꢀmethylation and coupling with 2,3,4ꢀ or 3,4,5ꢀ
trimethoxybenzaldehydes yielded benzhydrols 2 and 3, which upon PDC oxidation to phenstatins 4 and
5
followed by Wittig methylenations yielded isocombretastatins 6 and 7. The corresponding
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
dihydroderivatives 14 and 15 were obtained by hydrogenation (Scheme 1). Alternatively,
transmetalation of 5ꢀbromoꢀNꢀmethylꢀ1Hꢀindole and coupling of the organolytic compound with the
sodium salts of the trimethoxybenzoic acids yielded the benzophenones directly in similar overall
yields. From these already known compounds, the preparation of the new derivatives carrying the
substitution at the indole 3ꢀposition can be completed by formylation of the highly reactive indole
nucleus. Under VilsmeierꢀHaack conditions, aldehydes 8 and 9 were readily produced, whereas for
isocombretastatins mixtures of the monoformylated (10 or 11) and diformylated (12 or 13) products
were formed by sequential reaction with the indole and with the electronꢀrich olefin under more forcing
conditions. As shown by NMR, 12 and 13 were obtained as Z+E mixtures.
a
Scheme 1: Synthesis of 3ꢀformylated indoleꢀphenstatins and indoleꢀisocombretastatins.
a. Reagents and conditions: (a) (i) NaOH (10%), nBu NHSO , CH Cl , 1 h, rt; (ii) IMe, 24 h. (b) (i)
4
4
2
2
6
ACS Paragon Plus Environment

Page 13 of 52
Journal of Medicinal Chemistry
BuLi (1.6M), THF, ꢀ40ºC, 1h; (ii) 3,4,5ꢀ or 2,3,4ꢀtrimethoxybenzaldehyde, THF, ꢀ40ºC→rt, 24 h. (c)
PDC, CH Cl , 0ºC, 4Å ms, 4 ꢀ 24 h. (d) (i) Ph PMeI, THF, ꢀ40ºC, BuLi (1.6M), 1 h; (ii) 4 or 5, THF, ꢀ
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
2
3
4
0 ºC → rt, 24 h. (e) (i) POCl , DMF, 0ºC, 1 h; (ii) 4 or 5 or 6 or 7, 60ºC, 2 h. (f) Pd/C (10%), EtOH, H
3
2
atmosphere, 24 h.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
With ketoꢀaldehydes 8 and 9 at hand, the first modification performed was the preparation of the
carboxylic acids in an attempt to improve water solubility and as a means of synthesizing further
derivatives. Direct oxidation of the isocombretastatins was complicated by the double bond, and
phenstatin 16 was produced after the oxidation of aldehyde 10. This problem was circumvented by
previous oxidation of aldehyde 8 to carboxylic acid 16, followed by its methylenation to 17 (Scheme 2).
Unexpectedly, acid 16 was extremely insoluble in many solvents (including THF, water, methanol,
CH Cl and DMSO), which made it very difficult to handle and led to very low reaction yields in all
2
2
reactions attempted. The presence of minor amounts of nonꢀoxidized ketoaldehyde 8 accounts for the
isolation of diolefin 18 as a byꢀproduct of the olefination. Although the yield of this process was very
low, attempts to prepare the carboxylic acid by alternative procedures (trifluoroacetylationꢀhydrolysis,
ethoxycarbonylationꢀhydrolysis) failed or resulted in the degradation of the reacting material.
a
Scheme 2: Synthesis of isocombretastatin indoleꢀ3ꢀcarboxylic acid derivatives
a. Reagents and conditions: (a) 8 or 10 in tBuOH, 2ꢀmethylꢀ2ꢀbutene, THF; add NaClO , NaH PO ,
2
2
4
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 52
H O, rt, 12 h. (b) (i) Ph PMeI, THF, ꢀ78ºC, BuLi (1.6M), 1 h; (ii) 16, THF, ꢀ78ºC→rt, 24 h.
2
3
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Hydrogenation of olefinꢀaldehyde 10 reduced both the double bond and the carbaldehyde to the
corresponding alkanes (19). Several transformations aimed at increasing solubility by means of
additional nitrogen or oxygen atoms and/or enabling further modifications that could in some cases also
regenerate the simple carbonyl group by hydrolysis were also carried on the carbaldehydes (Scheme 3).
Thus, hydrazones 20 and 21 and oximes 22 and 23 were prepared as stereoisomeric Z+E mixtures,
because isomerization of the isomers isolated was observed. In the case of 21, simultaneous hydrazone
formation and bridge reduction occurred under the conditions assayed. The oximes can be used to easily
prepare other derivatives at the 3ꢀposition: the acetyloximes and the nitriles. By treatment with acetic
anhydride in pyridine, compounds 24ꢀ27 were produced by acetylation and acetylationꢀelimination of
the oximes, as well as a minor amount of ketoꢀnitrile 28. Compound 24 was also used to prepare amide
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
9. Finally, aldehydes 10 and 11 were modified by reduction to the 3ꢀhydroxymethyl derivatives 30 and
1. In view of the high potency displayed by the nitrile derivative 24, the synthesis of the dinitrile
3
derivative 32 was tackled from dialdehyde 12 (Scheme 4) following the same procedure.
a
Scheme 3: Preparation of isocombretastatin indoleꢀ3ꢀcarbaldehyde derivatives.
8
ACS Paragon Plus Environment

Page 15 of 52
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a. Reagents and conditions: (a) Pd/C (10%), EtOH, H atmosphere, 24 h. (b) NH NH ꢁH O, MeOH,
2
2
2
2
HOAc (2 droplets), rt, 24 h. (c) NH OHꢁHCl, MeOH, pyridine (2 droplets), rt, 24 h. (d) Ac O, pyridine,
2
2
rt, 4 h. (e) NaOH (10%), MeOH, rt, 16 h. (f) NaBH , MeOH, rt, 30 min.
4
a
Scheme 4: Preparation of dinitrile 32.
a. Reagents and conditions: (a) (i) NH OHꢁHCl, MeOH, pyridine (2 droplets), rt, 24 h; (ii) Ac O,
2
2
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 52
pyridine, rt, 4 h.
Simultaneously
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
to
the
synthesis
of
indole
phenstatins,
isocombretastatins
and
dihydroisocombretastatins, which in general display high potency in the inhibition of tubulin
polymerization and cytotoxicity assays (see below), we decided to replace the indole nucleus by related
bicyclic systems, such as benzimidazole and benzoxazole. Accordingly, 5(6)ꢀbromoꢀ1ꢀmethylꢀ1Hꢀ
benzimidazoles were obtained as a mixture of regioisomers by treatment of 4ꢀbromoꢀoꢀphenylene
diamine with methyl orthoformate followed by methylation of the 5ꢀbromoꢀ1Hꢀbenzimidazole thus
obtained. The mixture of bromides was treated with nꢀbutylꢀlithium and reacted with 3,4,5ꢀ
trimethoxybenzaldehyde, yielding products 33 and 34, arising from the metalation at the 2ꢀposition
instead of those derived from the metalꢀhalogen exchange. Oxidation of this reaction mixture produced
the ketoꢀimidazole derivatives 35 and 36. Under identical conditions, 5ꢀbromobenzoxazole (obtained
from 4ꢀbromoꢀ2ꢀnitrophenol by reduction to 4ꢀbromoꢀ2ꢀaminophenol and treatment with methyl
orthoformate) yielded alcohol 37, which was oxidized to the benzoxazoleꢀphenstatin 38. Owing to the
low tubulin polymerization inhibitory activity and cytotoxicity displayed by the latter compound, the
synthesis of the 5ꢀ(3,4,5ꢀtrimethoxybenzoyl)ꢀ1ꢀmethylꢀ1Hꢀbenzimidazole was not further pursued.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
Scheme 5: Synthesis of benzimidazole and benzoxazole derivatives 33ꢀ38.
10
ACS Paragon Plus Environment

Page 17 of 52
Journal of Medicinal Chemistry
a. Reagents and conditions: (a) (i) BuLi (1.6M), THF, ꢀ78ºC, 1 h; (ii) 3,4,5ꢀtrimethoxybenzaldehyde,
THF, ꢀ78ºC→rt, 24 h. (b) PDC, CH Cl , 0ºC, 4Å ms, 24h. (c) KMnO , nBuNSO , H O, CH Cl , rt, 24
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
2
4
4
2
2
2
h.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The 1Hꢀindole derivatives without substitution at 1ꢀposition were synthesized according to Scheme 6,
using the Nꢀbenzenesulphonylꢀprotected starting material to prepare benzhydrol 39, oxidation to
phenstatin 40, olefination to isocombretastatin 41 and deprotection to 42. The latter was formylated to
dihydroisocombretastatin 44 , through previous hydrogenation to 43, and directly to the dicarbaldehyde
derivative 45 (Z+E).
a
Scheme 6: Preparation of 1Hꢀindole isocombretastatin and dihydroisocombretastatin derivatives.
a. Reagents and conditions: (a) (i) 5ꢀbromoꢀ1,2,3ꢀtrimethoxybenzene, BuLi (1.6M), THF, ꢀ60ºC, 1.5
h; (ii) 1ꢀ(phenylsulfonyl)ꢀ1Hꢀindoleꢀ5ꢀcarbaldehyde, THF, ꢀ60ºC→rt, 24 h. (b) PDC, CH Cl , 0ºC→rt,
2
2
4
Å ms, 12 h. (c) (i) Ph PMeI, THF, ꢀ40ºC, BuLi (1.6M), 1.5 h; (ii) 40, THF, ꢀ40ºC→rt, 12 h. (d) NaOH
3
(
5M), EtOH, reflux, 20 h. (e) Pd/C (10%), EtOH, H atmosphere, 24 h. (f) (i) POCl , DMF, 0ºC, 30 min;
2
3
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 52
(
ii) 42 or 43, 60ºC, 2 h.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Following these methodologies a large number of analogues can be prepared by standard procedures,
but these compounds are sufficiently representative to accomplish assays aimed at elucidating the
effects of the modification at the 3ꢀposition at the indole nucleus of this class of antimitotics on tubulin
and on cytotoxicity.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Biology
The compounds synthesized were assayed by the XTT method for growth inhibition against four
types of human cancer cell lines: Aꢀ549 human lung carcinoma, HeLa human cervix epithelioid
carcinoma, HLꢀ60 human acute myeloid leukemia, and HTꢀ29 human colon adenocarcinoma. The
results are summarized in Table 1. Some of these compounds inhibited the proliferation of the indicated
ꢀ9
ꢀ10
human cancer cell lines at very low drug concentrations, with IC values in the 10 M and 10
M
5
0
range. These IC values are up to 10ꢀ100 times lower than those shown by doxorubicin (IC at a range
5
0
50
ꢀ8
of 10 M) (Fig. 3).
12
ACS Paragon Plus Environment

Page 19 of 52
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. Effects of compound 14 and doxorubicin on the cell proliferation of HeLa cells. Cells were
incubated with compound 14 and doxorubicin at the indicated concentrations. After 3 days, cell
proliferation was determined by the XTT assay and plotted as a percentage of untreated control cells.
Results are mean values ± SD of a representative experiment carried out in triplicate, out of three
performed.
Most of the compounds assayed were cytotoxic at micromolar concentrations or lower (Table 1), thus
indicating that the common 5ꢀ(trimethoxybenzyl)indole is a very favorable scaffold for cytotoxicity.
The benzophenones (phenstatins) with 3ꢀsubstituted indoles (e.g. 8, 9, and 28) displayed potencies in the
hundreds of nanomolar concentrations and were less potent than the corresponding isocombretastatins
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 52
(
1,1ꢀdiarylethenes, e.g. 24, 26, and 29), the most potent of them being cytotoxic in the nanomolar range.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
This result is in agreement with our previous findings showing that indoleisocombretastatins are more
7
13
potent than indolephenstatins and suggests that, contrary to what was previously considered, the
carbonyl oxygen of phenstatins is not an essential pharmacophoric anchor point. Binding at the
colchicine site of tubulin is usually considered to involve a geometric arrangement with two nonꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
3
coplanar aromatic rings that must disrupt the conjugation of the carbonyl group with the aromatic
rings and impose a more severe conformational penalty on binding for the carbonyls than for the olefins.
Irrespective of the bridge and the nature of the indole substitution, the 2,3,4ꢀtrimethoxyphenyl
analogues were less potent than the 3,4,5ꢀtrimethoxyphenyl ones (e.g. compare 8 to 9, 10 to 11, and 24
7
to 26), similar to what has been reported previously for the unsubstituted indoles. This is probably a
consequence of the scaffold geometry, with the oneꢀcarbon bridge placing the two aromatic moieties
when the compound is in a complex with tubulin similar to the aromatic rings of podophyllotoxin, and
the cleft for the A ring hosting the methoxy groups at the 3, 4 and 5 positions better. Allocating a 2,3,4ꢀ
trimethoxyphenyl ring at this site would require tilting of the indole ring, which does not seem to be
possible in sight of the lower potency observed, thus suggesting a scarce adaptability of the B ring
pocket in the colchicine site. The introduction of substituents at the indole three position increases the
bulk and likely aggravates this difficulty for the 2,3,4ꢀtrimethoxyphenyl analogues. For these reasons,
we selected the complex between tubulin and 1sa1.pdb for the molecular docking studies (see below)
and below we mainly discuss the results for the more potent 3,4,5ꢀtrimethoxyphenyl analogues.
Overall, the introduction of substituents at the indole three position results in a decrease in cytotoxic
potency, in good agreement with the low expandability of the B ring pocket at the colchicine site.
Consistent with the previously discussed effect of the bridge, this decrease was more pronounced in the
indolephenstatins (comparing unsubstituted 4 with aldehyde 8, with acid 16, or with nitrile 28, revealed
1
0ꢀ to 100ꢀfold potency decreases) than in the isocombretastatins with the same substituents (comparing
unsubstituted 6 with aldehyde 10 or with acid 17 caused a 10ꢀ to 100ꢀfold potency decrease, but potency
14
ACS Paragon Plus Environment

Page 21 of 52
Journal of Medicinal Chemistry
remained similar for nitrile 24). Further exploration of the isocombretastatins series led to amide 29,
which is an improvement over the nanomolar potencies of the parent indoles. The presence of hydrogenꢀ
bond donors and acceptors in 29 seemed to improve the cytotoxic potency with respect to substituents
with only acceptors (e.g. aldehyde 10 or nitrile 24) or with only hydrogenꢀbond donors (e.g. hydrazones
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
0), although in this latter case the presence of two stereoisomers might also contribute to the loss of
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
potency. This might also be the case for oximes 22, which despite their similarity to the amide
arrangement of hydrogenꢀbond donors and acceptors showed a lower cytotoxic potency, although they
represented an improvement with respect to their parent aldehyde 10, with only a hydrogenꢀbond
acceptor. This observation is interesting, since if they behaved as prodrugs releasing the aldehyde, one
would expect at most a potency equal to that of aldehyde 10 (which is in fact a more potent inhibitor of
tubulin polymerization, see Table 1 and below). Therefore, if they behave as prodrugs of the aldehyde
they must have other advantageous properties, such as for instance better solubility or cell uptake,
therefore attracting more interest and suggesting a starting point for potential double prodrug
derivatization of the aldehydes. The charged carboxylate resulted in a lower cytotoxic potency, which
might reflect a combination of poor solubility (despite of its charge), cell uptake, and tubulin
polymerization inhibitory activity (even if the carboxylate is close to the ammonium group of lysine
3
50β, see later). Reduction of aldehydes 10 and 11 to methanols 30 and 31 respectively, produced a
sharp decrease in potency, thus confirming the importance of a hydrogenꢀbond acceptor moiety placed
in the same plane as the indole ring. The more potent 3ꢀsubstituted indoleisocombretastatins, such as 10,
2
4 and 29, provide anchor points for further derivatization and provide interesting opportunities for
1
7
increases in solubility or drug targeting.
Taking into account the apparent importance of the overall indole bulk on cytotoxic potency, we
explored the effect of removing the methyl group on the indole nitrogen and also replaced the Nꢀ
methylindole by a benzoxazole, which also incorporates a less bulky hydrogenꢀbond acceptor in the ring
plane. However, the methyl group seemed to be important for the activity of the substituted derivatives,
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 52
since the compounds with a free indole amino group (compare 6 to 42, or 14 to 43) showed reduced
potencies, as did benzoxazole 38.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Considering the previously discussed potentially deleterious effect on cytotoxicity of the conjugation
2
of the aromatic rings with the sp bridges we decided to explore the effect of reducing the olefin bridge
to ethyl analogues. The unsubstituted hydrogenated derivative 14 proved to be very potent, but
substitution of the indole caused a sharp decrease in potency (e.g. 19) and therefore was not further
pursued. The introduction of other modifications on the olefinic bridge of the isocombretastatins did not
result in a consistent improvement with respect to the unsubstituted olefins, although the dinitrile
derivative 32 displayed consistent nanomolar potencies in all four cell lines studied.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
In order to interpret the results on cytotoxicity, the effect of the synthesized compounds 2 ꢀ 45 at a
concentration of 20 ꢀM on in vitro tubulin assembly was studied with bovine brain tubulin. For
compounds showing more than 50 % inhibition at this concentration the IC50 values of tubulin
polymerization inhibition (TPI) were calculated in order to compare potencies. Overall, there was good
agreement between the TPI and cytotoxic potencies, mainly for HeLa cells (Supplementary material:
Fig. S1). However, slightly different effects against Aꢀ549, HLꢀ60 and HTꢀ29 were observed for some
highly potent inhibitors of tubulin polymerization, such as with 22. The TPI results show that aldehyde,
oxime, nitrile and amides at position three of the indole are acceptable for tubulin binding, regardless of
the differences they cause in cytotoxic potencies against different cell lines. These results suggest that
both hydrogenꢀbond donors and acceptors are suitable modifications to obtain indoleꢀbased tubulin
inhibitors, in agreement with previous results with colchicine site analogues incorporating both types of
1
4
functionalities, as already mentioned.
The binding site and the association constant of the most cytotoxic and potent inhibitors of tubulin
18
19
assembly were obtained using the MTC displacement assay. MTC is a bonaꢀfide fluorescent probe of
1
7
the colchicine site, which largely increases its fluorescence upon binding. For phenstatins and
isocombretastatins, displacement of MTC from its binding site, which can be observed by the decay of
16
ACS Paragon Plus Environment

Page 23 of 52
Journal of Medicinal Chemistry
the fluorescence of the MTC upon ligand addition, was used. The observed displacement confirms
binding at or close to the colchicine binding site. Consistent with the TPI activities, the strongest
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
7
ꢀ1
association constant was for the cyanoꢀsubstituted indole isocombretastatin 24 (1.8±0.8 10 M ),
whereas the unsubstituted indolic phenstatin 4, isocombretastatin 6 and formylated isocombretastatin 10
6
ꢀ1
displayed association constants in the range of 10 M . In addition, a positive correlation was observed
between the cytotoxicity of the compounds studied in HLꢀ60 cells and their affinity for the colchicineꢀ
binding site of tubulin (Supplementary material: Fig. S2). The affinity range achieved with this ligand
series spanned almost three orders of magnitude.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
In order to ascertain that the cytotoxic effects of the assayed compounds were due to tubulin
inhibition, we analyzed the effect of the different compounds on cell cycle analysis by flow cytometry.
We used different drug concentrations (5, 10, 50 and 100 nM) and the HeLa cancer cell line in timeꢀ
course experiments (18, 24, 48 and 72 h incubation times). Compound 24, as one of the most active
ꢀ8
compounds, totally arrested the HeLa cell cycle at the G /M phase (> 83%) when used at 10 M for 18
2
h (Fig. 4). Similarly, compounds 14, 29 and 32, which together with compound 24 were the most active
ꢀ8
compounds, arrested practically all the cells at G /M phase of the cell cycle (> 85%) when used at 10
2
M after 18 h incubation (data not shown).
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 52
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ꢀ
8
Figure 4. Effects of compound 24 on the cell cycle in HeLa cells. Cells were incubated with 10 M
compound 24 for 18 h and stained with propidium iodide (PI). Their DNA content was analyzed by
fluorescence flow cytometry. Control untreated cells were run in parallel. The positions of the G /G
0
1
and G /M peaks are indicated by arrows. Data shown are representative of three experiments performed.
2
This cell cycle arrest was followed by the onset of apoptosis at later incubation times, as detected by
the appearance of cells with DNA contents at the subꢀG /G region of cell cycle analysis, a significant
0
1
percentage of apoptotic cells being reached at 72 h incubation (e.g.: 20% apoptosis following incubation
ꢀ8
of HeLa cells with 10 M compound 24). Compounds that interfere with the microtubule system
2
0
typically cause mitotic arrest at the G /M phase before the induction of an apoptotic cell death,
2
consistent with our results. Furthermore, confocal microscopy studies showed severe disruption of the
microtubule system of HeLa cells treated with 10 nM of the compounds (Fig. 5), further confirming that
the observed effects were due to interaction with tubulin, resulting in a prolonged mitotic arrest that
18
ACS Paragon Plus Environment

Page 25 of 52
Journal of Medicinal Chemistry
2
1
ultimately leads to cell death.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5. Effects of compounds 10, 11, 14, 22, 24, 29 and 32 on the microtubule network of HeLa cells.
ꢀ8
Cells were incubated in the absence (Control) or in the presence of 10 M of compounds 10, 11, 14, 22,
4, 29 and 32 for 18 h and then fixed and processed to study the immunofluorescence of microtubules
2
(red fluorescence) and nuclei (blue fluorescence), as described in Experimental Section. Bar: 25 ꢂm.
The photomicrographs shown are representative of at least three independent experiments performed.
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 52
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The type of cell killing induced by the most potent inhibitors of tubulin polymerization and cytotoxic
compounds was investigated by Western Blot, measuring the activation of caspaseꢀ3 (Fig. 6) in HeLa
cells after incubation with the compounds. Following 24ꢀh of incubation, caspaseꢀ3 activation was
detected and after 72 h cells started to accumulate in the subꢀG /G region, indicating the induction of
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
0
1
apoptosis. We found that the most active compounds, namely 14, 24, 29 and 32, induced caspaseꢀ3
activation in HeLa cells (Fig. 6) as assessed by the processing of procaspaseꢀ3 into the p19/p17 active
forms, and the cleavage of the typical caspaseꢀ3 substrate PARP, using a polyclonal antihuman caspaseꢀ
3
antibody that recognized the p19/p17 forms of the active caspaseꢀ3, and the antiꢀPARP C2.10
monoclonal antibody that detected both the 116ꢀkDa intact form and the 85ꢀkDa cleaved form of the
caspaseꢀ3 substrate PARP. A similar type of behavior was detected with the other active dugs assayed in
this work (data not shown). In addition, the incubation of HeLa cells with the cell permeable panꢀ
caspase inhibitor zꢀVADꢀfmk (50 ꢂM) prevented the induction of apoptosis triggered by compound 24
(19.5% vs. 5% apoptosis in 24ꢀtreated HeLa cells incubated in the absence or presence of zꢀVADꢀfmk
for 72 h), further supporting the involvement of apoptosis in the cell killing process.
Figure 6. Caspaseꢀ3 activation following treatment of HeLa cells with compounds 14, 24, 29 and 32.
Cells were treated with 50 nM of compounds 14, 24, 29 and 32 for the indicated times and analyzed by
immunoblotting with antiꢀactivated caspaseꢀ3 (SDSꢀ12% polyacrylamide gels) and antiꢀPARP
antibodies (SDSꢀ8% polyacrylamide gels). Untreated control cells (C) were run in parallel. The
migration positions of activated caspaseꢀ3, as well as of fullꢀlength PARP and its cleavage product p85,
20
ACS Paragon Plus Environment

Page 27 of 52
Journal of Medicinal Chemistry
are indicated. Data shown are representative of three experiments performed.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 1. Tubulin polymerization inhibitory activity, antineoplastic cytotoxic activity against human
tumor cell lines and binding parameters.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
a
a
a
TPI (%)
TPI
IC50 (nM)
HL-60
IC50 (nM)
A-549
IC50 (nM)
HeLa
IC50 (nM)
HT-29
Compound
Kass
at 20ꢀM IC50 (ꢀM)
6
4
5
98
100
96
97
61
73
100
98
88
96
98
100
87
2
7.9
0.4
27 ± 1
190 ± 30
380 ± 10
28 ± 1
25 ± 6
27 ± 4
2.16 ± 0.3 x 10
240 ± 10
2.8 ± 0.4
280 ± 10
310 ± 10
780 ± 120
28 ± 2
30 ± 1
240 ± 20
3.1 ± 0.1
310 ± 10
3000 ± 200
5300 ± 900
62 ± 8
6
6
0.7
2.7 ± 0.2
32 ± 1
3.8 ± 2 x 10
6
7
4.7
730 ± 60
7.6 ± 1 x 10
4
5
8
16.9
> 10
320 ± 10
780 ± 110
20 ± 3
3.1 ± 2 x 10
4
5
9
> 10
2.6 ± 1 x 10
5
10
0.9
4.0
≈7
3000 ± 100
1800 ± 200
340 ± 10
320 ± 10
350 ± 40
6.8 ± 0.9
470 ± 20
3.6 ± 3 x 10
4
1
1
32 ± 1
32 ± 1
230 ± 20
220 ± 10
170 ± 10
300 ± 10
3.8 ± 0.2
330 ± 80
<10
b
5
1
2
33 ± 2
31 ± 1
2.4 ± 0.5 x 10
1
2Z
29 ± 1
30 ± 2
b
0
4
1
3
5.9
1
300 ± 10
0.33 ± 0.05
370 ± 60
300 ± 10
5.3 ± 2 x 10
14
15
0.64 ± 0.08
310 ± 10
7000 ± 1600
880 ± 50
310 ± 10
38 ± 1
5
3.1 ± 2 x 10
4
4
4
4
1
1
1
6
7
9
> 10
> 10
> 10
< 10
91
90
12
0
≈5
1100 ± 200
280 ± 10
180 ± 20
1800 ± 100
21 ± 4
1100 ± 500
320 ± 10
550 ± 140
310 ± 10
300 ± 10
3200 ± 100
33 ± 1
b
4
2
0
3200 ± 100
5.4 ± 4 x 10
4
4
21
>10
430 ± 20
6.2 ± 0.5
34 ± 1
< 10
b
6
2
2
2
3
93
95
93
89
40
93
32
2.6
4.0
1.2
2.2
30 ± 1
290 ± 10
8.4 ± 0.2
360 ± 10
2000 ± 400
1.8 ± 0.4
3200 ± 100
8.6 ± 4 x 10
b
5
270 ± 10
1.3 ± 0.4
27 ± 3
320 ± 10
3.9 ± 0.2
32 ±1
5.4 ± 3 x 10
7
2
4
2.9 ± 0.1
32 ± 1
1.8 ± 0.8 x 10
6
2
6
4.0 ± 3 x 10
28
29
290 ± 10
0.11 ± 0.01
250 ± 40
270 ± 10
0.25 ± 0.05
310 ± 10
310 ± 20
0.51 ± 0.07
2000 ± 100
0.9
4
3
0
<10
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 52
a
a
a
a
TPI (%)
TPI
IC50 (nM)
HL-60
IC50 (nM)
A-549
IC50 (nM)
HeLa
IC50 (nM)
HT-29
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Compound
Kass
at 20ꢀM IC50 (ꢀM)
6
31
32
220 ± 10
4.8 ± 0.7
440 ± 20
4.9 ± 0.8
300 ± 10
3.5 ± 0.1
460 ± 60
2.1 ± 0.5
1.0 ± 0.6 x 10
b
3
2
90
2
4.7
4
4
4
4
3
3
4
5
>10
>10
>10
>10
4
4
4
4
3
3
22
37
37
0
>10
>10
>10
>10
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
29 ± 3
400 ± 30
190 ± 10
2300 ± 200
340 ± 30
340 ± 20
4
36
2400 ± 100
470 ± 40
~10
2700 ± 300
2800 ± 200
5100 ± 700
3300 ± 600
5800 ± 300
4
3
4
4
4
8
2
3
4
~10
36
5
3600 ± 500
3400 ± 300
5400 ± 400
9000 ± 1100
4900 ± 800
9100 ± 200
2100 ± 500
2300 ± 600
5800 ± 700
0
a
Values are derived from concentrationꢀresponse curves using the XTT assay as described in the
Experimental section. Data are shown as the main values ± S.D. of three experiments performed in
b
triplicate. Assayed as Z+E mixtures.
Theoretical calculations were performed in order to explain the relative biological activity of the
different compounds. Figure 7 shows the superimpositions of the DFT optimized indolephenstatin 4 and
indoleisocombretastatin 6 onto podophyllotoxin (carbon atoms colored in cyan), revealing an extensive
overlapping of the aromatic rings. This is in good agreement with the pharmacophoric models for the
colchicine site, which assign an essential role to the spatial arrangement of the two aromatic rings.
However, the higher potency of isocombretastatins when compared to the phenstatins, despite these
latter having one more potential hydrogenꢀbond acceptor point due to the carbonyl oxygen atom, could
be explained in terms of the greater preference of the carbonyl to remain conjugated to the phenyl rings.
This preference results in a less favorable disposition of the phenstatins when bound to tubulin.
22
ACS Paragon Plus Environment

Page 29 of 52
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
A)
B)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7. Superimposition of the most stable DFT conformers of 3ꢀunsubstituted compounds (A) 4 and
, and formylated analogues (B) 8 and 10 onto podophyllotoxin (carbons in cyan). Only those
6
conformations which place the phenyl rings close to those of podophyllotoxin are shown.
When DFTꢀminimized 3ꢀsubstituted indolephenstatin (8, R=CHO) and indoleisocombretastatin (10,
R=CHO) were superimposed onto podophyllotoxin (carbon atoms colored in cyan), there was a similar
overlapping of the aromatic rings (figure 7) as that described for the unsubstituted analogues. The
substituents protrude from the podophyllotoxin pocket towards its walls. For the phenstatins, the tilted
aromatic ring directs the substituents towards the upper and lower walls, while in the case of the
isocombretastatins they are in the podophyllotoxin benzodioxole plane, which might allow for their
better interaction with tubulin.
2
2
Attempts to dock the ligands into the rigid unmodified podophyllotoxin site of 1SA1.pdb resulted in
poses which did not place the trimethoxyphenyl rings in close proximity to those of podophyllotoxin
and colchicine, although the unsubstituted indoles did. These results pointed to the new substituents
being responsible. We therefore attempted to take site flexibility into account by allowing the sidechains
of the surrounding residues to rotate. This resulted in a more diffuse binding pattern for all the ligands
studied, including podophyllotoxin, colchicine, the unsubstituted indolic phenstatins and
isocombretastatins, and other ligands previously studied by us. Despite these unsuccessful results, which
suggested a different binding mode extending to additional regions of the site, as has also been
suggested for other families, but considering that all the available tubulin Xꢀray structures with
colchicine site ligands present ligands with very similarly sized B rings and that the structure – activity
23
ACS Paragon Plus Environment