Structural requirements for the interaction of combretastatins with tubulin: how important is the trimethoxy unit?

Article abstract of DOI:10.1039/b306878a

A series of combretastatins possessing both a trimethoxy unit and other substituents on ring A has been synthesised and tested for cytotoxicity and their ability to interact with the protein tubulin. All previous studies have indicated that the trimethoxy

Full text of DOI:10.1039/b306878a

View Article Online / Journal Homepage / Table of Contents for this issue
Structural requirements for the interaction of combretastatins
with tubulin: how important is the trimethoxy unit?†
Keira Gaukroger,a John A Hadfield,*a,b Nicholas J Lawrence,c,d Steven Nolana
and Alan T McGowna,b
a Drug Development Section, Paterson Institute for Cancer Research,
Christie Hospital NHS Trust, Wilmslow Road, Manchester, M20 4BX
b Centre for Molecular Drug Design, Cockcroft Building, University of Salford,
Manchester, M5 4WT
c Department of Chemistry, UMIST, PO Box 88, Manchester, M60 1QD
d Department of Chemistry, Cardiff University, PO Box 912, Cardiff, CF10 3TB
Received 19th June 2003, Accepted 9th July 2003
First published as an Advance Article on the web 24th July 2003
A series of combretastatins possessing both a trimethoxy unit and other substituents on ring A has been synthesised
and tested for cytotoxicity and their ability to interact with the protein tubulin. All previous studies have indicated
that the trimethoxy unit is essential for interaction with tubulin. The studies herein show that molecules possessing
functionalities other than trimethoxy can also interact with tubulin. Importantly a trimethyl substituted agent 52a
has shown reduced cytotoxicity, but increased potency in its ability to inhibit the assembly of tubulin.
the colchicine site on tubulin, is a trimethoxy aryl unit.1,2 It
Introduction
has previously been noted that alteration of the oxygenation
The combretastatins are a group of stilbenes which have
pattern from a trimethoxy motif on ring A of the combreta-
shown1–10 promising antimitotic activity. Combretastatin A-4
(1) is a naturally occurring stilbene which was originally
statins adversely affects their biological properties.2–4 This
methoxylation pattern of ring A appears to be critical for
isolated11 from the South African tree Combretum caffrum.
efficient binding to tubulin.3–6,9,13–15 The functionalities on ring
This phenolic stilbene 1 has proved to be the most potent
C of colchicine (2) and ring B of the combretastatins are the
agent which can interact9,11 with the colchicine site on tubulin.
key atoms which interact with tubulin and inhibit its
Furthermore combretastatin A-4 (1) has shown12 the ability to
assembly.14,15 Previous work has shown that small groups
shut down tumour vasculature at 10% of its maximum tolerated
dose without affecting the normal vasculature. Owing to its
promise during preclinical trials a phosphate derivative of
combretastatin A-4 (1) has now entered Phase II clinical trials.
(e.g. methyl, methoxy, dimethylamino etc.) on the 4-position of
the B ring of the combretastatins are essential for biological
activity.1,3,4,13 The trimethoxy unit on ring A is thought to
make an additive contribution to the strength of binding
A
major structural feature of several of the known
of colchicine 2 to tubulin and serves as an anchor that
maintains the whole molecule in the proper orientation within
the binding locus.14
antimitotic agents (combretastatin A-4 (1), colchicine (2),
steganacin (3), podophyllotoxin (4), Fig. 1), which bind at
Our group has extensive1,2,5–9,16–18 experience of the chemistry
and biochemistry of the combretastatins and was not totally
convinced of the necessity of a trimethoxyaryl group for
biological activity. The trimethoxy unit on the naturally
occurring combretastatins probably comes from the shikimate
biosynthetic pathway. Shikimic acid provides sequentially
phenylalanine, cinnamic acid and eventually trioxycinnamic
acids. The trimethoxy unit of colchicine is thought19 to be
derived via this pathway. It is conceivable that these antimitotic
natural products possess a trimethoxy aryl unit as a con-
sequence of their biogenesis. Similar sized functionalities on
the A ring of the combretastatins have not been explored and
we considered it possible that such derivatives might have
comparable biological activity to the natural products. Herein
are described the synthesis and biological activities of a
range of combretastatins where the trimethoxy motif has been
substituted by other functionalities.
Results and discussion
Synthesis
Fig. 1 Structures of combretastatin A-4 (1), colchicine (2), steganacin
(3) and podophyllotoxin (4).
A series of previously unknown trimethoxy substituted stil-
benes were synthesised as well as a series of triethoxy, trifluoro,
dimethyl, trimethyl and unsubstituted A ring combretastatins.
In most cases a Wittig reaction of the appropriate phos-
phonium bromide with the required arylaldehyde afforded a
† Electronic supplementary information (ESI) available: experimental
details. See http://www.rsc.org/suppdata/ob/b3/b306878a/
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 0 3 3 – 3 0 3 7
3033

View Article Online
mixture of Z-(a series) and E-stilbenes (b series) which were
separated by column chromatography. The stereochemistry of
the olefins was determined from the coupling constants in the
proton NMR (for E compounds J ≈ 16 Hz, for Z compounds
J ≈ 12 Hz). In the case of silyl ether protected phenols, fluoride
was used to generate the required carbinol. These methods
produced the trimethoxy 6–11 (Scheme 1), triethoxy 13–18
(Scheme 2), trifluoro 22–25 (Scheme 3), dimethyl 28–36
(Scheme 4) and unsubstituted A ring combretastatins 38–41
(Scheme 5).
The trifluorostilbene 25 was prepared via cinnamic acid 26
using a Perkin-type condensation reaction (Schemes 6–8). The
stereochemistry of this acid 26 was determined using UV
spectroscopy. The extinction coefficients for stilbenes with the
aryl groups cis to each other tend to be in the region of 10,000
Scheme
4
Synthesis of dimethylstilbenes 28–36. Reagents and
conditions: (i) n-BuLi–THF Ϫ20 ЊC; (ii) RCHO Ϫ20 ЊC to rt.
Scheme 5 Synthesis of unsubstituted stilbenes 38–41. Reagents and
conditions: (i) n-BuLi–THF Ϫ20 ЊC; (ii) RCHO Ϫ20 ЊC to rt.
Scheme
1
Synthesis of trimethoxystilbenes 6–11. Reagents and
conditions: (i) n-BuLi–THF Ϫ20 ЊC; (ii) RCHO Ϫ20 ЊC to rt.
Scheme 6 Synthesis of benzaldehydes 48 and 49. Reagents and
conditions: (i) 1 M NaOH–octafluorotoluene or pentafluoropyridine.
Scheme
2
Synthesis of triethoxystilbenes 13–18. Reagents and
conditions: (i) n-BuLi–THF Ϫ20 ЊC; (ii) RCHO Ϫ20 ЊC to rt.
Scheme
3
Synthesis of trifluorostilbenes 22–25. Reagents and
conditions: (i) ArCH2PPh3Br–n-BuLi Ϫ20 ЊC; (ii) 4-methoxyphenyl-
acetic acid–Ac2O–Et3N; (iii) Cu–quinoline 200 ЊC.
Scheme 7 Synthesis of trimethylstilbenes 50, 51, 53 and 54. Reagents
and conditions: (i) RCHO–Ac2O–Et3N; (ii) Cu–quinoline 200 ЊC.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 0 3 3 – 3 0 3 7
3034

View Article Online
Table 1 Cytotoxicity (K562 cells, HUVECs/µM), inhibition of tubulin
assembly (MA/µM), colchicine displacement (CD/µM) of trimethoxy-
stilbenes 7, 8, 10 and 11. ND means not determined
Compound
IC50 (K562)
IC50 MA
IC50 CD
IC50 HUVEC
1
0.001
0.04
35.8
0.001
7.83
0.01
0.14
22
0.175
2
>10
0.4
>10
0.085
>10
>10
3
>25
>25
10
>25
2.8
0.0026
ND
ND
0.067
ND
0.004
ND
7a
7b
8a
8b
10a
11a
11b
>25
>25
ND
Table 2 Cytotoxicity (K562 cells, HUVECs/µM), inhibition of tubulin
assembly (MA/µM), colchicine displacement (CD/µM) of triethoxy
stilbenes 13–16 and 18. ND means not determined
Compound
IC50 (K562)
IC50 MA
IC50 CD
IC50 HUVEC
Scheme 8 Synthesis of trimethylstilbene 52. Reagents and conditions:
(i) NaOMe–DMF.
13a
13b
14a
14b
15a
15b
16a
16b
18a
18b
0.6
>50
0.5
>50
0.044
>50
0.45
7
>10
>10
>10
>10
1.25
>10
>10
>10
0.50
>10
>25
>25
>25
>25
>25
>25
>25
>25
15.5
>25
ND
ND
ND
ND
0.19
ND
ND
ND
0.05
ND
whilst their trans isomers have extinction coefficients of 30,000.
NOESY spectra of similar acids (reduced to carbinols) have
also elucidated the stereochemistry of this type of cinnamic
acid.18 Copper-mediated decarboxylation18 of 26 afforded
a separable E/Z mixture (4 : 5) of the trifluorostilbenes 25.
Similarly the trimethylstilbenes 50,51 were prepared from
3,4,5-trimethylphenylacetic acid (42) via the cinnamic acids
43,44. However reaction of acid 42 with 3-hydroxy-4-methoxy-
benzaldehyde failed to produce the required cinnamic acid 45.
The perfluorinated aryl ethers 46,47 were prepared from
reaction of acid 42 with the aldehydes 48,49. Decarboxylation
of 46,47 provided the O-protected Z-stilbenes 53a,54a, respect-
ively. Treatment of the perfluorotolyl ether 53a with sodium
methoxide failed to give phenol 52a. However similar treatment
of the perfluoropyridyl ether did afford the required phenolic
stilbene 52a.
0.018
0.2
Table 3 Cytotoxicity (K562 cells/µM), inhibition of tubulin assembly
(MA/µM), colchicine displacement (CD/µM) of trifluorostilbenes
22–25
Compound
IC50 (K562)
IC50 MA
IC50 CD
22a
22b
23a
24a
24b
25a
25b
0.12
22
>50
0.13
5
4.2
40
5.5
>10
>10
4.5
>10
>10
>10
>25
>25
>25
>25
>25
>25
>25
Biochemistry
The sets of stilbenes were tested for cytotoxicity against the
human leukaemia K562 cell line and for their ability to interact
with tubulin using an assembly and colchicine displacement
assay.
Table 4 Cytotoxicity (K562 cells/µM), inhibition of tubulin assembly
(MA/µM), colchicine displacement (CD/µM) of stilbenes 38–41
In the trimethoxy series it has been noted3,7 that the
4-position on the B ring can be a small group (e.g. methoxy,
methyl, ethyl or dimethylamino). However, no A ring tri-
methoxy stilbene with just a proton on the 4-position of the B
ring had previously been tested for antimitotic activity. Phenol
11a described here is similar to the lead compound com-
bretastatin A-4 (1); it simply lacks the B ring 4-methoxy group.
This phenol 11a was, however, considerably less cytotoxic than
combretastatin A-4 (1) and neither inhibited tubulin assembly
or could displace colchicine from its binding site on tubulin.
The dimethyl compound 7 showed moderate cytotoxicity, but
failed to interact with tubulin. The halogenated compounds
8a,10a showed potent cytotoxicity as well as the ability to
interact with tubulin. Interestingly the fluoro derivative 10a,
although ten times less cytotoxic than combretastatin A-4 (1),
was more potent in the tubulin assays.
Using an in vitro method to measure the potential ability of
agents to interact with the vascular system, the halogenated
stilbenes 8a,10a were tested for their ability to inhibit the
growth of human umbilical vein endothelial cells (HUVECs).
This assay measures the permeability of HUVECS after treat-
ment with possible antivascular agents. Increased permeability
is indicative of the ‘rounding up’ of the cells which, in vivo,
would lead to blockage of the tumour vasculature.20 In this
assay the fluoro compound 10a was only slightly less active
than combretastatin A-4 (1) (Table 1).
Compound
IC50 (K562)
IC50 MA
IC50 CD
38a
38b
39a
39b
40a
40b
41a
41b
5
14
2.5
>50
0.5
40
>10
>10
>10
>10
4.75
>10
2.5
>25
>25
>25
>25
>25
>25
>25
>25
0.5
1
>10
In the triethoxy series, only the phenol 18a and fluoro
compound 15a showed good cytotoxicity and the ability to
interact with tubulin. The phenol 18a was also particularly
potent in the HUVEC assay (Table 2).
The unsubstituted 38–41 and trifluoro substituted stilbenes
22–25 were synthesised as lipophilic alternatives to the standard
trimethoxy substituted compounds. None of these agents
exhibited high cytotoxicity, although in the trifluoro series the B
ring fluoro and hydroxy compounds 22a, 24a showed a small
interaction with tubulin. Similarly, in the unsubstituted series,
the phenol 41a exhibited a slight interaction with tubulin
(Tables 3 and 4).
The dimethyl substituted stilbenes, in general, showed poor
cytotoxicity and little interaction with tubulin. However, the
phenol 36a, and the fluoro compound 32a did exhibit consider-
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 0 3 3 – 3 0 3 7
3035

View Article Online
able cytotoxicity to the K562 cell line and some ability to
interact with tubulin (Table 5). In the trimethyl series all the
agents 50a,51a,54a were less cytotoxic than combretastatin
A-4 (1). Interestingly, the phenolic compound 52a, which
has the B ring substitution pattern of combretastatin A-4 (1),
is twenty-fold less toxic than combretastatin A-4 (1) in
the cytotoxicity assay but more active in the tubulin assay
(Table 6).
Cell cycle analysis was carried out on those agents which
had shown some activity in the tubulin assays. Antimitotic
agents tend to accumulate their DNA in the G2/M phase of
the cell cycle. All the compounds in Table 7 showed at least
a moderate ability to block cells in the G2/M phase. As well as
combretastatin A-4 1, the fluoro and phenolic compounds
10a,15a,18a,51a,54a and the stilbene 50a showed potent ability
to block cells in the G2/M phase of the cell cycle.
Immunohistochemistry was used to visualise the effects of
several of these agents on microtubules of VERO cells. In the
control (Fig. 2a), the dark nuclei are clearly visible with intense
staining occurring around the perinuclear area from which
the microtubule network radiates out into the cytoplasm. When
the cells were treated with combretastatin A-4 (1), a complete
loss of fine structure was noticeable (Fig. 2b) and the staining
became almost uniform over the cellular cytoplasm. The cells
round-up and detach from their substratum. The immuno-
histochemistry sections for the trimethyl agents 50a–52a reflect
the results obtained in the cytotoxicity assays. The microtubule
structures in the cells which have been treated with 50a (Fig. 2c)
and 51a (Fig. 2d) are clearly visible and undamaged. However,
in the cells treated with phenol 52a (Fig. 2e), the microtubules
show a similar disruption pattern to those treated with com-
bretastatin A-4 (1). The triethoxy agents 15a (Fig. 2f ) and 18a
(Fig. 2g) also disrupt the microtubule structures in VERO cells.
The microtubule structures (Fig. 2h) in the cells which have
been treated with dimethyl phenol 36a show some disruption
but not to the extent of its trimethyl equivalent 52a. Interest-
ingly following treatment with 36a, there appear to be several
star-like structures present around the nuclei which may be
asters. Asters are structures which are often seen when micro-
tubules are allowed to repolymerise following treatment with
a reversible antimitotic agent and it is from these structures
that microtubule organising centres act as the nucleus for
microtubule growth.
Conclusions
It is already well-documented that the Z-stilbenes are more
potent microtubule interactors than their E-counterparts.1–3,7,14
This fact is again borne out in this study. In this study the
agents which have shown the greatest ability to interact with
tubulin have been those possessing a 3-hydroxy substituent on
the B ring. This again tallies with previous findings that small
groups on the 4-position of the B ring along with a vacant or
3-hydroxy substituent are essential for microtubular interaction.
At present combretastatin A-4 (1) is in clinical trial as its
water-soluble disodium phosphate salt. This phosphate has a
short plasma half-life and dephosphorylates10,21 to the phenol
1. The discovery herein, that the B ring fluorinated stilbenes
10a,15a,51a also show potent activities, may avoid this type of
event.
The binding of colchicine (2) and combretastatin A-4 (1) to
tubulin is reversible; so it is unlikely that a covalent bond is
formed between agents 1, 2 and the protein. One way in which
the trimethoxy unit on combretastatin A-4 1 may interact with
its binding site on tubulin is through the oxygen atoms
hydrogen bonding with the protein. This tallies with the results
described herein for the triethoxy compounds 15a and 18a.
However the microtubule disruption caused by the dimethyl
36a and more significantly by the trimethyl substituted agent
52a, indicates that hydrogen bonding of the A ring to tubulin is
unlikely to be a significant mechanism of interaction.
Table 5 Cytotoxicity (K562 cells, µM), inhibition of tubulin assembly
(MA, µM), colchicine displacement (CD, µM) of dimethylstilbenes
28–34 and 36
Compound
IC50 (K562)
IC50 MA
IC50 CD
28a
28b
29a
29b
30a
30b
31a
31b
32a
32b
33a
33b
34a
34b
36a
3.4
>50
1.8
24
1.6
>50
1.15
>50
0.09
2.5
0.39
>50
1.5
>10
>10
>10
>10
>10
>10
>10
>10
6.57
>10
>10
>10
>10
>10
3.09
>25
>25
>25
>25
>25
>25
>25
>25
>25
>25
>25
>25
>25
>25
12.5
>50
0.07
Table 6 Cytotoxicity (K562 cells/µM), inhibition of tubulin assembly
(MA/µM), colchicine displacement (CD/µM) of trimethylstilbenes
50–52
What is particularly exciting about agent 10a is that it is ten
times less cytotoxic than combretastatin A-4 (1), but more
active in depolymerising microtubules. Of the non-trimethoxy
agents the trimethylphenol 52a is again less toxic (twentyfold)
than combretastatin A-4 (1) and again more active in inhibiting
the assembly of tubulin.
Compound
IC50 (K562)
IC50 MA
IC50 CD
IC50 HUVEC
50a
51a
52a
0.31
0.14
0.020
0.650
0.700
0.120
>25
10
10
>1
0.46
0.05
Table 7 Cell cycle analysis of K562 cells treated with stilbenes 1, 10, 15, 18, 24, 32, 36, 41, 50, 51 and 52
% cells with DNA content ≥ 2n
Compound
% cells with DNA content < 2n
% cells in G0–G1 phase
% cells in S phase
% cells in G2M phase
1
7
5
24
4
38
28
31
25
15
23
6
3
5
7
6
16
18
6
39
33
27
24
10
17
6
92
79
75
92
49
57
65
68
87
77
91
10a
15a
18a
24a
32a
36a
41a
50a
51a
52a
3
12
10
8
8
3
6
3
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 0 3 3 – 3 0 3 7
3036

View Article Online
Fig. 2 (a) Immunohistochemistry sections of VERO cells [Control (DMSO)]. (b) After treatment with Combretastatin A-4 1. (c) After treatment
with trimethylstilbene 50a. (d) After treatment with trimethylstilbene 51a. (e) After treatment with phenolic trimethylstilbene 52a. (f ) After treatment
with triethoxystilbene 15a. (g) After treatment with triethoxystilbene 18a. (h) After treatment with phenolic dimethylstilbene 36a.
9 A. T. McGown and B. W. Fox, Anti-Cancer Drug Des., 1989, 3, 249.
10 G. M. Tozer, C. Kanthou, C. S. Parkins and S. A. Hill, Int. J. Exp.
Acknowledgements
Pathol., 2002, 83, 21.
The authors would like to thank the Association for
International Cancer Research and Cancer Research UK
for supporting this study.
11 G. R. Pettit, S. B. Singh, M. R. Boyd, E. Hamel, R. K. Pettit,
J. M. Schmidt and F. Hogan, J. Med. Chem., 1995, 38, 1666.
12 G. G. Dark, S. A. Hill, V. E. Prise, G. M. Tozer, G. R. Pettit and
D. J. Chaplin, Cancer Res., 1997, 57, 1829.
13 Z. Getahun, L. Jurd, P. S. Chu, C. M. Lin and E. Hamel, J. Med.
Chem., 1992, 35, 1058.
14 J. M. Andreu, B. Perez-Ramirez, M. J. Gorbunoff, D. Ayala and
S. N. Timasheff, Biochemistry, 1998, 37, 8356.
References
1 J. A. Hadfield, S. Ducki, N. Hirst and A. T. McGown, Prog. Cell
Cycle Res., 2003, 5, 309.
2 A. M. Jordan, J. A. Hadfield, N. J. Lawrence and A. T. McGown,
Med. Res. Rev., 1998, 259–96.
3 M. Cushman, D. Nagarathnam, D. Gopal, H-M. He, C. M. Lin and
E. Hamel, J. Med. Chem., 1992, 35, 2293.
4 M. Cushman, D. Nagarathnam, D. Gopal, A. K. Chakraborti,
C. M. Lin and E. Hamel, J. Med. Chem., 1991, 34, 2579.
5 J. A. Hadfield and A. T. McGown, Synth. Commun., 1998, 28,
1421.
15 Q. Shi, K. Chen, S. L. Morris-Natschke and K-H. Lee, Curr. Pharm.
Des., 1998, 4, 219.
16 S. Zhao, J. V. Moore, M. L. Waller, A. T. McGown, J. A. Hadfield,
G. R. Pettit and D. L. Hastings, Eur. J. Nucl. Med., 1999, 26, 231.
17 N. J. Lawrence, F. A. Ghani, L. A. Hepworth, J. A. Hadfield,
A. T. McGown and R. G. Pritchard, Synthesis, 1999, 1656.
18 K. Gaukroger, J. A. Hadfield, L. A. Hepworth, N. J. Lawrence and
A. T. McGown, J. Org. Chem., 2001, 66, 8135.
6 K. Aleksandrzak, A. T. McGown and J. A. Hadfield, Anti-Cancer
Drugs, 1998, 9, 545.
7 J. A. Woods, J. A. Hadfield, A. T. McGown, G. R. Pettit and
B. W. Fox, Br. J. Cancer, 1995, 71, 705.
8 R. T. Brown, B. W. Fox, J. A. Hadfield, A. T. McGown,
S. P. Mayalarp, G. R. Pettit and J. A. Woods, J. Chem. Soc., Perkin
Trans. 1, 1995, 577.
19 J. Mann in Natural Products Their Chemistry and Biological
Significance, Longman, UK, 1984, pp. 419–20.
20 M. E. Watts and M. Woodcock, Eur. J. Cancer, 1992, 28A, 1628.
21 A. Dowlati, K. Robertson, M. Cooney, W. P. Petros, M. Stratford,
J. Jesberger, N. Rafie, B. Overmoyer, V. Makkar, B. Stambler,
A. Taylor, J. Waas, J. S. Lewin, K. R. McCrae and S. C. Remick,
Cancer Res., 2002, 62, 3408.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 0 3 3 – 3 0 3 7
3037