Rationally simplified bistramide analog reversibly targets actin polymerization and inhibits cancer progression in vitro and in vivo

Article abstract of DOI:10.1021/ja101811x

We describe structure-based design and chemical synthesis of a simplified analog of bistramide A, which potently and reversibly binds monomeric actin with a Kd of 9.0 nM, depolymerizes filamentous actin in vitro and in A549 (nonsmall cell lung cancer) cells, inhibits growth of cancer cell lines in vitro at submicromolar concentrations, and significantly suppresses proliferation of A549 cells in a nude mice tumor xenograft model in terms of both tumor growth delay and average tumor volume. This study provides a conceptual framework for the design and development of new antiproliferative compounds that target cytoskeletal organization of cancer cells in vivo by a combination of reversible G-actin binding and effective F-actin severing.

Full text of DOI:10.1021/ja101811x

Published on Web 05/10/2010
Rationally Simplified Bistramide Analog Reversibly Targets Actin
Polymerization and Inhibits Cancer Progression in Vitro and in Vivo
Syed Alipayam Rizvi,† Song Liu,† Zhonglei Chen,† Colleen Skau,‡ Matthew Pytynia,§
David R. Kovar,‡ Steven J. Chmura,§ and Sergey A. Kozmin*,†
UniVersity of Chicago, Department of Chemistry, Department of Molecular Genetics and Cell Biology, Department
of Cellular and Radiation Oncology, Chicago, Illinois 60637
Received March 3, 2010; E-mail: skozmin@uchicago.edu
Cytoskeletal organization is strongly altered upon malignant
transformation and tumor invasion, which includes significant
changes in expression of actin and actin-binding proteins.1 Small
molecules that target the actin cytoskeleton could be potentially
employed to study such alterations in ViVo and ultimately combat
cancer.2 Bistramides comprise a polyketide-based family of natural
products with potent antiproliferative activities, which was isolated
from tunicates L. bistratum and T. cyclops.3-6 Chemical synthesis
of bistramide A (1, Figure 1)7,8 enabled subsequent determination
of the mechanism of action, which entailed specific modulation of
the actin cytoskeleton9,10 and covalent modification of the protein
target.11 Importantly, the actin-severing activity of this agent was
not dependent on the covalent protein modification.11 This observa-
tion suggested that rationally designed compounds, which do not
react covalently with actin, may be able to target the actin
cytoskeletion of cancer cells in ViVo. Indeed, bistramides D and K
that lack the reactive enone subunit were shown to inhibit
proliferation of nonsmall cell broncho-pulmonary carcinoma im-
planted in nude mice.12 In this communication, we describe
structure-based design and synthesis of a simplified analog of
Figure 1. Design of a simplified analog of bistramide A. Liphophilic
subunit A and hydrophilic subunit B are retained to preserve actin-binding
and severing activity. Chemically reactive subunit C is eliminated to
minimize in ViVo toxicity.
bistramide A, which reversibly binds monomeric actin, efficiently
depolymerizes filamentous actin, inhibits growth of cancer cell lines
in Vitro, and suppresses A549 (nonsmall cell lung cancer) tumor
growth in ViVo.
interactions of 2 with actin, which would in turn maximize the
potency of this compound.
The structure of bistramide A can be dissected into three main
fragments, including C(19)-C(38) spiroketal-bearing subunit A,
C(13)-C(18) bisamide subunit B, and C(1)-C(13) enone-contain-
ing subunit C (Figure 1). Subunits A and B are required for binding
of the natural product to its target by forming extended networks
of hydrogen-bonding and van der Waals interactions, respectively.10
The 1,3-enone moiety present in subunit C is responsible for
covalent interaction of bistramide A with G-actin but not required
for actin-severing activity.11 Thus, removal of the highly reactivity
enone-containing subunit C was expected to provide access to a
simplified congener 2 (Figure 1) that could reversibly interact with
the actin cytoskeleton and inhibit tumor growth in Vitro and in ViVo.
The main challenge, however, was to preserve the potent bioactivity
of this chemical agent while significantly simplifying the structure
of the parent natural product. Indeed, while we previously
demonstrated that structural simplifications of the bistramide
framework were possible, such modifications typically decreased
actin-binding and actin-severing activity of the resulting analogs,
in some cases substantially.11 Retention of the fully elaborated
spiroketal subunit A, as well as the intact central amide fragment
B, was expected to preserve the majority of hydrophobic and polar
The assembly process followed our general strategy for rapid
polyketide assembly7,13,14 starting with ring-opening of cyclopro-
pene 4 with alkene 3 in the presence of the Grubbs catalyst 5 (Figure
2A).15 Removal of the acetal followed by cross-metathesis with
alkene 6 delivered dienone 7. Hydrogenation with concomitant
removal of both benzyl and silyl protecting groups and chemose-
lective installation of the phthalimide moiety, followed by
Dess-Martin oxidation, afforded aldehyde 8; Cr-based olefination
and oxazoborolidine reduction afforded phthalimide 9. The endgame
of the synthesis entailed removal of the phthalimide group with
methyl amine, PyBOP-promoted coupling with acid 10,7 Fmoc-
deprotection, and amide coupling using activated ester 12. This
convergent sequence efficiently produced the target compound 2
(11 steps from alkene 3). It is noteworthy that while the longest
linear sequence of the assembly process is comparable to that in
our initial synthesis of bistramide A,7 the total number of steps
required for construction of 2 is substantially smaller due to
simplification of the subunit C.
Isothermal titration calorimetry established the dissociation
constant (Kd) value of 9.0 nM for binding of 2 to G-actin, which
was comparable to that of the parent natural product 1 (Kd ) 6.8
nM)9 and proved superior to other synthetic bistramide deriva-
tives.11 The high actin-binding affinity of 2 was attributed to the
† Department of Chemistry.
‡ Department of Molecular Genetics and Cell Biology.
§ Department of Cellular and Radiation Oncology.
9
7288 J. AM. CHEM. SOC. 2010, 132, 7288–7290
10.1021/ja101811x  2010 American Chemical Society

C O M M U N I C A T I O N S
Figure 2. (A) Synthesis of a structurally simplified bistramide analog. This sequence was employed to produce a sufficient amount of 2 (110 mg) for all
subsequent studies. (B) Visualization of severing of actin filaments in Vitro by time-lapse TIRF. Time-lapse micrographs with time in seconds indicated in
the upper right. Black arrow at time 0 indicates introduction of one flowcell volume of (top) buffer only or (bottom) buffer containing 1.5 µM of bistramide
analog 2. White arrowheads indicate break points in the filaments. Scale bar )2.5 µm. Movies are available as Supporting Information. (C) Time- and
dose-dependent growth inhibition of A549 cells by 2. Cell growth was measured using an ATP-monitoring luciferase-based assay. (D) Growth inhibition of
A549 tumors in athymic mice by compound 2, which was administered i.p. in three doses of 20 mg/kg each on third, fifth, and seventh day after the initial
tumor challenge. (E) Depolymerization of F-actin in A549 treated with 2 for 2 h at variable concentrations of the drug shown. Following treatment with 2,
cells were fixed with paraformaldehyde, stained with DAPI (nuclei, blue) and Alexa-fluor 488 phalloidin (actin, green), and imaged using a Leica TCS SP2
AOBS Laser Scanning Confocal microscope. Scale bar )20 µm.
van der Waals interactions between several lipophilic amino acid
side chains of actin and a fully extended C(31)-C(38) terminus of
the spiroketal-bearing subunit A. This result further confirmed that
pyran-containing fragment C did not significantly contribute to the
actin-binding affinity of bistramide-based agents.
fluorescence (TIRF) microscopy to examine the effects of this
compound on depolymerization of filamentous actin in Vitro.16 We
initially assembled individual actin filaments by elongating them
from a pool of 1.0 µM Mg-ATP-actin monomers supplemented
with 0.5 µM Mg-ATP-actin monomers labeled with Oregon green
for visualization.17 We then flowed into the observation chamber
either a solution of buffer or a 1.5 µM solution of 2 and imaged
Encouraged by the initially observed potent G-actin binding
affinity of 2, we next employed time-lapse total internal reflection
9
J. AM. CHEM. SOC. VOL. 132, NO. 21, 2010 7289

C O M M U N I C A T I O N S
with TIRF microscopy (Figure 2B). As expected, addition of buffer
alone did not substantially impact filament depolymerization. On
the other hand, treatment of F-actin with 2 resulted in rapid filament
disassembly. The mechanism of actin depolymerization by 2
entailed formation of multiple filaments breaks, which created new
free barbed ends that rapidly disassemble. This actin-severing
mechanism was identical to that displayed by the parent natural
product.11
We next evaluated the effect of 2 on proliferation of a range of
human cancer cell lines grown in culture. This analysis revealed
several cell lines that elicited increased sensitivity toward the
growth-inhibitory action of this agent, including A549 cells. We
next examined the effects of prolonged exposure of 2 on prolifera-
tion of this nonsmall lung cancer cell line after 24, 48, 72, and
96 h following initial incubation with this agent (Figure 2C). Cell
growth was monitored by measuring the amount of ATP produced
using a well-established luciferase-based protocol. This study
revealed efficient, dose-dependent growth inhibition. Indeed, the
growth of the A549 nonsmall cell lung cancer cell line was inhibited
by 70% at 0.65 µM of 2 and by 85% at 2.5 µM of 2.
Our work demonstrated that a detailed knowledge of biochem-
istry of the bistramide family of natural products enabled rational
design and practical chemical synthesis of a simplified chemical
agent 2, which proved to be highly effective at depolymerizing
filamentous actin and inhibiting cancer growth in Vitro and in ViVo.
Our studies provide a conceptual framework for the design and
development of new antiproliferative compounds that target cy-
toskeletal organization of cancer cells by reversible binding to
monomeric actin and effective severing of actin filaments. This work
sets the stage for continued comprehensive pharmacological evalu-
ation of this class of actin-targeting agents.18
Acknowledgment. Financial support of this work was provided
by the American Cancer Society (RSG-04-017-CDD).
Supporting Information Available: Experimental details and
movies related to time lapses shown in Figure 2. This information is
available free of charge via the Internet at http://pubs.acs.org.
References
(1) Van Troys, M.; Vandekerckhove, J.; Ampe, C. Protein ReV. 8; Springer
Science and Business Media, LLC: New York, NY, 2008.
(2) (a) Fenteany, G.; Zhu, S. Curr. Top. Med. Chem. 2003, 3, 593. (b) Spector,
I.; Braet, F.; Shochet, N. R.; Bubb, M. R. Microsc. Res. Tech. 1999, 47,
18. (c) Allingham, J. S.; Klenchin, V. A.; Rayment, I. Cell. Mol. Life Sci.
2006, 63, 2119.
(3) Gouiffe`s, D.; Moreau, S.; Helbecque, N.; Bernier, J. L.; Henichart, J. P.;
Barbin, Y.; Laurent, D.; Verbist, J. F. Tetrahedron 1988, 44, 451–459.
(4) Degnan, B. M.; Hawkins, C. J.; Lavin, M. F.; McCaffrey, E. J.; Parry,
D. L.; Watters, D. J. J. Med. Chem. 1989, 32, 1354–1359.
(5) Biard, J. F.; Roussakis, C.; Kornprobst, J. M.; Gouffes-Barbin, D.; Verbist,
J. F. J. Nat. Prod. 1994, 57, 1336–1345.
We also examined the ability of 2 to depolymerize F-actin in
live cells. The actin cytoskeleton can be readily visualized by
fluorescent microscopy following Alexa Fluor 488 phalloidin
staining. This experiment revealed that 2 depolymerized F-actin in
A549 cells in a dose-dependent manner (Figure 2E). Importantly,
noticeable actin disassembly was detected even at 250 nM of 2,
which further verified the notable potency of this agent. A similar
effect of 2 on F-actin depolymerization was also observed in the
DU145 prostate cancer cell line (not shown).
(6) Murphy, B. T.; Cao, S.; Brodie, P.; Maharavo, J.; Andriamanantoanina,
H.; Ravelonandro, P.; Kingston, D. G. I. J. Nat. Prod. 2009, 72, 1338–
1340.
Finally, we examined the ability of 2 to inhibit the growth of
A549 cells in ViVo (Figure 2D). Initially, we established that 2 had
no observed toxicity in mice up to a 50 mg/kg single i.p. dose or
3 × 20 mg/kg i.p. using 3 mice per group with escalating doses
starting at 1/10 of those employed for subsequent in ViVo experi-
ments. Toxicity was assessed by daily weight measurements and
mouse behavior compared to noninjected mice. No weight loss
(>5%) or behavioral changes were observed. Next, we established
tumor xenografts by injecting 1 × 107 A549 cells subcutaneously
into two groups of athymic mice. The first group (10 animals) was
injected only with the DMSO control. The second group of 12
animals received three i.p. doses of analog 2 (20 mg/kg each) on
the third, fifth and seventh day after the initial tumor challenge
using the same amount of DSMO as the control group. Tumor
growth was monitored over a period of five subsequent weeks every
other day. We observed exponential tumor growth in the control
mice population, which reached an average size of 731 mm3 at the
end of 5 weeks. Tumor growth was substantially inhibited by analog
2. The average tumor size for this population was 308 mm3 after
5 weeks. This result is particularly noteworthy since 2 was
administered to animals only during the first week after the initial
tumor challenge and no drug was given during the last 5 weeks of
the study. These studies demonstrate both a tumor growth delay of
2 weeks and a significant reduction in tumor growth using regression
analysis (p < 0.001).
(7) Statsuk, A. V.; Liu, D.; Kozmin, S. A. J. Am. Chem. Soc. 2004, 126, 9546–
9547.
(8) For other synthetic approaches of bistramides, see: (a) Wipf, P.; Uto, Y.;
Yoshimura, S. Chem.sEur. J. 2002, 8, 1670–1681. (b) Wipf, P.; Hopkins,
T. D. Chem. Commun. 2005, 3421–3423. (c) Crimmins, M. T.; DeBaillie,
A. C. J. Am. Chem. Soc. 2006, 128, 4936–4937. (d) Lowe, J. T.; Wrona,
I. E.; Panek, J. S. Org. Lett. 2007, 9, 327–330. (e) Yadav, J. S.; Chetia, L.
Org. Lett. 2007, 9, 4587–4589. (f) Wrona, I. E.; Lowe, J. T.; Turbyville,
T. J.; Johnson, T. R.; Beignet, J.; Beutler, J. A.; Panek, J. S. J. Org. Chem.
2009, 74, 1897–1916.
(9) Statsuk, A. V.; Bai, R.; Baryza, J. L.; Verma, V. A.; Hamel, E.; Wender,
P. A.; Kozmin, S. A. Nat. Chem. Biol. 2005, 1, 383–388.
(10) Rizvi, S. A.; Tereshko, V.; Kossiakoff, A. A.; Kozmin, S. A. J. Am. Chem.
Soc. 2006, 128, 3882–3883.
(11) Rizvi, S. A.; Courson, D. S.; Keller, V. A.; Rock, R. S.; Kozmin, S. A.
Proc. Natl. Acad. Sci. USA. 2008, 105, 4088–4092.
(12) Riou, D.; Roussakis, C.; Biard, J. F.; Verbist, J. F. Anticancer Res. 1993,
13, 2331–1334.
(13) Marjanovic, J.; Kozmin, S. A. Angew. Chem., Int. Ed. 2007, 46, 9010–
9013.
(14) Matsumoto, K.; Kozmin, S. A. AdV. Synth. Catal. 2008, 350, 557–560.
(15) Chatterjee, A. K.; Morgan, J. P.; Scholl, M.; Grubbs, R. H. J. Am. Chem.
Soc. 2000, 122, 3783–3784.
(16) (a) Amann, K. J.; Pollard, T. D. Proc. Natl. Acad. Sci. U.S.A. 2001, 98,
15009–15013. (b) Kovar, D. R.; Pollard, T. D. Proc. Natl. Acad. Sci. U.S.A.
2004, 101, 14725–14730. (c) Kovar, D. R.; Harris, E. S.; Mahaffy, R.;
Higgs, H. N.; Pollard, T. D. Cell 2006, 124, 423–435.
(17) Neidt, E. M.; Skau, C. T.; Kovar, D. R. J. Biol. Chem. 2008, 283, 23872–
23883.
(18) Compound 2 has been selected for preclinical pharmacological study by
the National Cancer Institute’s (NCI’s) Rapid Access to Interventional
Development (RAID) program; NSC Number: D751592-H.
JA101811X
9
7290 J. AM. CHEM. SOC. VOL. 132, NO. 21, 2010