Synthesis of 2-alkoxy and 2-benzyloxy analogues of estradiol as anti-breast cancer agents through microtubule stabilization

Article abstract of DOI:10.1016/j.bmc.2014.01.012

2-Methoxyestradiol (2ME2) is an investigational anticancer drug. In the present study, 2-alkoxyesters/acid and 2-benzyloxy analogues of estradiol have been synthesized as analogues of 2ME2. Three of the derivatives exhibited significant anticancer activity against human breast cancer cell lines. The best analogue of the series i.e. 24 showed stabilization of tubulin polymerisation process. It was substantiated by confocal microscopy and molecular docking studies where 24 occupied 'paclitaxel binding pocketg€ to stabilize the polymerisation process. Compound 24 significantly inhibited MDA-MB-231 cells (IC₅₀: 7 μM) and induced arrest of cell cycle and apoptosis in MDA-MB-231 cells. In acute oral toxicity, 24 was found to be non-toxic and well tolerated in Swiss albino mice up to 1000 mg/kg dose.

Full text of DOI:10.1016/j.bmc.2014.01.012

Bioorganic & Medicinal Chemistry xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Review
Taccalonolide microtubule stabilizers
Jing Li a,c, April L. Risinger a,c, Susan L. Mooberry a,b,c,
⇑
a
Department of Pharmacology, University of Texas Health Science Center at San Antonio, TX 78229, United States
Department of Medicine, University of Texas Health Science Center at San Antonio, TX 78229, United States
Cancer Therapy & Research Center, University of Texas Health Science Center at San Antonio, TX 78229, United States
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online xxxx
This review focuses on a relatively new class of microtubule stabilizers, the taccalonolides. The taccalon-
olides are highly oxygenated pentacyclic steroids isolated from plants of the genus Tacca. Originally iden-
tified in a cell-based phenotypic screen, the taccalonolides have many properties similar to other
microtubule stabilizers. They increase the density of interphase microtubules, causing microtubule bun-
dling, and form abnormal multi-polar mitotic spindles leading to mitotic arrest and, ultimately, apopto-
sis. However, the taccalonolides differ from other microtubule stabilizers in that they retain efficacy in
taxane resistant cell lines and in vivo models. Binding studies with the newly identified, potent taccalon-
olide AJ demonstrated covalent binding to b-tubulin at or near the luminal and/or pore taxane binding
site(s) which stabilizes microtubule protofilaments in a unique manner as compared to other microtu-
bule stabilizers. The isolation and semi-synthesis of 21 taccalonolides helped to identify key structure
activity relationships and the importance of multiple regions across the taccalonolide skeleton for opti-
mal biological potency.
Keywords:
Tacca
Taccalonolide
Microtubule
Tubulin
Microtubule stabilizer
Taxol
Paclitaxel
Laulimalide
Ó 2014 Elsevier Ltd. All rights reserved.
Contents
1
2
3
4
5
6
7
8
9
1
.
.
.
.
.
.
.
.
.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
Chemical isolation and structure elucidation of natural taccalonolides A–Y . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
Identification of the taccalonolides A and E as microtubule stabilizers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
Mechanistic uniqueness of the taccalonolides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
Isolation and semi-synthesis of the potent new taccalonolides AF and AJ and their direct effects on tubulin polymerization . . . . . . . . . . . . . . . 00
Identification of the taccalonolide binding site on tubulin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
In vivo antitumor effects of taccalonolides AF and AJ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
Optimization of the C15 hydrolysis reaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
Isolation of new natural taccalonolides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
0. Structure activity relationships of the taccalonolides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
10.1. E-ring constituents at C20–C23 play a key role in taccalonolide potency. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
10.2. A large, bulky group at C1 provides a significant increase in potency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
10.3. There are complex indications with regard to the importance of the C5 hydroxy group to potency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
10.4. The C5-8 substituents on the lower portion of the B ring can affect taccalonolide potency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
11. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
References and notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00
1
. Introduction
Microtubule disrupting agents have been isolated from a wide
variety of sources, including marine sponges, mycobacteria, cyano-
bacteria, and plants. Moreover, new microtubule active com-
pounds with significant biological activities continue to be
discovered from nature. Plant-derived compounds that destabilize
microtubules include the vinca alkaloids, colchicine and the com-
bretastatins. The first microtubule stabilizer identified, taxol, was
isolated from the bark of the Pacific Yew. Plants of the genus Tacca
Nature has provided a wide range of compounds that affect the
mammalian cytoskeleton, including compounds that bind to tubu-
lin and disrupt the structure and function of cellular microtubules.
⇑
Corresponding author. Tel.: +1 210 567 4788; fax: +1 210 567 4300.
E-mail address: Mooberry@uthscsa.edu (S.L. Mooberry).
0
968-0896/$ - see front matter Ó 2014 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmc.2014.01.012
Please cite this article in press as: Li, J.; et al. Bioorg. Med. Chem. (2014), http://dx.doi.org/10.1016/j.bmc.2014.01.012

2
J. Li et al. / Bioorg. Med. Chem. xxx (2014) xxx–xxx
have a history of use in traditional medicines for a wide variety of
ailments, and while the chemical structures of the taccalonolides
were elucidated in the 1980’s, their microtubule stabilizing and
antitumor activities were only discovered in the last 10 years.
Taccalonolides A and E cause bundling of interphase microtubules
and mitotic arrest of cancer cells with multiple aberrant spindles
1
2
that initiate apoptosis in a manner similar to paclitaxel. However,
the taccalonolides retain efficacy in cells containing mutations in
the paclitaxel binding site as well as those expressing P-glycopro-
1
2
tein (Pgp), a drug efflux pump which contributes to the clinical
2
. Chemical isolation and structure elucidation of natural
1
3,14
resistance of paclitaxel and docetaxel.
Additional studies with
taccalonolides A–Y
the taccalonolides A and E, as well as the semi-synthetic deriva-
tives B and N, demonstrated their ability to overcome resistance
due to expression of Pgp, the bIII-isotype of tubulin or the MRP7
A compound named taccalin was initially isolated from the tu-
bers of Tacca leontopetaloides by the Scheuer laboratory in 1963,
1
5
but the structure was not solved.1 Twenty-five years later, Chen’s
drug efflux transporter. Most importantly, the taccalonolides A
and E were found to have excellent in vivo antitumor activity in
a Pgp-expressing, paclitaxel and doxorubicin resistant syngeneic
group isolated and elucidated the structures of the taccalonolides
2
A and B from the rhizomes of Tacca plantaginea. The taccalonolide
1
5
mammary tumor model. Although the in vitro antiproliferative
potencies of the taccalonolides A, E and N were at least 100-fold
lower than paclitaxel, antitumor trials demonstrated that they
had unexpectedly high in vivo potency comparable to or better
backbone consists of a highly oxygenated pentacyclic steroidal
skeleton which contains a C2–C3 epoxide group and an enol-
c-lac-
tone. Other natural taccalonolides were isolated by various labora-
2
–6
tories from 1987 to 2008, including the taccalonolides A–M
and
1
5,16
7
8
9
than paclitaxel.
Thus, in spite of the low in vitro potencies of
W–Y from Tacca plantaginea, taccalonolides N and R–V from Tac-
ca paxiana, and taccalonolides O–Q from Tacca subflabellata.
1
0,11
taccalonolides A and E, their excellent in vivo potency and efficacy
coupled with their ability to overcome taxane resistance prompted
further interest in this class of stabilizers.
Only the taccalonolides A and G–K were reported to exhibit weak
cytotoxic activity against P-388 leukemia cells in vitro.
Studies were conducted to identify the mechanism of microtu-
bule stabilization of the taccalonolides A and E. Surprisingly, in a
comprehensive study of 19 structurally diverse agents that were
reported to have microtubule stabilizing activity, Buey and col-
leagues found that the taccalonolides A and E were unable to inter-
act directly with microtubules or to enhance the polymerization of
3
. Identification of the taccalonolides A and E as microtubule
stabilizers
The ability of Tacca chantrieri extracts to cause bundling of
1
7
interphase microtubules was discovered in a cell-based screen of
crude extracts. Bioassay-guided fractionation led to the isolation
of taccalonolides A and E, the most abundantly occurring second-
ary metabolites, as new microtubule stabilizers.12 The highly acet-
ylated pentacyclic skeleton of the taccalonolides makes them
structurally distinct from all other microtubule stabilizers. The
only structural difference between the taccalonolides A and E is
purified tubulin. Furthermore, taccalonolide A was unable to en-
hance the polymerization of tubulin even in cellular extracts that
1
8
contained a full complement of cytosolic proteins. These initial
studies suggested that the taccalonolides A and E stabilize microtu-
bules in cells independently of direct microtubule binding. How-
ever, later studies showed that the significantly more potent
taccalonolides, AF and AJ, directly interacted with microtubules.19
It is interesting to speculate that the taccalonolides A and E are
1
2
the lack of an acetoxy group at C11 in taccalonolide E (Fig. 1).
21
2
2
OAc
OAc
O
OAc
H
OAc
H
1
8
20
7
16
R1
OAc 11 13
AcO
OAc
AcO
OAc
1
O
O
O
O
O
1
9
H14
24
26
H
H
O
O
9
H
O
H
O
H
O
H
O
1
H
1
0
28
R2 27
H
H
H
OH
H
O
OH
H
O
H
OH
H
O
H
OH
O
H
H
OH
O
R
O
R2
R2
4
6
OH
R1
R1
H
H
HO
HO
O
taccalonolide A : R1= R2=OAc
taccalonolide AF : R=OAc
taccalonolide AJ : R=OH
taccalonolide Z : R =OH, R =OAc
1 2
taccalonolide T: R =R =OAc
taccalonolide AM: R =R =OH
1 2
1
2
taccalonolide B : R1=OAc, R2=OH
taccalonolide E : R1=H, R2=OAc
taccalonolide N : R1=H, R2=OH
taccalonolide AA : R =R =OAc
1
2
taccalonolide AB : R =R =OH
1
2
2
1
OOH
OAc
OAc
OAc18
OAc
H
R
AcO
OAc
O
O
20
2
2
O
OAc
OH
OAc 11 13
17
H
H
H
H
19
H
24
H
H
H
O
O
15
O
O
H
H
1
1
0
8
28
H
OAc
OH
O
H
O
H
H
26
O
H
O
OH
O
H
O
H
OH
O
H
O
H
OH
R2
3
5
O
27
OH
OH
OH
OH
R1
H
HO
H
H
O
taccalonolide R : R1=R2=OAc
taccalonolide AL : R1=R2=OH
taccalonolide AN
taccalonolide AO: R=OAc
taccalonolide AK: R=H
taccalonolide AC
OAc
OAc
OAc
AcO
OAc
AcO
AcO
OAc
AcO
O
O
O
O
OAc
OAc
OAc
H
H
H
H
H
O
H
O
O
O
H
H
H
H
OH
H
O
H
OH
OH
O
H
H
O
OH
O
H
H
O
OH
O
H
O
H
OAc
O
OAc
OAc
OAc
OH
H
OH
H
OH
OH
taccalonolide I
taccalonolide AD
taccalonolide AE
taccalonolide H2
Figure 1. The Structures of taccalonolides described in this study.
Please cite this article in press as: Li, J.; et al. Bioorg. Med. Chem. (2014), http://dx.doi.org/10.1016/j.bmc.2014.01.012

J. Li et al. / Bioorg. Med. Chem. xxx (2014) xxx–xxx
3
prodrugs that are converted into binding-competent taccalono-
lides in the intact cell where they exert microtubule stabilizing ef-
fects and in vivo antitumor efficacy.
The identification of the potent taccalonolides AF and AJ facili-
tated binding studies to identify the interaction of the taccalono-
lides with microtubules. Although the taccalonolides are similar
to paclitaxel and laulimalide in their ability to enhance the extent
of microtubule polymerization from purified tubulin, they differ
from these microtubule stabilizers in several regards. The taccalon-
olides are less efficient than either paclitaxel or laulimalide in their
ability to initiate microtubule polymerization, as indicated by a
prolonged, 5–9 min lag period that occurs before the initiation of
4
. Mechanistic uniqueness of the taccalonolides
In addition to the inability of taccalonolides A and E to stimulate
tubulin polymerization in vitro, several other properties were
noted that distinguished this class of microtubule stabilizers from
classic taxane-site binding agents. Another major difference is that
taccalonolide A causes bundling of interphase microtubules at low
antiproliferative concentrations while paclitaxel does not cause
gross microtubule bundling until concentrations over 30-fold high-
er than its IC50 are used.18 This finding is interesting in light of re-
cent evidence that the interphase effects of microtubule targeted
2
3
tubulin polymerization. This is in contrast to the almost immedi-
ate polymerization observed by the addition of other microtubule
stabilizers and suggests that the taccalonolides are less effective at
1
9
nucleating microtubules.
However, once microtubules are
formed in the presence of taccalonolide AF or AJ, they exhibit an
unprecedented level of cold and mechanical stability as compared
2
0,21
19
agents likely play an important role in their anticancer actions.
to paclitaxel or laulimalide. The microtubule structures formed
Taccalonolide A also has a high degree of cellular persistence based
on the observation that exposure of cells to antiproliferative con-
centrations for only 4 h effectively eliminated any subsequent col-
in the presence of each of these drugs look essentially identical,
1
9
as visualized by electron microscopy, indicating that the taccal-
onolides are likely imparting increased stability to a fairly normal
microtubule structure in a manner that is distinct from the other
major classes of microtubule stabilizers.
1
8
ony formation. The strong interphase effects of the taccalonolides
combined with their high degree of cellular persistence were pro-
posed to partially explain why they are much more potent in vivo
than would be expected from cell-based studies.1
8
6. Identification of the taccalonolide binding site on tubulin
The mitotic spindles formed by the taccalonolides are also quite
distinct from those initiated by other classes of microtubule stabi-
Initial binding studies demonstrated that once bound, the taccal-
onolides AF and AJ could not be extracted from microtubules under a
variety of denaturing conditions, suggesting that the taccalonolides
lizers, including paclitaxel and laulimalide.1
2,22
Studies evaluating
the mitotic signaling pathways leading from microtubule stabiliza-
tion to initiation of mitotic arrest and subsequent apoptosis showed
differences among microtubule stabilizers.22 At antiproliferative
concentrations, the taccalonolides cause the formation of multipo-
lar mitotic spindles, often with 5–6 punctate mitotic spindle asters,
whereas paclitaxel treatment causes the appearance of 2–3 diffuse
1
9
covalently bind to microtubules. This was confirmed by mass
spectrometry studies with taccalonolide AJ that identified an in-
crease in the mass of the 212–230 peptide of b-tubulin by the exact
mass of the drug. Although technical limitations have so far pre-
cluded mapping the exact residue(s) of taccalonolide binding, it is
important to note that this peptide contains the N226 residue in
the luminal taxane binding pocket and the T218 residue in the pore
by which taxane-site binding agents gain access to the microtubule
1
2,22
spindle asters.
Unlike other classes of microtubule stabilizers,
the taccalonolides also inhibit centrosome disjunction and separa-
tion at antiproliferative concentrations. These centrosomal defects
initiated by the taccalonolides appear to be due to their ability to
inhibit Polo-like kinase 1 (Plk1), a mitotic kinase that regulates cen-
trosome disjunction and separation.22 Additionally, the taccalono-
lides cause premature and amplified expression of Aurora A and
Aurora A interacting proteins while the opposite is seen in paclit-
2
2
axel-treated cells. These studies demonstrated the significant dif-
ferences in mitotic signaling events that occur after treatment of
cells with the taccalonolides and other microtubule stabilizers.
5
. Isolation and semi-synthesis of the potent new
taccalonolides AF and AJ and their direct effects on tubulin
polymerization
The distinct effects of the major taccalonolides A and E led to a
search for additional minor taccalonolides to define structure
activity relationships. Taccalonolide AF was isolated as a rare con-
stituent from one fraction from Tacca plantaginea.23 This new tac-
calonolide has potent antiproliferative and microtubule
stabilizing activities with an IC50 of 23 nM. Interestingly, the only
difference between AF and the major plant metabolite taccalono-
lide A is the presence of an epoxide group at C22-23, which in-
creased the potency over 200-fold (Fig. 1).23 Due to its low
abundance and similarity to taccalonolide A, a semi-synthetic ap-
proach to produce AF was explored. Taccalonolide AF was gener-
ated from taccalonolide A with 100% yield when reacted with
Figure 2. The allosteric effects on microtubule stability elicited by the binding of
docetaxel (left) or taccalonolide AJ (right) as determined by hydrogen/deuterium
exchange mass spectrometry are depicted in red. Note in the lower panels that the
H7 and M-loop, which are affected by docetaxel binding, are not affected by
taccalonolide AJ. The covalent binding of taccalonolide AJ to microtubules was
mapped to the 212–220 peptide on b-tubulin (yellow), which contains the T218 and
N226 residues that covalently bind cyclostreptin. This is distinct, but near to where
docetaxel (purple) binds.
2
3
dimethyldioxirane. This one-step epoxidation reaction was also
applied to taccalonolide B to produce taccalonolide AJ, which re-
sulted in a dramatic 743-fold increase in potency.23 Biochemical
experiments with the potent taccalonolides AF and AJ demon-
strated, for the first time, a direct interaction with tubulin that pro-
moted microtubule polymerization.23
Please cite this article in press as: Li, J.; et al. Bioorg. Med. Chem. (2014), http://dx.doi.org/10.1016/j.bmc.2014.01.012

4
J. Li et al. / Bioorg. Med. Chem. xxx (2014) xxx–xxx
lumen, which have both been shown to covalently bind to the micro-
tubule stabilizer cyclostreptin.19,24 Another microtubule stabilizer
8. Optimization of the C15 hydrolysis reaction
that binds covalently to N226 and H227 in the taxane pocket but
The extensive binding and in vivo studies conducted with the
C22–C23 epoxidated form of taccalonolide B, taccalonolide AJ, led
to efforts to optimize the hydrolysis conditions to generate taccal-
onolide B from A. Conditions were identified that provided an 80%
yield with a 1:1:1 ratio of taccalonolide A: MeOH: 0.05 M NaHCO
not to the pore site is zampanolide.2
5,26
The taccalonolides were
found to be synergistic with either paclitaxel or laulimalide, further
suggesting they interact with microtubules in a distinct manner.
Interestingly, taccalonolide AJ was not displaced from microtubules
by either paclitaxel or laulimalide, although the covalent binding of
AJ to microtubules was sufficient to decrease paclitaxel, but not lau-
limalide, binding. The ability of the taccalonolides to displace paclit-
axel is consistent with their binding at or near the luminal and/or
pore taxane binding site(s) on microtubules.
In addition to identifying the covalent attachment of taccalono-
lide AJ to a peptide of b-tubulin, hydrogen/deuterium exchange
mass spectrometry was used to map the allosteric effects of taccal-
onolide binding on microtubules.19 As expected from its biochem-
ical and cellular effects on microtubule polymerization,
taccalonolide AJ imparted increased lateral inter-protofilament
stability. However, the mechanism of this stabilization differed
markedly from classical taxane-site binding agents in that it lar-
3
3
0
for 20 h. Similarly, taccalonolide N was produced from E under
the same hydrolysis conditions. In addition to taccalonolides B
and N, five new taccalonolides, AO, AK, AL, AM and AN, were ob-
3
0
tained as products of these reactions. These hydrolysis reactions
confirmed that the taccalonolides are very unstable in alkaline
1
6
solutions, consistent with a previous report. Biological evalua-
tions indicated that none of the new hydrolysis products were po-
tent, with all having IC50 values greater than 1
taccalonolides AO and AK having no activity up to 50
l
M and the
lM. The inac-
tive taccalonolides AO and AK are the C22–C23 keto-enol tauto-
meriztion products of taccalonolides and N, respectively,
B
resulting from the opening and reclosing of the lactone ring be-
tween C15 and C26 to form a d-lactone ring at the bottom of the
molecule under basic hydrolysis conditions. These results addi-
tionally demonstrated that the C1 acetoxy group of taccalonolide
E can be hydrolyzed to form AN, opening up the possibility of intro-
ducing different functional groups at C1, which we have shown en-
gely involved stabilization of adjacent
a
-tubulin protofilaments
and did not involve structuring of the M-loop of b-tubulin or the
H7 helix at the intradimer interface (Fig. 2).1
9,26–29
Further studies
are needed to clearly define the residues on b-tubulin to which the
taccalonolides bind covalently.
3
0
hances potency (discussed in detail below). Interestingly, this
hydrolysis was not observed with taccalonolide A, suggesting that
an acetoxy group at C11 interferes with this reaction.
The profound microtubule stabilization caused by covalent
binding of the taccalonolides to tubulin explains many of their pre-
1
8
viously observed effects, including their high cellular persistence,
ability to avoid efflux by ATP-dependent transporters
12,15
and their
9
. Isolation of new natural taccalonolides
ability to overcome drug resistance mediated by mutations in the
1
2
taxane binding site.
In addition to taccalonolide AF a number of other rare taccalon-
olides have been isolated from three Tacca species, Tacca chantrieri,
Tacca integrifolia and Tacca plantaginea, using bioassay-guided frac-
tionation. In 2011, three new taccalonolides designated Z, AA and
AB, and two known taccalonolides R and T, were isolated from
Tacca chantrieri and Tacca integrifolia (Fig. 1). The taccalonolides
Z, AA and AB are similar in structure to taccalonolide A, but each
contains a hydroxy group at C5. They have a wide range of poten-
7
. In vivo antitumor effects of taccalonolides AF and AJ
The taccalonolides are the first class of covalently bound micro-
tubule stabilizers whose antitumor activities have been exten-
sively explored.
16
1
5,16,19
As previously mentioned, their excellent
in vivo potency is likely a direct result of their ability to covalently
bind microtubules. The antitumor effects of the taccalonolides AF
and AJ were recently evaluated in the MDA-MB-231 breast cancer
xenograft model. The administration of 2 mg/kg AF twice a week
in 5% EtOH: 95% PBS (final concentration) caused significant anti-
tumor effects that were comparable to 10 mg/kg paclitaxel in
cies from low micromolar AB (1.2
lM) to low nanomolar AA
(
32 nM) (Table 1). The identification of taccalonolide AA, which
1
9
has an IC50 of 32 nM, together with taccalonolide AF with an IC50
of 23 nM led to the continued search for rare and potent natural
taccalonolides. The isolation and biological evaluations of the tac-
calonolides T and R, which also contain a C5 hydroxy but are sim-
ilar to taccalonolide E in that they lack the C11 acetoxy group,
provided significant SAR information. Taccalonolide T contains a
bulky isovalerate group at C1 and has an IC50 of 340 nM, while tac-
calonolide R, which contains a C1 acetoxy group, has 38-fold lower
potency (Table 1). The fraction from Tacca plantaginea that yielded
1
(
0% Cremophor: 90% PBS. A higher, 2.5 mg/kg, dose of AF for 2 days
1 and 5) caused total tumor regression, but was associated with
unacceptable delayed toxicity that defined this dose as the LD20
.
Four doses of 2 mg/kg taccalonolide AJ caused weight loss, but
there was no indication of any antitumor effects. Further studies
using a variety of doses and schedules of AJ showed slight antitu-
mor effects, but they were associated with unacceptable toxicities.
Chemical stability measurements of these taccalonolides in PBS
showed that while taccalonolide AJ was quite stable for over
2
0 h, taccalonolide AF had a t1/2 of 9 h in PBS and was rapidly con-
Table 1
Antiproliferative potencies of taccalonolides1
6,23,30
verted into AJ. These studies suggested the possibility that AF
might also convert to AJ in vivo and that this conversion may play
a role in the narrow therapeutic window observed for AF.1
Although excellent antitumor effects were obtained with the
taccalonolides A, E, N and AF, they correlate with weight loss, indi-
cating a narrow therapeutic window. However, differences have
been noted among the taccalonolides with regard to the degree
of toxicity that is associated with therapeutic doses, suggesting
the feasibility of identifying a taccalonolide with an acceptable
therapeutic window for clinical development.19 For instance, it is
intriguing to hypothesize that the toxicity observed with AF may
be diminished by minimizing its conversion to AJ.
Compound
IC50
(
l
M)
Compound
IC50 (lM)
9
Taccalonolide A
Taccalonolide B
Taccalonolide E
Taccalonolide N
Taccalonolide I
Taccalonolide AO
Taccalonolide AK
Taccalonolide AL
Taccalonolide AM
Taccalonolide AN
Taccalonolide R
5.3 ± 0.2
3.1 ± 0.2
40 ± 5
8.5 ± 0.4
49 ± 3
>50
>50
34 ± 8
Taccalonolide T
Taccalonolide Z
Taccalonolide AA
Taccalonolide AB
Taccalonolide AC
Taccalonolide AD
Taccalonolide AE
Taccalonolide AF
Taccalonolide AJ
Taccalonolide H2
0.34 ± 0.02
0.120 ± 0.008
0.032 ± 0.002
2.7 ± 0.1
>50
3.4 ± 0.2
5.0 ± 0.2
0.023 ± 0.003
0.0042 ± 0.0003
0.73 ± 0.02
2.0 ± 0.1
1.5 ± 0.1
13 ± 1
Please cite this article in press as: Li, J.; et al. Bioorg. Med. Chem. (2014), http://dx.doi.org/10.1016/j.bmc.2014.01.012

J. Li et al. / Bioorg. Med. Chem. xxx (2014) xxx–xxx
5
taccalonolide AF also yielded four additional new taccalonolides in
trace quantitities, AC-AE and H2.23
C5-hydroxy group in taccalonolide Z as compared to taccalonolide
A. Moreover, the presence of the C5-hydroxy group did not have a
major effect on the potencies of taccalonolides B and AB. Therefore,
the presence of the C5-hydroxy group in itself does not have a pre-
dictable effect on potency although this modification can be
important when combined with other modifications. For instance,
compounds lacking the C5-hydroxy group show increased potency
when the C15-acetoxy group is hydrolyzed, such as when A or E
are hydrolyzed to generate B and N, respectively. However, this
same hydrolysis gives rise to decreased potency in molecules con-
taining the C5-hydroxy group as evidenced by a comparison of Z to
AB. These data suggest that multiple interactions across the taccal-
onolide molecule can interplay in complex ways to affect potency.
1
0. Structure activity relationships of the taccalonolides
The vast array of natural and semi-synthetic taccalonolide ana-
logues characterized in our laboratory has allowed for rigorous
structure activity relationship (SAR) analyses that will greatly facil-
itate modeling of the interactions between the taccalonolides and
microtubules and possibly aid in the identification of an optimal
clinical candidate. The current SAR findings clearly demonstrate
that there are complex relationships among multiple sites on the
taccalonolide backbone, however, the following conclusions can
been drawn with regard to the key moieties of the taccalonolide
structure that are critical for optimal efficacy and potency of this
class of compounds.
1
0.4. The C5-8 substituents on the lower portion of the B ring
can affect taccalonolide potency
1
0.1. E-ring constituents at C20–C23 play a key role in
One example of the importance of the C5-8 region of the taccal-
onolide backbone is evidenced by the two natural taccalonolides A
and I. The only difference between taccalonolides A and I is a
switch between the C6 ketone and the C7-hydroxy. This difference
resulted in a 9.25-fold decrease in potency for I, indicating this
arrangement is not optimal for biological activity. However, when
these same substituents are present in combination with a double
bond at C5–C6, as in taccalonolide AD, the potency is slightly in-
creased as compared to A, suggesting that this double bond plays
a role in maintaining activity. Additionally, when a second hydroxy
group is present at the C7 of taccalonolide A to give the rare gem-
inal diol group on taccalonolide AE, the potency was also un-
changed, further indicating that there is some degree of
flexibility for modifications in this region. The 7-fold increase in
potency of taccalonolide H2 afforded by a C7-C8 double bond com-
pared to taccalonolide A demonstrates that a conjugated B ring
with a ketone at C6 position and a double bond at C7–C8 may be
optimal for potency in this region.
taccalonolide potency
As discussed previously, analogues such as AO and AK, which
have significant structural rearrangements in the C20-23 region,
show no antiproliferative or microtubule stabilizing activities at
concentrations up to 50
served for taccalonolide AC, which contains an
group at C20 (Fig. 1), as opposed to the majority of other taccalon-
olides which contain an -methyl group at this site, indicating that
this additional -hydroperoxyl group is not optimal for bioactivity.
In contrast, the epoxidation of the C22–C23 double bond was
shown to dramatically increase taccalonolide potency as evidenced
by the generation of the potent taccalonolides AF and AJ from their
parent taccalonolides A and B, respectively (Table 1). Ongoing
studies demonstrate that this C22-23 epoxidation increases the po-
tency of other natural and semisynthetic taccalonolides. The sub-
stantial changes in potency afforded by small modifications to
this region of the taccalonolide backbone strongly suggest that this
area is critical for optimal drug binding to microtubules.
30
l
M. A similar lack of potency was ob-
a-hydroperoxyl
a
a
11. Conclusion
1
0.2. A large, bulky group at C1 provides a significant increase in
The taccalonolides are a novel class of microtubule stabilizing
agents with the ability to circumvent clinically relevant forms of
taxane resistance both in vitro and in vivo. The ability of the taccal-
onolides to covalently bind b-tubulin contributes toward their high
cellular persistence and affords in vivo administration at low con-
centrations in aqueous vehicles. Detailed SAR data from a large
number of natural and semisynthesized taccalonolides provides
guidance for further modifications that will facilitate more detailed
modeling of the taccalonolide binding site and may lead to the
identification of a clinical candidate.
potency
The first indication that C1 modifications can dramatically im-
pact potency came from the finding that replacing the C1 acetoxy
in R with an isovalerate group in T resulted in a 38-fold increase in
1
6
potency. This was confirmed with the 17-fold increase in potency
observed by the replacement of the acetoxy group at C1 in taccalon-
olide AL with an isovalerate group in taccalonolide AM.3 Interest-
ingly, a decrease in bulk at C1 offered by the hydrolysis of the
acetoxy group in AL to generate AN in combination with removal
of the C5 hydroxy group also offers a similar increase in potency.
More highly potent new taccalonolides with isovalerate or other
bulky modifications at C1 can likely be isolated or produced
semi-synthetically as evidenced by the selective hydrolysis of C1
in the absence of a C11 acetoxy group as described above. These
data highlight the importance of the C1 position in facilitating
taccalonolide activity and the opportunity to improve potency by
hydrolyzing this acetoxyl group with the potential to add steric bulk
at this site.
0
Acknowledgements
This work was supported by NCI CA121138 (S.L.M.) and NIDCR
DE 14318 (J.L.). We would like to thank Dr. David Schriemer for
assistance with creation of the taccalonolide binding figure and
Cristina Rohena and Andrew Robles for their critical reading of
the manuscript.
References and notes
1
0.3. There are complex indications with regard to the
1. Scheuer, P. J.; Swanholm, C. E.; Madamba, L. A.; Hudgins, W. R. Lloydia 1963, 26,
133.
importance of the C5 hydroxy group to potency
2
3
.
.
Chen, Z.; Wang, B.; Chen, M. Tetrahedron Lett. 1987, 28, 1673.
Chen, Z.; Wang, B.; Shen, J. Phytochemistry 1988, 27, 2999.
Upon comparison of the taccalonolides N and AL, the C5-hydro-
xy group is the only structural difference, which leads to a 4-fold
decrease in potency for taccalonolide AL. By contrast, a significant
4. Chen, Z. L.; Shen, J. H.; Gao, Y. S.; Wichtl, M. Planta Med. 1997, 63, 40.
5
6
7
.
.
.
Shen, J.; Chen, Z.; Gao, Y. Chin. J. Chem. 1991, 9, 92.
Shen, J.; Chen, Z.; Gao, Y. Phytochemistry 1996, 42, 891.
Yang, J.-Y.; Zhao, R.-H.; Chen, C.-X.; Ni, W.; Teng, F.; Hao, X.-J.; Liu, H.-Y. Helv.
Chim. Acta 2008, 91, 1077.
4
4-fold increase in potency was observed by the presence of a
Please cite this article in press as: Li, J.; et al. Bioorg. Med. Chem. (2014), http://dx.doi.org/10.1016/j.bmc.2014.01.012

6
J. Li et al. / Bioorg. Med. Chem. xxx (2014) xxx–xxx
8
.
.
Muehlbauer, A.; Gehling, M.; Velten, R.; Andersch, W.; Erdelen, C.; Harder, A.;
Marczok, P.; Nauen, R.; Turberg, A.; Tran, V. S.; Adam, G.; Liu, J.
WO2001040256A1, 2001, 113.
21. Sackett, D. L.; Fojo, T. Cell Cycle 2011, 10, 3233.
22. Rohena, C. C.; Peng, J.; Johnson, T. A.; Crews, P.; Mooberry, S. L. Biochem.
Pharmacol. 2013, 85, 1104.
23. Li, J.; Risinger, A. L.; Peng, J.; Chen, Z.; Hu, L.; Mooberry, S. L. J. Am. Chem. Soc.
2011, 133, 19064.
24. Buey, R. M.; Calvo, E.; Barasoain, I.; Pineda, O.; Edler, M. C.; Matesanz, R.;
Cerezo, G.; Vanderwal, C. D.; Day, B. W.; Sorensen, E. J.; Lopez, J. A.; Andreu, J.
M.; Hamel, E.; Diaz, J. F. Nat. Chem. Biol. 2007, 3, 117.
25. Field, J. J.; Pera, B.; Calvo, E.; Canales, A.; Zurwerra, D.; Trigili, C.; Rodriguez-
Salarichs, J.; Matesanz, R.; Kanakkanthara, A.; Wakefield, S. J.; Singh, A. J.;
Jimenez-Barbero, J.; Northcote, P.; Miller, J. H.; Lopez, J. A.; Hamel, E.;
Barasoain, I.; Altmann, K. H.; Diaz, J. F. Chem. Biol. 2012, 19, 686.
26. Prota, A. E.; Bargsten, K.; Zurwerra, D.; Field, J. J.; Diaz, J. F.; Altmann, K. H.;
Steinmetz, M. O. Science 2013, 339, 587.
9
Muehlbauer, A.; Seip, S.; Nowak, A.; Tran, V. S. Helv. Chim. Acta 2003, 86, 2065.
1
1
1
0. Huang, Y.; Liu, J.-K.; Muhlbauer, A.; Henkel, T. Helv. Chim. Acta 2002, 85, 2553.
1. Huang, Y.; Muehlbauer, A.; Henkel, T.; Liu, J. K. Chin. Chem. Lett. 2003, 14, 68.
2. Tinley, T. L.; Randall-Hlubek, D. A.; Leal, R. M.; Jackson, E. M.; Cessac, J. W.;
Quada, J. C., Jr.; Hemscheidt, T. K.; Mooberry, S. L. Cancer Res. 2003, 63, 3211.
3. Fojo, T.; Menefee, M. Ann. Oncol. 2007, 18, v3.
1
1
4. Horwitz, S. B.; Cohen, D.; Rao, S.; Ringel, I.; Shen, H. J.; Yang, C. P. J. Natl. Cancer
Inst. Monogr. 1993, 55.
1
5. Risinger, A. L.; Jackson, E. M.; Polin, L. A.; Helms, G. L.; LeBoeuf, D. A.; Joe, P. A.;
Hopper-Borge, E.; Luduena, R. F.; Kruh, G. D.; Mooberry, S. L. Cancer Res. 2008,
68, 8881.
1
1
6. Peng, J.; Risinger, A. L.; Fest, G. A.; Jackson, E. M.; Helms, G.; Polin, L. A.;
Mooberry, S. L. J. Med. Chem. 2011, 54, 6117.
7. Buey, R. M.; Barasoain, I.; Jackson, E.; Meyer, A.; Giannakakou, P.; Paterson, I.;
Mooberry, S.; Andreu, J. M.; Diaz, J. F. Chem. Biol. 2005, 12, 1269.
8. Risinger, A. L.; Mooberry, S. L. Cell Cycle 2011, 10, 2162.
27. Huzil, J. T.; Chik, J. K.; Slysz, G. W.; Freedman, H.; Tuszynski, J.; Taylor, R. E.;
Sackett, D. L.; Schriemer, D. C. J. Mol. Biol. 2008, 378, 1016.
28. Khrapunovich-Baine, M.; Menon, V.; Yang, C. P.; Northcote, P. T.; Miller, J. H.;
Angeletti, R. H.; Fiser, A.; Horwitz, S. B.; Xiao, H. J. Biol. Chem. 2011, 286, 11765.
29. Xiao, H.; Verdier-Pinard, P.; Fernandez-Fuentes, N.; Burd, B.; Angeletti, R.; Fiser,
A.; Horwitz, S. B.; Orr, G. A. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 10166.
30. Li, J.; Peng, J.; Risinger, A. L.; Mooberry, S. L. J. Nat. Prod. 2013, 76, 1369.
1
1
9. Risinger, A. L.; Li, J.; Bennett, M. J.; Rohena, C. C.; Peng, J.; Schriemer, D. C.;
Mooberry, S. L. Cancer Res. 2013, 73, 6780.
2
0. Komlodi-Pasztor, E.; Sackett, D.; Wilkerson, J.; Fojo, T. Nat. Rev. Clin. Oncol.
2
011, 8, 244.
Please cite this article in press as: Li, J.; et al. Bioorg. Med. Chem. (2014), http://dx.doi.org/10.1016/j.bmc.2014.01.012