Synthesis and biological evaluation of spongistatin/altohyrtin analogues: E-ring dehydration and C46 side-chain truncation

Article abstract of DOI:10.1039/b212651f

Simplified analogues of the potent antimitotic marine macrolide spongistatin 1/altohyrtin A were synthesised and evaluated as growth inhibitory agents against a range of human tumour cell lines, including Taxol-resistant strains, revealing that E-ring deh

Full text of DOI:10.1039/b212651f

Synthesis and biological evaluation of spongistatin/altohyrtin analogues:
E-ring dehydration and C46 side-chain truncation†
Ian Paterson,* Jose L. Aceña, Jordi Bach, David Y.-K. Chen and Mark J. Coster
University Chemical Laboratory, Lensfield Road, Cambridge, UK CB2 1EW. E-mail: ip100@cam.ac.uk;
Fax: +44 1223 336362; Tel: +44 1223 336407
Received (in Cambridge, UK) 23rd December 2002, Accepted 15th January 2003
First published as an Advance Article on the web 23rd January 2003
Simplified analogues of the potent antimitotic marine
macrolide spongistatin 1/altohyrtin A were synthesised and
evaluated as growth inhibitory agents against a range of
human tumour cell lines, including Taxol-resistant strains,
revealing that E-ring dehydration leads to enhanced cyto-
toxicity at the low picomolar level while truncation of the
side-chain at C46 results in a drastic decrease in activity.
Smith et al. have described two such simplified analogues,
bearing a model F ring and retaining the C44–C51 triene side-
chain.6 These displayed growth inhibitory activity against
several cancer cell lines, although only at the micromolar level.
Whereas Uckun and co-workers have reported the design,
synthesis and surprising level of biological activity of a greatly
simplified analogue, based solely on the AB-spiroacetal motif.7
Recently, this claim has been questioned, as the Smith group
prepared this same analogue, as well as another AB-spiroacetal
mimic, and found neither had significant cytotoxic or anti-
tubulin activity.8 With these contradictory results in mind, we
decided to explore less drastic simplifications of 1 to gather
SAR data and help identify the spongipyran pharmacophore.
Herein, we report the synthesis and in vitro biological
evaluation of two novel analogues, where E-ring dehydration at
C35, as in 2, leads to enhanced cytotoxicity at the low picomolar
level, while truncation of the side-chain at C46, as in 3, results
in a drastic decrease in activity.
The spongipyrans are a unique family of bis(spiroacetal)
macrolides, obtained from marine sponges in trace amounts by
bioassay-guided isolation.1,2 They display exceptional cytotox-
icity against a wide variety of human cancer cell lines, the mode
of action stemming from inhibition of mitosis by binding in the
vinca domain of tubulin and blocking microtubule assembly.
Spongistatin 1 (1 · altohyrtin A) is typical of the series, as
reported by the Pettit1a–d and Kobayashi groups,1ef having sub-
nanomolar growth inhibitory activity in the US National Cancer
Institute primary screen.
Due to the extremely scarce natural supply and the painstak-
ing isolation procedure, lack of material has halted the
preclinical development of these compounds in cancer chemo-
therapy. This has stimulated numerous synthetic efforts,
culminating in several completed total syntheses,2,3 including
one by ourselves.3a Despite this interest, relatively little is
known regarding structure–activity relationships (SAR) for
these architecturally complex marine natural products.4–8 In
particular, is it possible to design progressively simplified
analogues of 1 that retain the exceptional cancer cell growth
inhibitory properties whilst increasing their synthetic accessi-
bility?5
The first analogue arose out of our recent total synthesis of
spongistatin 1 (1),3a where the final deprotection step, per-
formed under acidic conditions (aqueous HF, MeCN), was
accompanied by the formation of a minor component. This
byproduct was purified by reverse-phase HPLC and showed
close homology to that of 1 in its 1H NMR spectra, with
differences only in the regions corresponding to resonances for
the E-ring protons. Following mass spectrometry and detailed
1H NMR analysis (500 MHz, COSY), the structure of this
analogue was determined as 2, corresponding to an E-ring
dehydrated version of 1, i.e. lacking the axial C35 hydroxy
group with a double bond between C35 and C36.
The planned synthesis of a side-chain truncated analogue 3
was also undertaken (Scheme 1). This otherwise complete
spongipyran was designed to evaluate the effect of deleting the
terminal chlorodiene and the C47 hydroxy group. The EF
subunit 43a was first treated with TBSCl, imidazole and Et3N in
DMF to silylate the C35 hydroxy group, followed by displace-
ment of the C29 chloride by exposure to Ph3P and NaI (iPr2NEt,
MeOH, MeCN) to give 5 (93%). By employing our optimised
conditions (LiHMDS, THF, HMPA),3a the challenging Wittig
coupling between the side-chain truncated EF phosphonium salt
5 and the ABCD aldehyde 63b was achieved in 71% yield
( > 97+3 Z+E). Subsequent removal of the three PMB protecting
groups (DDQ) and the trichloroethyl ester (TCE) moiety (Zn,
THF, NH4OAc) provided seco-acid 7, as a ca. 1+1 mixture of
methyl acetal and hemiacetal forms (R = Me, H respectively).
Macrolactonisation of 7 under modified Yamaguchi conditions,
engaging the C41 hydroxyl group in a completely selective
manner, was followed by global deprotection of the resulting
42-membered macrocycle by its exposure to aqueous HF in
MeCN, providing the desired side-chain truncated analogue 3.
With spongistatin analogues 2 and 3 in hand, the effects on
cytotoxicity of E-ring dehydration and removal of the C47–C51
chlorodiene allylic alcohol were now investigated. Listed in
Table 1 are the results of growth inhibition experiments9 for
paclitaxel/Taxol (8), spongistatin 1 (1),3a E-ring dehydrated
analogue 2 and side-chain truncated analogue 3 against a set of
five different human cancer cell lines, including the MIP101
colon and 1A9PTX22 ovarian, paclitaxel-resistant strains. In all
† Electronic supplementary information (ESI) available: spectroscopic data
for compounds 2 and 3. See http://www.rsc.org/suppdata/cc/b2/b212651f/
462
CHEM. COMMUN., 2003, 462–463
This journal is © The Royal Society of Chemistry 2003

Scheme 1 Synthesis of C46 side-chain truncated spongipyran 3. Reagents and conditions: (i) TBSCl, Im, Et3N, DMF, 20 °C, 16 h, 95%; (ii) PPh3, NaI,
iPr2NEt, MeCN–MeOH, D, 20 h, 98%; (iii) LiHMDS, THF–HMPA, 278 °C, 10 min; 6, 278 ? 20 °C, 40 min, 71%; (iv) DDQ, CH2Cl2–pH 7 buffer, 0
°C, 90 min, 69%; (v) Zn, THF–NH4OAc(aq), 20 °C, 30 min, 62%; (vi) 2,4,6-trichlorobenzoyl chloride, Et3N, THF, 20 °C, 3 h; DMAP, PhMe, 100 °C, 16
h, 47%; (vii) HF(aq), MeCN, 0 °C, 30%.
Table 1 Growth inhibition against human cancer cell linesa
Financial support was provided by the EPSRC (GR/L41646),
EC (Marie Curie Fellowship to J. L. A.), Cambridge Com-
IC50 values (nM)
8
1
2
3
monwealth Trust (M. J. C.), KingAs College and the Sims Fund,
Cambridge (D.Y.-K. C.). We thank Merck, AstraZeneca and
Novartis Pharmaceuticals for generous support, and Dr
Kennneth W. Bair, Dr Frederick R. Kinder, Jr., Dr Peter T.
Lassota and Dr Erik F. Sorensen at Novartis for providing the
biological data.9
MIP101 colon (Pgp-1 overexpressing)
HCT116 colon
200
0.1
0.08
587
407
0.3 0.05 0.02
1A9PTX22 ovarian (mutation in b-tubulin) 47
0.03 0.007 > 632
0.03 0.007 > 632
0.07 0.04
1A9 ovarian (parental)
1
6
A549 non-small cell lung
> 632
a Cells were plated in 96 well plates at 4 3 103 cells per well for MIP101,
HCT116 and A549 cell lines, and at 2 3 104 cells per well for 1A9 and
1A9PTX22 cell lines. Compounds dissolved in DMSO were added to the
wells at 5 fold serial dilutions starting from 1 mg ml21. No compound was
added to control wells. After 72 h incubation, the number of viable cells was
assessed using an MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazo-
lium bromide) assay.10 After processing, the plates were read in a Molecular
Devices 96-well plate reader at 540 nm and IC50 values (concentrations in
nM causing 50% inhibition of cell growth) were calculated.
Notes and references
1 (a) G. R. Pettit, Z. A. Cichacz, F. Gao, C. L. Herald, M. R. Boyd, J. M.
Schmidt and J. N. A. Hooper, J. Org. Chem., 1993, 58, 1302; (b) G. R.
Pettit, Z. A. Cichacz, F. Gao, C. L. Herald and M. R. Boyd, J. Chem.
Soc., Chem. Commun., 1993, 1166; (c) R. L. Bai, Z. A. Cichacz, C. L.
Herald, G. R. Pettit and E. Hamel, Mol. Pharmacol., 1993, 44, 757; (d)
R. L. Bai, G. F. Taylor, Z. A. Cichacz, C. L. Herald, J. A. Kepler, G. R.
Pettit and E. Hamel, Biochemistry, 1995, 34, 9714; (e) M. Kobayashi, S.
Aoki, H. Sakai, K. Kawazoe, N. Kihara, T. Sasaki and I. Kitagawa,
Tetrahedron Lett., 1993, 34, 2795; (f) M. Kobayashi, S. Aoki, K. Gato
and I. Kitagawa, Chem. Pharm. Bull., 1996, 44, 2142; (g) N. Fusetani,
K. Shinoda and S. Matsunaga, J. Am. Chem. Soc.,, 1993, 115, 3977.
2 Review: J. Pietruszka, Angew. Chem., Int. Ed., 1998, 37, 2629.
3 For leading references, see: (a) I. Paterson, D. Y.-K. Chen, M. J. Coster,
J. L. Aceña, J. Bach, K. R. Gibson, L. E. Keown, R. M. Oballa, T.
Trieselmann, D. J. Wallace, A. P. Hodgson and R. D. Norcross, Angew.
Chem., Int. Ed., 2001, 40, 4055; (b) I. Paterson, D. J. Wallace and R. M.
Oballa, Tetrahedron Lett., 1998, 39, 8545; (c) D. A. Evans, B. W.
cases, 1 was found to be substantially more active (6- to
2000-fold) than paclitaxel, and was particularly effective
against the MIP101 colon carcinoma cell line, indicating that it
is a poor substrate for the P-glycoprotein (Pgp) drug efflux
pump. Given the already exceptional cytotoxicity displayed by
1,1a–f,2 we were gratified to find that the E-ring dehydrated
analogue 2 was generally (2- to 4-fold) more potent than the
parent natural product. Analogue 2 had low picomolar IC50
values, in the range 0.007–0.08 nM, against this set of cancer
cell lines. This indicates that the C35 hydroxyl of 1 is
unnecessary for biological activity, and that its removal leads to
an increase in potency. In contrast, the dramatic attenuation of
cytotoxicity for analogue 3, against all cell lines employed in
these assays (e.g. 0.587 and 0.407 mM against the MIP101 and
HCT116 colon carcinoma cell lines), reveals that the C47–C51
chlorodiene allylic alcohol moiety is an essential structural
feature. These results suggest that the full C44–C51 triene side-
chain is a crucial part of the spongipyran pharmacophore,
consistent with the findings of the Smith group6 but not with the
claims made by Uckun and co-workers.7
In conclusion, we have prepared two fully synthetic ana-
logues of the bis(spiroacetal) macrolide spongistatin 1 (1).
Evaluation of their growth inhibitory activity in vitro against a
range of human cancer cell lines reveals that dehydration in the
E ring actually leads to enhanced potency—at the low
picomolar level—highlighting the extraordinary cytotoxicity of
these compounds, particularly against Taxol-resistant cell lines.
Additionally, truncation of the side-chain at C46 leads to a
drastic decrease in activity (!5 3 103-fold less than 1). While
E-ring modification can be tolerated but not F-ring side-chain
truncation, the design of simplified synthetic analogues of the
spongipyrans, retaining the exceptional antimitotic potency,
clearly requires a great deal more SAR work.
ˆ
Trotter, P. J. Coleman, B. Coté, L. C. Dias, H. A. Rajapakse and A. N.
Tyler, Tetrahedron, 1999, 55, 8671; (d) M. M. Hayward, R. M. Roth, K.
J. Duffy, P. I. Dalko, K. L. Stevens, J. S. Guo and Y. Kishi, Angew.
Chem., Int. Ed., 1998, 37, 192; (e) A. B. Smith III, Q. Y. Lin, V. A.
Doughty, L. Zhuang, M. D. McBriar, J. K. Kerns, C. S. Brook, N.
Murase and K. Nakayama, Angew. Chem., Int. Ed., 2001, 40, 196; (f) M.
T. Crimmins, J. D. Katz, D. G. Washburn, S. P. Allwein and L. F.
McAtee, J. Am. Chem. Soc., 2002, 124, 5661.
4 A C50 chlorine substituent (as in spongistatin 1/altohyrtin A) shows
somewhat greater cell growth inhibitory activity than the des-chloro
version (spongistatin 2/altohyrtin C), see ref. 2.
5 This approach has been used successfully by Kishi and co-workers for
the halichondrins, see: M. J. Towle, K. A. Salvato, J. Budrow, B. F.
Wels, G. Kuznetsov, K. K. Aalfs, S. Welsh, W. J. Zheng, B. M.
Seletsky, M. H. Palme, G. J. Habgood, L. A. Singer, L. V. Di Pietro, Y.
Wang, J. J. Chen, D. A. Quincy, A. Davis, K. Yoshimatsu, Y. Kishi, M.
J. Yu and B. A. Littlefield, Cancer Res., 2001, 61, 1013.
6 A. B. Smith III, Q. Y. Lin, G. R. Pettit, J.-C. Chapuis and J. M. Schmidt,
Bioorg. Med. Chem. Lett., 1998, 8, 567.
7 (a) F. M. Uckun, C. Mao, A. O. Vassilev, H. Huang and S.-T. Jan,
Bioorg. Med. Chem. Lett., 2000, 10, 541; (b) H. Huang, C. Mao, S.-T.
Jan and F. M. Uckun, Tetrahedron Lett., 2000, 41, 1699.
8 A. B. Smith III, R. M. Corbett, G. R. Pettit, J.-C. Chapuis, J. M. Schmidt,
E. Hamel and M. K. Jung, Bioorg. Med. Chem. Lett., 2002, 12, 2039.
9 The biological assays were performed by Novartis Pharmaceuticals
Corporation, Summit, New Jersey.
10 M. C. Alley, C. M. Pacula-Cox, M. L. Hursey, L. R. Rubenstein and M.
R. Boyd, Cancer Res., 1991, 51, 1247.
CHEM. COMMUN., 2003, 462–463
463