Discovery of (E)-9,10-dehydroepothilones through chemical synthesis: On the emergence of 26-trifluoro-(E)-9,10-dehydro-12,13-desoxyepothilone B as a promising anticancer drug candidate

Article abstract of DOI:10.1021/ja046992g

We provide a full account of the discovery of the (E)-9,10-dehydro derivatives of 12,13-desoxyepothilone B (dEpoB), a new class of antitumor agents with promising in vivo preclinical properties. The compounds, which are to date not available by modification of any of the naturally occurring epothilones, were discovered through total chemical synthesis. We describe how our investigations of ring-closing metathesis reactions in epothilone settings led to the first and second generation syntheses of (E)-9,10-dehydro-12,13- desoxyepothilone congeners. With further modifications, the synthesis was applied to reach a 26-trifluoro derivative compound (see compound 7). To conduct such studies and in anticipation of future development needs, the total synthesis which led to the initial discovery of compound 7 was simplified significantly. The total synthesis methodology used to reach compound 7 was then applied to reach more readily formulated compounds, bearing hydroxy and amino functionality on the 21-position (see compounds 45,62, and 63). Following extensive in vitro evaluations of these new congeners, compound 7 was nominated for in vivo evaluations in xenograft models. The data provided herein demonstrate a promising therapeutic efficacy, activity against large tumors, nonrelapseability, and oral activity. These results have identified compound 7 as a particularly promising compound for clinical development. The excellent, totally synthetic, route to 7 makes such a program quite feasible.

Full text of DOI:10.1021/ja046992g

Published on Web 08/13/2004
Discovery of (E)-9,10-Dehydroepothilones through Chemical
Synthesis: On the Emergence of
26-Trifluoro-(E)-9,10-dehydro-12,13-desoxyepothilone B as a
Promising Anticancer Drug Candidate
Alexey Rivkin,† Fumihiko Yoshimura,† Ana E. Gabarda,† Young Shin Cho,†
Ting-Chao Chou,‡ Huajin Dong,‡ and Samuel J. Danishefsky*,†,§
Contribution from the Laboratory for Bioorganic Chemistry, Preclinical Pharmacology Core
Facility, Sloan-Kettering Institute for Cancer Research, 1275 York AVenue,
New York, New York 10021, and Department of Chemistry, Columbia UniVersity,
HaVemeyer Hall, New York, New York 10027
Received May 21, 2004; E-mail: s-danishefsky@ski.mskcc.org.
Abstract: We provide a full account of the discovery of the (E)-9,10-dehydro derivatives of 12,13-
desoxyepothilone B (dEpoB), a new class of antitumor agents with promising in vivo preclinical properties.
The compounds, which are to date not available by modification of any of the naturally occurring epothilones,
were discovered through total chemical synthesis. We describe how our investigations of ring-closing
metathesis reactions in epothilone settings led to the first and second generation syntheses of (E)-9,10-
dehydro-12,13-desoxyepothilone congener 6. With further modifications, the synthesis was applied to reach
a 26-trifluoro derivative compound (see compound 7). To conduct such studies and in anticipation of future
development needs, the total synthesis which led to the initial discovery of compound 7 was simplified
significantly. The total synthesis methodology used to reach compound 7 was then applied to reach more
readily formulated compounds, bearing hydroxy and amino functionality on the 21-position (see compounds
45, 62, and 63). Following extensive in vitro evaluations of these new congeners, compound 7 was nominated
for in vivo evaluations in xenograft models. The data provided herein demonstrate a promising therapeutic
efficacy, activity against large tumors, nonrelapseability, and oral activity. These results have identified
compound 7 as a particularly promising compound for clinical development. The excellent, totally synthetic,
route to 7 makes such a program quite feasible.
Introduction
Epothilones A (1) and B (2) are naturally occurring cytotoxic
macrolides which were initially isolated by Ho¨fle and co-
workers from the mycobacterium Sorangium cellulosum (Figure
1).1 Interest in epothilones originated from the discovery that
their mode of antitumor activity mimicked that of the established
clinically useful taxoids (3 and 4).2 It has been shown that the
taxoids3 and epothilones interrupt the dynamic mechanism of
microtubule assembly/disassembly via microtubulin stabilization.
† Laboratory for Bioorganic Chemistry, Sloan-Kettering Institute.
‡ Preclinical Pharmacology Core Facility, Sloan-Kettering Institute.
§ Department of Chemistry, Columbia University.
(1) (a) Ho¨fle, G. H.; Bedorf, N.; Steinmetz, H.; Schomburg, D.; Gerth, K.;
Reichenbach, H. Angew. Chem., Int. Ed. Engl. 1996, 35, 1567-1569. (b)
Gerth, K.; Bedorf, N.; Ho¨fle, G.; Irschik, H.; Reichenbach, H. J. Antibiot.
1996, 49, 560-563. (c) Ho¨fle, G.; Bedorf, N.; Gerth, K.; Reichenbach H.
(GBF). DE-B 4138042; Chem. Abstr. 1993, 120, 52841.
(2) (a) Bollag, D. M.; McQueney, P. A.; Zhu, J.; Hensens, O.; Koupal, L.;
Liesch, J.; Goetz, M.; Lazarides, E.; Woods, C. M. Cancer Res. 1995, 55,
2325-2333. (b) Kowalski, R. J.; Terhaar, E.; Longley, R. E.; Gunasekera,
S. P.; Lin, C. M.; Day, B. V.; Hamel, E. Mol. Biol. Cell 1995, 6, 2137.
(3) (a) Nicolaou, K. C.; Dai, W.-M.; Guy, R. K. Angew. Chem., Int. Ed. Engl.
1994, 33, 15-44. (b) Rowinsky, E. K. Annu. ReV. Med. 1997, 48, 353-
374. For reviews, see: (c) Taxane Anticancer Agents; Georg, G. I., Chen,
T. T., Ojima, I., Vyas, D. M., Eds.; American Chemical Society: San Diego,
CA, 1995. (d) The Chemistry and Pharmacology of Taxol and its
DeriVatiVes; Farin, V., Ed.; Elsevier: New York, 1995. (e) Taxol: Sciences
& Applications; Suffness, M., Ed.; CRC Press: Boca Raton, FL, 1995.
Figure 1. Structures of epothilones and taxoids.
This mode of drug induced intervention initiates cell death
through apoptosis. In contrast to paclitaxel, the epothilones seem
to exhibit near imperviousness to the defenses of otherwise
9
10.1021/ja046992g CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 10913-10922
10913

A R T I C L E S
Rivkin et al.
Figure 2. Anticancer drug candidates.
multidrug resistant cells.4 It was this property, as well as more
promising formulatibility, which prompted great interest in
epothilones as anticancer drug candidates.5
Following extensive multidisciplinary research directed to the
biology, chemistry, pharmacology, toxicology, and biosynthesis
of the epothilones, three agents, including one from our program
(vide infra), have already been advanced to phase I and phase
II clinical trials.6
Taking advantage of our then pioneering total syntheses of
epothilones A and B,7 preliminary biological studies with
selected probe structures suggested that the 12,13-oxido linkage
of the macrolactone is a locus of nontumor selective toxicity.8
This perception led us to an extended evaluation of 12,13-
desoxyEpoB (dEpoB, 5) and related desoxy congeners (Figure
2). The desoxy congeners indeed seemed to exhibit less
nontumor directed toxicity9 relative to epothilone B and a far
broader apparent therapeutic index than either EpoB or the
taxoids. The advantages of dEpoB relative to paclitaxel, at the
level of nude mouse xenograft models, is particularly dramatic
with resistant tumors. dEpoB (5) is currently in phase II clinical
trials.
More recently our laboratory reported that incorporation of
E-9,10 unsaturation in the macrolide framework of compound
5 (dEpoB) resulted in a marked increase in potency and in
metabolic stability.10 These properties gave rise to favorable
outcomes in xenograft tumor studies with nude mice. From this
family, 26-trifluoro-(E)-9,10-dehydroepothilone (7) has emerged
as a most promising candidate for drug development.10b Herein
we provide a full account of the novel synthetic chemistry which
led to the discovery of compound 7 and its striking performance
with human cancers in murine models.
Our epothilone program has been directed to two intertwined
goals. The first was the study of the structure-activity relation-
ship of carefully crafted analogues with enhanced biological
properties. The second goal was the exploration of various
strategies to develop a practical total synthesis of drug candi-
dates. Our laboratory does not have access to epothilones of
natural origin. Only through total synthesis could we enter into
a sustainable discovery and development program. Moreover,
we hoped to evaluate structures of either enhanced or diminished
complexity which could not readily be reached from naturally
occurring epothilones. Hence, the total synthesis we practiced
should be adaptable to gaining access to chemical “space” not
accessible from naturally occurring epothilones.
(4) Giannakakou, P.; Sackett, D. L.; Kang, Y.-K.; Zhan, Z.; Buters, J. T.; Fojo,
T.; Poruchynsky, M. S. J. Biol. Chem. 1997, 272, 17118-17125 and
references therein.
(5) (a) Rowinsky, E. K.; Eisenhauer, E. A.; Chaudhry, V.; Arbuck, S. G.;
Donehawer, R. C. Semin. Oncol. 1993, 20, 1-15. (b) Fletcher, B. S.;
Kujubadu, D. A.; Perrin, D. M.; Herschman, H. R. J. Biol. Chem. 1992,
267, 4338-4344. (c) Tsuji, M.; Dubois, R. N. Cell 1995, 83, 493-501.
(d) Essayan, D. M.; Kagey-Sobotka, A.; Colarusso, P. J.; Lichtenstein, L.
M.; Ozols, R. F.; King, E. D. J. Allergy Clin. Immunol. 1996, 97, 42-46.
(6) For reviews of epothilone chemistry and biology, see: (a) Harris, C. R.;
Danishefsky, S. J. J. Org. Chem. 1999, 64, 8434-8456. (b) Nicolaou, K.
C.; Roschangar, F.; Vourloumis, D. Angew. Chem., Int. Ed. 1998, 37, 2014-
2045. (c) Altmann, K.-H. Mini-ReV. Med. Chem. 2003, 3, 149-158. (d)
Nicolaou, K. C.; Ritze´n, V. A.; Namoto, K. Chem. Commun. 2001, 1523-
1535. (e) Harris, C. R.; Kuduk, S. D.; Danishefsky, S. J. Chemistry for the
21st Century 2001, 8. (f) He, L.; Orr, G. A.; Horwitz, S. B. Drug DiscoVery
Today 2001, 6, 1153-1164. (g) Rivkin, A.; Cho, Y. S.; Gabarda, A. E.;
Fumihiko, Y. J. Nat. Prod. 2004, 67, 139-143. (h) Florsheimer, A.;
Altmann, K. H. Expert Opin. Ther. Pat. 2001, 11, 951-968. (i) Wartmann,
M.; Altmann, K.-H. Curr. Med. Chem. 2002, 2, 123-148. (j) Altmann, K.
H.; Wartmann, M.; O’Reilly, V. Biochim. Biophys. Acta 2000, 1470, M79-
M91. (k) Wessjohann, L. A. Curr. Opin. Chem. Biol. 2000, 4, 303-309.
(7) (a) Balog, A.; Meng, D. F.; Kamenecka, T.; Bertinato, P.; Su, D.-S.;
Sorensen, E. J.; Danishefsky, S. J. Angew. Chem., Int. Ed. Engl. 1996, 35,
2801-2803. (b) Su, D.-S.; Meng, D. F.; Bertinato, P.; Balog, A.; Sorensen;
E. J.; Danishefsky, S. J.; Zheng, Y. H.; Chou, T.-C.; He, L. F.; Horwitz, S.
B. Angew. Chem., Int. Ed. Engl. 1997, 36, 757-759. (c) Meng, D. F.;
Bertinato, P.; Balog, A.; Su, D.-S.; Kamenecka, T.; Sorensen, E. J.;
Danishefsky, S. J. J. Am. Chem. Soc. 1997, 119, 10073-10092. For initial
reports of other epothilone syntheses, see: (d) Yang, Z.; He, Y.; Vour-
loumis, D.; Vallberg, H.; Nicolaou, K. C. Angew. Chem., Int. Ed. Engl.
1997, 36, 166-168. (e) Nicolaou, K. C.; Sarabia, F.; Ninkovic, S.; Yang,
Z. Angew. Chem., Int. Ed. Engl. 1997, 36, 525-527. (f) Nicolaou, K. C.;
Winssinger, N.; Pastor, J.; Nincovic, S.; Sarabia, F.; He, Y.; Vourloumis,
D.; Yang, Z.; Li, T.; Giannakakou, P.; Hamel, E. Nature 1997, 387, 268-
272. (g) Schinzer, D.; Limberg, A.; Bauer, A.; Bohm, O. M.; Cordes, M.
Angew. Chem., Int. Ed. Engl. 1997, 36, 523-524. (h) May, S. A.; Greico,
P. A. Chem. Commun. 1998, 1597-1598. (i) Sawada, D.; Shibasaki, M.
Angew. Chem., Int. Ed. Engl. 2000, 39, 209-213. (j) Martin, H. J.; Drescher,
M.; Mulzer, J. Angew. Chem., Int. Ed. 2000, 39, 581-583. (k) White, J.
D.; Carter, R. G.; Sundermann, K. F. J. Org. Chem. 1999, 64, 684-685.
(l) Zhu, B.; Panek, J. S. Org. Lett. 2000, 2, 2575-2578, (m) Bode, J. W.;
Carreira, E. M. J. Am. Chem. Soc. 2001, 123, 3611-3612. (n) Fu¨rstner,
A.; Mathes, C.; Grela, K. Chem. Commun. 2001, 12, 1057-1059. (o)
Taylor, R. E.; Chen, Y. Org. Lett. 2001, 3, 2221-2224. (p) Valluri, M.;
Hindupur, R. M.; Bijoy, P.; Labadie, G.; Jung, J. C.; Avery, M. A. Org.
Lett. 2001, 3, 3607-3070. (q) Ermolenko, M. S.; Potier, P. Tetrahedron
Lett. 2002, 43, 2895-2898.
Our particular interest in synthesizing and examining
epothilones with additional nuclear unsaturation, was first
prompted by epothilone 490 (12, (E)-10,11-dehydro-dEpoB).
Epothilone 490, a recently isolated natural product, which
(9) Chou, T.-C.; O’Connor, O. A.; Tong, W. P.; Guan, Y.; Zhang, Z.-G.;
Stachel, S. J.; Lee, C.; Danishefsky, S. J. Proc. Natl. Acad. Sci. U.S.A.
2001, 98, 8113-8118. For more information about clinical trials of dEpoB,
visit: www.kosan.com.
(10) (a) Rivkin, A.; Yoshimura, F.; Gabarda, A. E.; Chou, T.-C.; Dong, H.;
Tong, W. P.; Danishefsky, S. J. J. Am. Chem. Soc. 2003, 125, 2899-2901.
(b) Chou, T.-C.; Dong, H.; Rivkin, A.; Yoshimura, F.; Gabarda, A. E.;
Cho, Y. S.; Tong, W. P.; Danishefsky, S. J. Angew. Chem., Int. Ed. 2003,
42, 4762-4767. (c) Yoshimura, F.; Rivkin, A.; Gabarda, A. E.; Chou, T.-
C.; Dong, H.; Sukenick, G.; Danishefsky, S. J. Angew. Chem., Int. Ed.
2003, 42, 2518-2521. For recent examples of alternative potent epothilone
analogues, see: (d) Altmann, K. H.; Bold, G.; Caravatti, G.; Flo¨rsheimer,
A.; Guagnano, V.; Wartmann, M. Bioorg. Med. Chem. Lett. 2000, 10,
2765-2768. (e) Nicolaou, K. C.; Scarpelli, R.; Bollbuck, B.; Werschkun,
M. M. A.; Pereira, M.; Wartmann, K.-H.; Altmann, D.; Zaharevitz, R.;
Gussio, P.; Giannakakou, Chem. Biol. 2000, 7, 593-599.
(8) (a) Chou, T.-C.; Zhang, X. G.; Balog, A.; Su, D.; Meng, D. F.; Savin, K.;
Bertino, J. R.; Danishefsky, S. J. Proc. Natl. Acad. Sci. U.S.A. 1998, 95,
9642-9647. (b) Chou, T.-C.; Zhang, X. G.; Harris, C. R.; Kuduk, S. D.;
Balog, A.; Savin, K. A.; Bertino, J. R.; Danishefky, S. J. Proc. Natl. Acad.
Sci. U.S.A. 1998, 95, 15798-15802.
9
10914 J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004

(E)-9,10-Dehydroepothilones as Anticancer Agent
A R T I C L E S
Scheme 1. Synthesis of Epothilone 490 via a Ring-Closing Metathesis
Scheme 2. Synthesis of 17-Membered Epothilones via a Ring-Closing Metathesis
showed highly favorable cell culture cytotoxicity profiles, is
available only in small quantities from fermentation.11 Our
previous synthesis of dEpoB could be redirected to yield
epothilone 490 by a novel RCM strategy (see progression of 8
f 10 f 12 (Scheme 1)). Remarkably the 10,11-olefin con-
structed through RCM emerged cleanly in the E-configuration.
Epothilone 490 (12) was then converted to dEpoB (5) by
regioselective diimide reduction of the E-10,11-olefin. Unfor-
tunately, the in vivo performance of epothilone 490, in
xenografts, turned out to be surprisingly poor. This difficulty
was soon traced to unfavorable pharmacokinetic features of the
compound.
Though epothilone 490 (12) was no longer a development
candidate in our program, we postulated that the synthetic route
we had demonstrated for its synthesis could be applied toward
the synthesis 26-trifluoro-epo490 (14) or, indeed, trifluoro-
dEpoB (15).12 Accordingly, we synthesized RCM substrate 13
and tested the proposed ring-closing metathesis reaction. Su-
prisingly, at the time, a range of conditions which had been
used for the ring-closing metathesis reactions of the parent
methyl substrate 10 failed to provide the desired 14.13 Clearly
the nonimplementability of the RCM reaction in the context of
compound 13 was due to the presence of the fluorines at the
26-position.
To probe the scope of the deleterious effect of the trifluoro-
methyl group on the RCM reaction, we inserted a one carbon
“spacer” between this group and olefin linkage. Indeed, the ring-
closing metathesis of 16 provided the desired 17-membered
epothilone macrolide 17 in 57% yield, again exclusively in the
E-10,11 configuration.14 This was for us a telling experiment.
It suggested that whatever was the precise reason by which the
CF3 group at C12 abrogates the proposed RCM in the case of
13, it could be countered by inclusion of appropriate spacers.
(12) For a discussion of the effect of fluorine substitutents on biological activity,
see: (a) Smart, B. E. J. Fluorine Chem. 2001, 109, 3-11. (b) Ojima, I.;
Inoue, T.; Chakravarty, S. J. Fluorine Chem. 1999, 97, 3-10. (c) Newman,
R. A.; Yang, J.; Finlay, M. R. V.; Cabral, F.; Vourloumis, D.; Stephens,
L. C.; Troncoso, P.; Wu, X.; Logothetis, C. J.; Nicolaou, K. C.; Navone,
N. M. Cancer Chemother. Pharmacol. 2001, 48, 319-326.
(13) Rivkin, A.; Biswas, K.; Chou, T.-C.; Danishefsky, S. J. Org. Lett. 2002, 4,
4081-4084.
(11) Biswas, K.; Lin, H.; Njardarson, J. T.; Chappell, M. D., Chou, T.-C.; Guan,
Y.; Tong, W. P., He, L.; Horwitz, S. B.; Danishefsky, S. J. J. Am. Chem.
Soc. 2002, 124, 9825-9832.
(14) Rivkin, A.; Njardarson, J. T.; Biswas, K.; Chou, T.-C.; Danishefsky, S. J.
J. Org. Chem. 2002, 67, 7737-7740.
9
J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004 10915

A R T I C L E S
Rivkin et al.
Figure 3. Retrosynthesis of the (E)-9,10-dehydroepothilones and 26-trifluoro-dEpoB.
To gain a quick assessment of the effect of the C12-trifluoro-
Alkylation of the oxazolidinone 2417 with the readily
methyl group on the likely biological profile of the epothilones,
we prepared the 17-membered epothilone macrolide 19 in high
yield using again the stereoselective RCM reaction to produce,
exclusively, the E-9,10-olefin (Scheme 2).
In vitro and mouse plasma stability studies of compounds of
17 and 19 revealed some interesting facts. The first was that
the introduction of the three fluorine atoms at the 26-position
had improved the plasma stability of the drug 2-fold. The second
teaching was that both the 17-membered lactones 17 and 19
had retained their cytotoxicity due to the presence of the E-9,-
10-olefin. This was a significant and suggestive finding, since
the 17-membered epothilone without the 9,10-olefin is virtually
inactive.15
We reasoned that the synthesis of 26-trifluoro-dEpoB (15)
containing the usual 16-membered ring could be accomplished
via a highly convergent strategy, related to that employed in
the synthesis of 27-trifluoro-[17]ddEpoB (17). This strategy
envisioned the formation of a E-9,10-olefin via a ring-closing
metathesis reaction (Figure 3).16 We anticipated that chemo-
selective reduction of the E-9,10-olefin of 6 and 7 would furnish
dEpoB (5) and the desired 26-trifluoro-dEpoB (15). The RCM
precursor would be prepared by the union of the two fragments
(21 or 22) and 23 through an esterification reaction. Coupling
partner 23 would be constructed by deletion of the methylene
spacer group (found in our earlier routes cf. compound 16)
between the secondary methyl groups at C8. As noted above,
the in vitro level findings with the 17-membered epothilones
containing the skipped diene arrangement (see 17 and 19)
underscored the need for a corresponding investigation of the
biological consequences of such a diene in the familiar
16-membered lactone setting. Given the presence of the cis
12,13-olefin in a dEpoB context, such a skipped diene would
necessarily contain a 9,10-double bond.
synthesized trifluoro and methyl allyl iodides allowed for the
C15 stereocenter to be set in the appropriate absolute config-
uration18 with high diastereomeric access (see products 25 and
26 in Scheme 3). The latter were converted to their correspond-
ing Weinreb amides19 and thence to 27 and 28 en route to 21
and 22 by nucleophilic methylation (MeMgBr) and appropriate
Horner-Wittig olefinations.20 While the use of such oxazoli-
dinones as chiral auxiliaries had been pioneered by Evans and
associates,18 its application to the synthesis of optically defined
glycolates by alkylation (rather than by hydroxylation) had not
been developed.
The synthesis of the polypropionate fragment 34 was enabled
by two critical aldol reactions, which established the relative
configuration of the C3, C6, and C7 stereocenters (Scheme 4).
The first aldol reaction involved reaction of the Z-enolate of
the ethyl ketone 29 with Roche aldehyde 3021 to provide the
desired 31 with high diastereoselectivity. Recourse to 30,
allowed by this synthesis, is a significant advantage over the
use of earlier aldehydes which required resolution for the
attainment of enantiomerically pure starting materials.
Protection of the C7-alcohol followed by hydrolysis of the
acetal provided the desired aldehyde 32 and set the stage for
the second aldol reaction. Reaction of 32 with Duthaler’s DAG
(diacetone glucose) Ti-enolate22 afforded the desired hydroxy
(17) Lee, C. B.; Wu, Z.; Zhang, F.; Chappell, M. D.; Stachel, S. J.; Chou, T.-
C.; Guan, Y.; Danishefsky, S. J. J. Am. Chem. Soc. 2001, 123, 5249-
5259.
(18) (a) Evans, D. A.; Morrissey, M. M.; Dorow, R. L. J. Am. Chem. Soc. 1985,
107, 4346-4348. (b) Paterson, I.; Bower, S.; McLeod, M. D. Tetrahedron
Lett. 1995, 36, 175-178.
(19) (a) Nahm, S.; Weinreb, S. M. Tetrahedron Lett. 1981, 22, 3815-3818. (b)
Levin, J. I.; Turos, E.; Weinreb, S. M. Synth. Commun. 1982, 12, 989-
993.
(20) (a) Lythgoe, B.; Nambudiry, M. E. N.; Ruston, S.; Tideswell, J.; Wright,
P. W. Tetrahedron Lett. 1975, 40, 3863-3866. (b) Lythgoe, B. Chem. Soc.
ReV. 1980, 449-475. (c) Toh, H. T.; Okamura, W. H. J. Org. Chem. 1983,
48, 1414-1417. (d) Baggiolini, E. G.; Iacobelli, J. A.; Hennessy, B. M.;
Batcho, A. D.; Sereno, J. F.; Uskokovic, M. R. J. Org. Chem. 1986, 51,
3098-3108.
(21) (a) Cohen, N.; Eichel, W. F.; Lopresti, R. J.; Neukom, C.; Saucy, G. J.
Org. Chem. 1976, 41, 3505-3511. (b) Nagaoka, H.; Kishi, Y. Tetrahedron
1981, 37, 3873-3888. (c) Roush, W. R.; Palkowitz, A. D.; Ando, K. J.
Am. Chem. Soc. 1990, 112, 6348-6359.
(15) Nicolaou, K. C.; Sarabia, F.; Ninkovic, S.; Ray, M.; Finlay, V.; Boddy, C.
N. C. Angew. Chem., Int. Ed. 1998, 37, 81-84. Introduction of the
corresponding 9,10-olefin in the context of a 18-membered epothilone did
not lead to retention of the cytotoxicity.
(16) For an early instance of macrocyclization in a complex setting via a RCM,
see: Sinha, S. C.; Sun, J. Angew. Chem., Int. Ed. 2002, 41, 1381-1383.
9
10916 J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004

(E)-9,10-Dehydroepothilones as Anticancer Agent
A R T I C L E S
Scheme 3. Synthesis of Fragments 21 and 22
reductions of the (E)-9,10-olefins validated the structures of the
various synthetic intermediates described above, thereby prompt-
ing re-evaluation of previous assignments in the literature.24
Examination of synthetic analogues (6, 7, and 45), in cell
culture settings (see Supporting Information), revealed them to
exert increased potency on various sensitive and MDR tumor
cell lines that are exhibited by our current clinical candidate,
dEpoB (5). The impressive cell growth inhibition exhibited by
epothilones 6, 7, and 45 across a range of various drug-resistant
tumors prompted determination of the blood plasma stability
of these new (E)-9,10 congeners. We recall that epothilone 490
(cf. compound 12 with a methyl group at C-12) exhibits very
poor plasma stability. Indeed, it was this plasma instability which
had blocked further development of 12. By contrast, on exposure
of 6, 7, and 45 to murine plasma, we observed a much slower
drug degradation as compared to dEpoB (5) (by a factor of ca.
7). This stability constitutes a substantial advance, from a drug
availability perspective, relative to dEpoB (5), not to speak of
epothilone 490 (12). Based on the preliminary cell culture and
pharmacokinetic data of the (E)-9,10-dehydro derivatives 6 and
7, it would be appropriate to advance them for in vivo
investigations. Such studies are, of course, rather more sensitive
to drug availability than in vitro measurements.10
Scheme 4. Synthesis of Acid 34
This requirement, and indeed the possible eventual need to
prepare multigram quantities of 9,10-dehdyro derivatives for
further development, prompted a significant reassessment of our
total synthesis route. Of course, the single most serious problem
was that the RCM reaction on 36-38 produced 39, 41, and 43
only as minor products. The major pathway involved an RCM
reaction which was strictly confined to the O-alkyl sector of
36-38, leading primarily to the undesired 40, 42, and 44.
Accordingly, it was decided to attempt to defer introduction of
the thiazole (by olefination) after the RCM (vide infra).
With eventual processiblity in multigram scales as our goal,
the syntheses of the alkyl and acyl fragments entering into the
RCM reaction were restructured (Scheme 6). Compound 48 was
readily synthesized as shown. It was used to alkylate oxazoli-
dinone 24 in high diastereomeric excess. Following deprotection
of the OTES group and nucleophilic methylation, compound
50 was in hand. This R-hydroxyketone would serve as the acyl
group acceptor in formation of the central ester to prepare all
the critical RCM precursors. An obvious concern was the
possible vulnerability of such a hydroxyketone as an acyl
acceptor to partial racemization or diversion to regioisomeric
R-ketols.
tert butyl ester 33 with very high (>95%) diastereoselectivity.
The latter was then converted to the desired acid 34 in
straightforward steps.
The allylic alcohols 21, 22, and the C1-C9 acid fragment 34
were united through an EDCI esterification protocol, thus
providing the RCM precursors 36 and 37, respectively (Scheme
5). Ring-closing metathesis reactions of 36 and 37 were carried
out using catalyst 1123 in toluene. These reactions did indeed
provide the trans isomers 39 and 41. However, the major
products were 40 and 42 which had arisen from the obvious
alternate RCM pathway. These unwanted RCM isomers pre-
dominated over the desired 39 and 41 by ratios of ca. 3:1.
Finally, deprotection of the silyl ethers of 39 and 41 with HF-
pyridine led to the desired (E)-9,10-dehydroepothilones 624 and
7. As planned, the latter were converted to dEpoB (5) and 26-
trifluoro-dEpoB (15) via diimide reduction of the E-9,10-
olefin.25
A serious problem in our earlier synthesis of the acid fragment
34 was the very expensive and technically demanding Duthaler
chemistry22 to generate the desired S stereochemistry at the C3.
To circumvent this problem, the aldol reaction was carried out
without any chiral auxiliary to provide a 1:1 mixture of the
corresponding â-hydroxy ketone (Scheme 7). Reagent-controlled
(24) (a) White, J. D.; Carter, R. G.; Sundermann, K. F.; Wartmann, M. J. Am.
Chem. Soc. 2001, 123, 5407-5413. (b) White, J. D.; Carter, R. G.;
Sundermann, K. F.; Wartmann, M. J. Am. Chem. Soc. (Addition/Correction)
2003, 125, 3190-3190. With compound 6 of rigorously proven structure
in hand, we were surprised to find that its spectral properties were not
congruent with those previously reported (see ref 24a) for a compound
presumed to be the same entity. The actual structure of the compound
previously assigned as 6 has now been re-evaluated (see ref 24b). In
retrospect, it is clear that 6 had not been previously prepared and, in fact,
the whole family of (E)-9,10-dehydroepothilones reported here is a new
genus.
By corresponding methodology, we synthesized the (E)-9,-
10-dehydro-dEpoF (45) (see Scheme 5). The selective diimide
(22) Duthaler, R. O.; Herold, P.; Lottenbach, W.; Oretle, K.; Reidiker, M. Angew.
Chem., Int. Ed. Engl. 1989, 28, 495-497.
(23) Initial report: (a) Scholl, M.; Trnka, T. M.; Morgan, J. P.; Grubbs, R. H.
Tetrahedron Lett. 1999, 40, 2247-2250. RCM reviews: (b) Grubbs, R.
H.; Miller, S. J.; Fu, G. C. Acc. Chem. Res. 1995, 28, 446-452. (c) Trnka,
T. M.; Grubbs, R. H. Acc. Chem. Res. 2001, 34, 18-29. (d) Alkene
Metathesis in Organic Chemistry; Fu¨rstner, A., Ed.; Springer: Berlin, 1998.
(e) Fu¨rstner, A. Angew. Chem., Int. Ed. 2000, 39, 3012-3043. (f) Schrock,
R. R. Top. Organomet. Chem. 1998, 1, 1-36.
(25) (a) Corey, E. J.; Mock, W. L.; Pasto, D. J. Tetrahedron Lett. 1961, 2, 347-
352. (b) Pasto, D. J.; Taylor, R. T. Org. React. 1991, 40, 91-155.
9
J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004 10917

A R T I C L E S
Rivkin et al.
Scheme 5. Synthesis of (E)-9,10-Dehydroepothilones
Scheme 6. Processible Synthesis of Fragment 50
toluene. This reaction, now uncomplicated by alternate metath-
esis pathways, provided exclusively the trans isomers 57 and
58 in high yields. Fortunately, installation of the thiazole moiety
via a Wittig reaction proceeded with high E/Z selectivity and
yield to provide 6 and 7 following deprotection of the two silyl
ethers.27
We considered whether the incorporation of C9-C10 olefin
in epothilone B (2, EpoB) would alter its biological profile in
the same direction as was the case with its 12,13-desoxy
counterparts. Toward this end, we studied the epoxidation of 6
with 2,2′-dimethydioxirane (DMDO). The reaction indeed
proceeded with high chemoselectivity at the more substituted
C12-C13 olefin. There was obtained an 87% yield of a 1:2.6
ratio of the (E)-9,10-dehydroepothilone B (59) and its diaste-
reomer bearing the R-12,13-oxirane (structure not shown).28 In
vitro studies with 59 (whose configurations at C12 and C13
was established by its reduction to afford Epo B) revealed it to
be roughly 2-4-fold more potent than the parent EpoB (2) in
various cell lines. While compound 59 proved to be the most
potent epothilone we encountered in our program, its narrow
therapeutic index in xenografts, as well as its difficult acces-
sibility (vide supra), served to reduce its priority for the
preclinical development. Interestingly, the unnatural R-oxirane
possessed a substantially lower in vitro activity.
asymmetric reduction of the derived keto function (see com-
pound 52) using Noyori conditions26 generated the desired S
stereochemistry at the C3 in high diastereomeric excess. The
now available â-hydroxy ester 53 was transformed to acid 34
in several steps following earlier protocols.10a
Remarkably, esterification of the resultant hydroxyketones
50 and 54 with the C1-C9 acid fragment 34 provided the
corresponding RCM cyclization precursors 55 and 56 (Scheme
8) without noticeable racemization at C15, or loss of integrity
of the initial R-ketol linkage. The ring-closing metathesis
reaction of 55 and 56 was carried out using catalyst 1123 in
Another advantage of the restructured (second generation)
synthesis described above is that a variety of heterocycles can
(27) Hindupur, R. M.; Panicker, B.; Valluri, M.; Avery, M. A. Tetrahedron
Lett. 2000, 42, 7341-7344. Attempts to apply Avery’s protocol for the
installation of the thiazole gave the desired product in low yield and with
poor E/Z selectivity.
(26) Noyori, R.; Ohkuma, T.; Kitamura, M.; Takaya, H.; Sayo, N.; Kumobayashi,
H.; Akutagawa, S. J. Am. Chem. Soc. 1987, 109, 5856-5858.
(28) Stachel, S. J.; Danishefsky, S. J. Tetrahedron Lett. 2001, 42, 6785-6787.
9
10918 J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004

(E)-9,10-Dehydroepothilones as Anticancer Agent
A R T I C L E S
Scheme 7. Processible Synthesis of Acid 34
Scheme 8. Processible Synthesis of (E)-9,10-Dehydroepothilones
Scheme 9. Diversification of C21 of (E)-9,10-Dehydroepothilones
be installed via ketone intermediates 57 and 58. This point is
well highlighted by the synthesis of (E)-9,10-dehydro-dEpoF
(Scheme 9). Wittig reaction of the ketone with the appropriate
phosphonium ylides afforded the desired (E)-9,10-dehydro-
dEpoF compounds 45 and 60 in high yield and with high E/Z
selectivity. Furthermore, we were able to efficiently convert the
21-hydroxy 60 to derivatives of the type 62 and 63, containing
amino functionality at C21 in several steps as shown in Scheme
9.
The fully synthetic epothilone analogues have been evaluated
against a variety of cell types to determine their antitumor
potential. As shown in tabular form in the Supporting Informa-
tion, all of the compounds exhibited high cytotoxic activity
against a variety of sensitive and resistant tumor cell lines.29
9
J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004 10919

A R T I C L E S
Rivkin et al.
Figure 4. Therapy of extra large MX-1 tumor xenograft. MX-1 tumor tissue (50 mg) was implanted sc on day 0. On day 22 (D22) when tumor size reached
960 ( 132 mg (about 3.4% of body weight), Fludelone 25 mg/kg, 6 hr iv infusion, Q3Dx5 was given on D22, D25, D28, D31, and D34 as indicated by
arrows. The second cycle of treatment, following 9 day rest, was given on D43, D46, D49, and D52. (a) Tumor size changes in the vehicle treated control
(b) and Fludelone treated group (0) (n ) 5 each). Observation was continued Q3D up to D180 (128 days following cessation of treatment on D52). b.
Photographs for the nude mice (one mouse each selected from the control group and the treated group) taken on D25, D31, D37, D43, and D52. No relapse
was observed on D180.
The most salient features of our findings are as follow: One
can expect a loss of ca. an order of magnitude in replacement
of the C12-C13 â-epoxide by an E-12,13 double bond
(compare EpoB and dEpoB in the sensitive CCRF-CEM cell
line). Another expectation, is that inclusion of an E-9,10 double
bond in addition to the Z-12,13-olefin leads to a significant
increase in cytotoxicity across several cell lines.
Still another instructive trend was seen in comparing 26-
trifluoro-(E)-9,10-dehydro-dEpoB (7) (Fludelone) with the cor-
responding (E)-9,10-dehydro compound 6. Inclusion of the three
fluorine atoms at C26 attenuates cytotoxicity of 7 by up to a
factor of 5 relative to 6 in CCRF-CEM leukemic cells. This
attenuation effect of the 12-trifluoromethyl function is also seen
in compounds lacking the 9,10-unsaturation (compare dEpoB
(5) and 26-trifluoro-dEpoB (15)).
Given these data, and given the accessibility of these 9,10-
dehydro compounds (including 12-trifluoromethyl congeners)
through chemical synthesis, we were in a position to initiate in
vivo experiments on our most promising compounds. We
describe here some particularly striking and promising results
with compound 7 (Fludelone), which has emerged as a most
exciting possibility for advancement to clinical evaluation. In
vivo experiments were carried out using human tumor xe-
nografts in immunodeficient nude mice. For all their imperfec-
tions, such models in oncology are widely used in evaluating30
the potential of antitumor lead compounds en route to clinical
development.
(29) See ref 9a-c for the preliminary therapeutic evaluations of (E-9,10-
dehydroepothilones. A comprehensive account of the therapeutic evaluations
of the (E)-9,10-dehydroepothilones will be reported in due course.
9
10920 J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004

(E)-9,10-Dehydroepothilones as Anticancer Agent
A R T I C L E S
Figure 5. Therapeutic effects against human mammary carcinoma MX-1 xenograft by Fludelone or paclitaxel (Taxol) via oral administration. Female nude
mice were used. Fludelone 30 mg/kg (0) (n ) 3) was given orally Q2Dx7 beginning D16 after tumor implantation and then Q2Dx9 on D32-48, as indicated
by arrows. All three mice’s tumor disappeared on D40, 45, and 48. For consolidation therapy, the third cycle of treatment was given Q2Dx5 from D58-66
when all mice were tumor free on D48. There was no relapse on D115 (49 days after stopping treatment). Control (b) (n ) 2) received vehicle only. Parallel
comparative experiment was carried out with Paclitaxel 30 mg/kg (4) (n ) 3), orally beginning D16, Q2Dx3 and then the dose was increased to 40 mg/kg,
Q2Dx3 (D22-26) and then to 60 mg/kg, Q2Dx3 (D28-40).
Figure 6. Therapeutic effects against the drug resistant human T-cell lymphoblastic leukemia CCRF-CEM/Paclitaxel xenograft by Fludelone and Paclitaxel.
Tumor tissue of CCRF-CEM/Paclitaxel (44-fold resistant in vitro), 50 mg/mouse was implanted sc into nude mice on day 0. Treatment of 6 hr iv infusion
started on D8 with Fludelone 15 mg/kg (O) (n ) 3) and 30 mg/kg (4) (n ) 4) or Paclitaxel 20 mg/kg (O) (n ) 4). Q2Dx7 (D8-D20) skipped D22 dose and
then resumed treatment Q2Dx5 on D24, 26, 28, 30, and 32, as indicated by arrows. The control group (O) (n ) 4) received vehicle only. For Fludelone at
1
3
15 mg/kg, /3 of mice’s tumor disappeared on D37, and at 30 mg/kg, /4 of mice’s tumor disappeared on D22 and 32).
Remarkably, treatment of MX-1 xenografts with 25 mg/kg
dosages of Fludelone resulted in complete tumor disappearance
and the absence of any relapse for over 4 months after
suspension of treatment (See Figure 4). Most importantly, these
therapeutic successes can be achieved either by 6 h iv infusion
or by oral administration (See Figure 5). On the other hand,
treatment of the MX-1 xenografts by oral administration of
paclitaxel did not shrink the tumor (see Figure 5). It goes without
saying that if this was translatable to the human clinical setting,
achievement of oral activity could be of significant advantage.
Paclitaxel-resistant tumor xenografts (Figure 6) as well as
human colon carcinoma (HCT-116, Figure 7) can also be cured
with Fludelone by iv infusion. The experiments using human
mammary carcinoma (MX-1) and human colon carcinoma
(HCT-116) xenografts in nude mice lasted 6.0 and 6.6 months,
respectively. There was no tumor relapse in either experiment
during 4.3 and 5.3 months, respectively, following the cessation
of treatment. For the HCT-116 experiment, paclitaxel and
Fludelone were compared at 20 mg/kg and both achieved tumor
disappearance. The paclitaxel-treated group relapsed 1.1 months
(30) For reports that suggest correlations between xenograft data and clinical
activity are good, see ref 30a. For an opposite view, see ref 30b. (a) Fiebig,
H. H.; Berger, D. P. Preclinical Phase II trials. In The Nude Mouse in
Oncology Research; Boven, E., Winograd, B., Eds.; CRC Press: Boca
Raton, FL, 1995; p 318. (b) Johnson, J. I.; Decker, S.; Zaharevitz, D. Br.
J. Cancer 2001, 84, 1424-1431.
9
J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004 10921

A R T I C L E S
Rivkin et al.
Figure 7. Therapeutic effects against human colon carcinoma HCT-116 xenograft by Fludelone and Paclitaxel. HCT-116 tumor tissue 50 mg/mouse was
implanted sc into nude mice on day 0. Treatment of Q2Dx4, 6 hr iv fusion for three cycles was carried out on (D9, 11, 13, 15), (D19, 21, 23, 25), and (D31,
34, 35, 37) with Fludelone 20 mg/kg (4), 30 mg/kg (0), and Paclitaxel 20 mg/kg (O) (n ) 4 each). Complete tumor disappearance in all mice occurred on
(D33, 35, 41, 45), (D21, 23, 33, 41), and (D33, 33, 41, 45) for Fludelone 20 mg/kg, 30 mg/kg and Paclitaxel 20 mg/kg, respectively. There was no tumor
relapse for both of the Fludelone treated groups on D200. However, the Paclitaxel treated group (O) suffered relapses on D71, 75, and 81 which represent
the relapses on the 34th, 38th, and 41st day after stopping treatment.
after treatment was discontinued, whereas Fludelone-treated
animals were tumor free for over 5.3 months.
synthesis,31 even in multistep settings. Extensive research into
the (E)-9,10-dehydroepothilone B family, in the context of
chemistry, biology and oncology, continues.
These results have involved a particularly long and thorough
therapeutic study using xenografts and report remarkably long
periods of complete remission with parenteral or oral admin-
istration of a single antitumor agent. Achievement of complete
tumor disappearance and long term remission by oral treatment
could well be of particularly great clinical significance if
translatable to the human setting.
Acknowledgment. This research was supported by the
National Institutes of Health (Grant Numbers: CA-28824
(S.J.D.), CA-08748 (T.C.C.). Postdoctoral fellowship support
is gratefully acknowledged be A.R. (NIH Cancer Pharmacology
Training Grant, T32-CA62948), F.Y. (Uehara Memorial Foun-
dation), and A.E.G. (Goodwin Fellow, Therapeutics Center of
Memorial Sloan-Kettering Cancer Center). The authors wish
to thank Dr. George Sukenick (NMR Core Facility, Sloan-
Kettering Institute, and CA-08748) for mass spectral and NMR
analysis.
Conclusion
Only advancement to a human clinical setting will establish
whether Fludelone will become a valuable resource for oncolo-
gists in the treatment of patients. That this compound was
discovered is a consequence of a close linkage between the
disciplines of organic chemistry, pharmacokinetics, and cancer
pharmacology. That a development program on Fludelone can
now be undertaken realistically speaks particularly well as to
the power of chemical synthesis. Parenthetically, we note that
the manner in which the (E)-9,10-dehydroepothilones were
discovered underscores the value of natural product leads in
drug discovery and the potential applicability of diverted total
Supporting Information Available: Experimental details for
preparation and spectral characteristics ofselected compounds
and in vitro cytotoxicity data in tabular form. This material is
available free of charge via the Internet at http://pubs.acs.org.
JA046992G
(31) cf. Njardarson, J. T.; Gaul, C.; Shan, D.; Huang, X.-Y.; Danishefsky, S. J.
J. Am. Chem. Soc. 2004, 126, 1038-1040.
9
10922 J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004