Synthesis and SAR of C12-C13-oxazoline derivatives of epothilone A

Article abstract of DOI:10.1016/j.bmcl.2009.04.112

The SAR of a series of new epothilone A derivatives with a 2-substituted-1,3-oxazoline moiety trans-fused to the C12-C13 bond of the deoxy macrocycle have been investigated with regard to tubulin polymerization induction and cancer cell growth inhibition.

Full text of DOI:10.1016/j.bmcl.2009.04.112

Bioorganic & Medicinal Chemistry Letters 19 (2009) 3760–3763
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Synthesis and SAR of C12–C13-oxazoline derivatives of epothilone A
*
Bernhard Pfeiffer, Kurt Hauenstein, Philipp Merz, Jürg Gertsch, Karl-Heinz Altmann
Swiss Federal Institute of Technology (ETH) Zürich, HCI H405, Wolfgang-Pauli-Str. 10, CH-8093 Zürich, Switzerland
a r t i c l e i n f o
a b s t r a c t
Article history:
The SAR of a series of new epothilone A derivatives with a 2-substituted-1,3-oxazoline moiety trans-fused
to the C12–C13 bond of the deoxy macrocycle have been investigated with regard to tubulin polymeri-
zation induction and cancer cell growth inhibition. Significant differences in antiproliferative activity
were observed between different analogs, depending on the nature of the substituent at the 2-position
of the oxazoline ring. The most potent compounds showed comparable activity with the natural product
epothilone A. Modeling studies provide a preliminary rationale for the observed SAR.
Ó 2009 Elsevier Ltd. All rights reserved.
Received 7 April 2009
Revised 22 April 2009
Accepted 23 April 2009
Available online 3 May 2009
Keywords:
Epothilone
Natural product
Tubulin
Anticancer
SAR
Epothilones A and B (Epo A and B; Fig. 1) are naturally occurring
microtubule-stabilizing macrolides that were first isolated from
the myxobacterium Sorangium cellulosum by Reichenbach, Höfle,
and co-workers in 1987.1,2 Epothilones are potent inhibitors of hu-
man cancer cell growth in vitro and in vivo2,3 and they have served
as successful lead structures for the development of new antican-
cer drugs.4 Thus, at least seven epothilone-derived agents have
entered clinical evaluation in humans (including the natural prod-
uct Epo B) and the most advanced of these analogs, the Epo B lac-
tam BMS-247550 (ixabepilone), has recently obtained FDA
approval for clinical use in cancer patients.5
compared with Epo A.9,10 In addition, the activity of these analogs
is independent of the geometry of the ring fusion (cis (2a) or trans
(2b)), if C13 is present in the natural S-configuration; the corre-
sponding 13R-isomers are substantially less potent.9,10
Rather surprisingly, these findings have not been followed up in
any systematic fashion and little is known at this point about the
SAR of bicyclic epothilone analogs related to 2. In light of this
R
R
O
13
S
S
12
Extensive studies on the epothilone SAR have led to a detailed
(empirical) understanding of the structural requirements for bio-
logical activity and modifications have been identified that are
associated with enhanced potency (relative to the natural prod-
ucts).4 These studies also revealed that deoxyepothilones (i.e.,
Epo C and Epo D; Fig. 1), which incorporate a Z double bond
between C12 and C13 in place of a cis-epoxide moiety, possess sim-
ilar antiproliferative activity as the epoxide-based parent
compounds Epo A and B.6,7 In contrast, Epo C/D analogs with a phe-
nyl8 or imidazole9 moiety fused to the C12–C13 bond have been
reported to be virtually inactive, thus indicating that conforma-
tional restriction of the C12–C13 bond to a cis-geometry by itself
does not suffice to maintain potent biological activity. Surprisingly,
however, the additional steric bulk associated with C12–C13-fused
rings appears to be much better tolerated for non-planar structures,
as the activity of cyclic acetals 2 (Fig. 2) is reduced less than 10-fold
HO
HO
N
N
15
O
1
O
O
O
OH
OH
O
O
R = H:
R = H:
Epo A
Epo B
Epo C
R = CH3:
Epo D
R = CH3:
Figure 1.
O
O
O
O
O
S
S
HO
HO
N
N
O
O
O
OH
OH
O
O
2a
2b
* Corresponding author.
E-mail address: karl-heinz.altmann@pharma.ethz.ch (K.-H. Altmann).
Figure 2.
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.04.112

B. Pfeiffer et al. / Bioorg. Med. Chem. Lett. 19 (2009) 3760–3763
3761
NH+
R = CH3: 1b
R = OCH3: 1a
R
2Cl
N
O
O
R
O
R
S
R =
:1c
N
R =
:1d
:1f
a)
HO
N
S
4
O
HO
N
R =
R =
:1e R =
O
OH O
O
1d-1h
O
1
OH
O
:1g
R =
:1h
Scheme 2. Reagents and conditions: (a) Compound 1d–f: DCE, EtOH (cat.), 90 °C, 2–
6 h, 17–23%. Compound 1g: DCE, EtOH (cat.), mw, 150 °C, 15 min, 21%. Compound
1h: DCE, EtOH (cat.), Et3N, mw, 150 °C, 15 min, 21%.
N
Figure 3.
Using this procedure 1d–h were obtained in moderate yields
(17–23%), but without the need for prior ortho ester preparation
(which had proven to be tedious). In the case of 1g and 1h micro-
wave heating was employed in an attempt to improve the yield of
the desired oxazolines. While no yield improvement could be
achieved, microwave heating at least led to significantly reduced
reaction times.
X
O
OH
S
S
HO
HO
N
N
O
O
a)
O
OH
O
O
OH O
Epothilone analogs 1a–h were evaluated for their ability to
induce the assembly of soluble tubulin heterodimers and to inhibit
human cancer cell proliferation in vitro. Compounds 1a and 1b,
which incorporate a small methoxy or methyl substituent at the
2-position of the oxazoline ring, respectively, did not show any
profound antiproliferative activity and were not further investi-
Epo A
X = N :
3
3
b)
X = NH :
4
2
R
O
N
S
N
HO
c)
gated (IC50 values >1 lM against the human cervix carcinoma cell
O
line KB-31; data not shown). In contrast, phenyl derivative 1c was
found to inhibit human cancer cell growth in vitro with IC50 values
around 20 nM (Table 1); thus, the activity of this analog is within a
10-fold range of the activity of Epo A and it is comparable with the
activity of cyclic acetals 2a and 2b (Fig. 2; IC50 values of 30 nM and
23 nM against the human cervix carcinoma cell line KB-31 have
been reported for 2a and 2b, respectively.9 The IC50 of 1c against
this cell line is 30 nM). The tubulin-polymerizing activity of 1c is
comparable with that of Epo A, at least under the experimental
conditions employed in this work (Table 1). It should be remem-
bered, however, that tubulin polymerization data of the type
shown in Table 1 are a relatively crude measure for the interaction
of ligands with the tubulin/microtubule system and do not allow to
differentiate between compounds with potent tubulin-polymeriz-
ing activity.
The introduction of a methyl substituent at the p-position of the
phenyl moiety leads to a substantial increase in cellular potency
and the antiproliferative activity of the corresponding analog 1d
is now comparable with that of Epo A (Table 1). A further increase
in the size of the p-substituent, however, is associated with a grad-
ual loss in biological potency, at the level of tubulin polymerization
as well as cancer cell growth inhibition, such that the p-tert-butyl
phenyl derivative 1g no longer exhibits any meaningful antiprolif-
1a-c
O
OH O
Scheme 1. Reagents and conditions: (a) LiN3, NH4Cl, DMF, 85 °C, 24 h, 38%; (b)
Ph3P, THF/H2O 15/1, rt, 88 h, 50%; (c) compound 1a: C(OCH3)4, DCE, 90 °C, 21%.
Compound 1b: CH3(OC2H5)3, DCE, TFA, 90 °C, 23%. Compound 1c: PhC(OC2H5)3,
DCE, 90 °C, 69%.
situation and in order to delineate the biological effects of other
five-membered heterocycles fused to C12–C13 in a non-planar
arrangement and, in particular, to assess the impact of substituents
on the five-membered ring, we have embarked on the synthesis
and biological evaluation of bicyclic epothilone analogs of the
general structure 1 (Fig. 3).
The synthesis of Epo A-derived oxazolines 1 was based on ami-
no alcohol 4 as the central intermediate (Scheme 1). As illustrated
in Scheme 1, 4 was obtained through nucleophilic ring-opening of
the epoxide moiety in Epo A with azide anion (to produce 3) and
subsequent reduction of the azide group under Staudinger condi-
tions (Ph3P/H2O).
The structure of azido alcohol 3 (with a 12-azido group) as the
major product of the reaction between Epo A and LiN3 in DMF in
the presence of NH4Cl (the conditions employed in this work)
was firmly established by means of NMR spectroscopy.11 The reg-
iochemical course of the epoxide opening reaction is thus identical
with that reported for the reaction of 12,13-bis-epi-Epo A with
NaN3 in EtOH.12
The elaboration of amino alcohol 4 into oxazolines 1a-c in-
volved reaction of this intermediate with commercially available
tetramethyl orthocarbonate, triethyl orthoacetate, and triethyl
orthobenzoate, respectively. Thus, heating of 4 (either as the free
base, for 1b, or as the hydroacetate, for 1a and 1c) produced ana-
logs 1a–c in yields between 21% and 69% (Scheme 1). Analogs
1d–h were obtained from amino alcohol 4 (as the hydroacetate)
and the appropriate crude imino ester hydrochlorides (prepared
from the corresponding nitriles by HCl-catalyzed methanolysis)
in refluxing dichloroethane (DCE) in the presence of catalytic
amounts of EtOH (Scheme 2; for structures cf. Fig. 3).
Table 1
Nubulin-polymerizing and antiproliferative activity of epothilone analogs 1c–h
Compound
% Tub-Pol.a (5
lM/2
lM)
IC50 A549b (nM)
IC50 MCF-7b (nM)
Epo A
1c
1d
1e
1f
89/83
80/65
88/71
36/32
15/14
7/0
2.2 0.3
16.4 2.7
2.9 0.3
378.5 14.8
4873 121
>10,000
2.6 0.3
19.2 1.4
1.9 0.1
297 5.2
5144 207
>10,000
1g
1h
93/78
0.92 0.01
1.3 0.1
a
Induction of polymerization of porcine brain microtubule protein by 5
M of test compound relative to the effect of 25 M Epo B.
Concentration required for 50% growth inhibition of the human lung and breast
carcinoma cells lines A549 and MCF-7, respectively (72 h exposure).
lM or
2
l
l
b

3762
B. Pfeiffer et al. / Bioorg. Med. Chem. Lett. 19 (2009) 3760–3763
Figure 4. (A) Structure of the taxol/epothilone binding cavity in b-tubulin with bound Epo A, as determined by electron crystallography (1TVK);14 (B) 1c docked into b-
TM
tubulin of 1TVK using Glide .13 For b-tubulin: surface (white), ribbon (cyan), V23, A231, F270, P358 as CPK (brown). For ligand structures: C, green; O, red; N, blue; S, yellow;
TM
polar H, white. The figure was produced with Maestro .17
erative activity (Table 1). Likewise, the compound has no signifi-
cant effect on tubulin assembly, which points to the existence of
a spatially restricted, well defined tubulin binding pocket for the
phenyl moiety at the 2-position of the oxazoline ring (vide infra).
The most potent analog investigated in this study appears to be
pyridine derivative 1h, whose antiproliferative activity is at least
comparable with that of Epo A or analog 1d. Thus, the replacement
of the phenyl ring in 1c by a 3-pyridyl moiety leads to a >10-fold
increase in cellular activity. As would be expected, 1h is also a po-
tent inducer of tubulin assembly, but the difference in polymeriza-
tion efficiency between 1c and 1h is clearly less pronounced than
the difference in antiproliferative activity (at least under the spe-
cific conditions of our experiments). It is, therefore, unclear to what
extent (if at all) the enhanced cellular activity of 1h (over 1c) may
be a result of higher affinity interactions with the tubulin/microtu-
bule system (possibly through H-bond formation between the pyr-
data shown in Table 1 are a direct measure of ligand affinity for
the tubulin/microtubule system, this is what is experimentally ob-
served. In agreement with this structural hypothesis, and quite
intriguingly, the inactive analog 1g could not be docked into the
taxol/epothilone binding pocket on b-tubulin with the aryl group
pointing into the hydrophobic cleft that is utilized by 1c–f.
In conclusion, we have prepared a series of new epothilone ana-
logs with a 2-substituted 1,3-oxazoline ring trans-fused to the
C12–C13 bond of the deoxy macrocycle. Two of these analogs show
cellular potencies that are at least comparable with Epo A, thus
confirming and extending previous findings on cyclic acetals 2. A
clear SAR could be discerned in the 2-phenyl oxazoline series with
regard to the size of the para-substituent and a preliminary struc-
tural model has been established that accounts for the observed
changes in tubulin polymerization activity. Future work will focus
on the investigation of a broader range of substituents on the oxaz-
oline ring and an improved understanding of the interactions of
these analogs with tubulin by computational and experimental
approaches.
idine nitrogen and
a donor group on the protein). Other
parameters, such as cellular uptake or intracellular distribution
might be equally (or even more) important for the difference in
cellular potency between 1c and 1h.
In an attempt to elaborate a rationale for the observed SAR in
the 2-phenyl-oxazoline series at the level of tubulin binding, ana-
logs 1c–g were docked into the taxol/epothilone binding cavity on
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2009.04.112.
13
TM
b-tubulin using the program Glide
and the b-tubulin structure
that has been derived from an Epo A/b-tubulin complex (as part
of a two-dimensional tubulin polymer sheet) by means of electron
crystallography (1TVK; Fig. 4A).14
References and notes
1. (a) Höfle, G.; Reichenbach, H. In Anticancer Agents from Natural Products; Cragg,
G. M., Kingston, D. G. I., Newman, D. J., Eds.; CRC: Boca Raton, 2005; pp 413–
450; (b) Gerth, K.; Bedorf, N.; Höfle, G.; Irschik, H.; Reichenbach, H. J. Antibiot.
1996, 49, 560.
2. For the discovery of the microtubule-stabilizing properties of epothilones see:
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.
3. (a) Altmann, K.-H.; Wartmann, M.; O’Reilly, T. Biochim. Biophys. Acta Rev. Cancer
2000, 1470, M79; (b) Kowalski, R. J.; Giannakakou, P.; Hamel, E. J. Biol. Chem.
1997, 272, 2534.
4. For reviews on the epothilone SAR see, e.g.: (a) Altmann, K.-H.; Pfeiffer, B.;
Arseniyadis, S.; Pratt, B. A.; Nicolaou, K. C. ChemMedChem. 2007, 2, 396; (b) Ref.
1a.; (c) Rivkin, A.; Chou, T. C.; Danishefsky, S. J. Angew. Chem., Int. Ed. 2005, 44,
2838; (d) Wartmann, M.; Altmann, K.-H. Curr. Med. Chem: Anti-Cancer Agents
2002, 2, 123; (e) Nicolaou, K. C.; Roschangar, F.; Vourloumis, D. Angew. Chem.,
Int. Ed. 1998, 37, 2014.
For each ligand a set of 100–150 distinct low energy structures
were generated by means of MD simulations, energy minimization,
and conformational averaging that were individually submitted to
the Glide 13 docking procedure.15 The docking itself was restricted
TM
by the requirement for a <3 Å deviation between the positions of
the thiazole ring in the docked structures and the experimental
Epo A/b-tubulin structure 1TVK.14
For analogs 1c–f the lowest energy scores yielded ligand orien-
tations as that depicted in Figure 4B for 1c, with the phenyl moiety
entering into a hydrophobic pocket that is defined by residues V23,
A231, F270, and P358 of b-tubulin. The same pocket has been
found to be occupied by the C-30 phenyl group of taxol in a previ-
ously reported electron crystallography structure of a taxol/b-
tubulin complex.16 With increasing steric bulk of the p-alkyl sub-
stituent on the phenyl ring, steric congestion within the pocket
would be predicted to increase, thus leading to a gradual loss in
binding affinity. If one assumes that the tubulin polymerization
5. Kaminskas, E.; Jiang, X.; Aziz, R.; Bullock, J.; Kasliwal, R.; Harapanhalli, R.; Pope,
S.; Sridhara, R.; Leighton, J.; Booth, B.; Dagher, R.; Justice, R.; Pazdur, R. Clin.
Cancer Res. 2008, 14, 4378.
6. (a) Nicolaou, K. C.; Winssinger, N.; Pastor, J.; Ninkovic, S.; Sarabia, F.; He, Y.;
Vourloumis, D.; Yang, Z.; Li, T.; Giannakakou, P.; Hamel, E. Nature 1997, 387,
268; (b) Nicolaou, K. C.; Vourloumis, D. T.; Li, H.; Pastor, J.; Winssinger, N.; He,

B. Pfeiffer et al. / Bioorg. Med. Chem. Lett. 19 (2009) 3760–3763
3763
Y.; Ninkovic, S.; Sarabia, F.; Vallberg, H.; Roschangar, F.; King, N. P.; Finlay, M. R.
V.; Giannakakou, P.; Verdier-Pinard, P.; Hamel, E. Angew. Chem., Int. Ed. 1997,
36, 2097.
9. Altmann, K.-H.; Bold, G.; Caravatti, G.; End, N.; Flörsheimer, A.; Guagnano, V.;
O’Reilly, T.; Wartmann, M. Chimia 2000, 54, 612.
10. Sefkow, M.; Kiffe, M.; Höfle, G. Bioorg. Med. Chem. Lett. 1998, 8, 3031.
11. For details see Supplementary data.
12. Regueiro-Ren, A.; Borzilleri, R. M.; Zheng, X.; Kim, S.-H.; Johnson, J. A.; Fairchild,
C. R.; Lee, F. Y. F.; Long, B. H. G.; Vite, D. Org. Lett. 2001, 3, 2693.
13. Glide, version 4.5, Schrödinger, LLC, New York, NY, 2007.
14. Nettles, J. H.; Li, H.; Cornett, B.; Krahn, J. M.; Snyder, J. P.; Downing, K. H. Science
2004, 305, 866.
15. Details of the computational protocols can be found in Supplementary data.
16. Nogales, E.; Wolf, S. G.; Downing, K.-H. Nature 1998, 391, 199.
17. Maestro, version 8.0, Schrödinger, LLC, New York, NY, 2007.
7. (a) Su, D.-S.; Meng, D.; Bertinato, P.; Balog, A.; Sorensen, E. J.; Danishefsky, S. J.;
Zheng, Y.-H.; Chou, T.-C.; He, L.; Horwitz, S. B. Angew. Chem., Int. Ed. 1997, 36,
757; (b) Meng, D. F.; Su, D. S.; Balog, A.; Bertinato, P.; Sorensen, E. J.;
Danishefsky, S. J.; Zheng, Y. H.; Chou, T. C.; He, L. F.; Horwitz, S. B. J. Am. Chem.
Soc. 1997, 119, 2733; (c) Su, D.-S.; Balog, A.; Meng, D. F.; Bertinato, P.;
Danishefsky, S. J.; Zheng, Y.-H.; Chou, T.-C.; He, L. F.; Horwitz, S. B. Angew.
Chem., Int. Ed. 1997, 36, 2093.
8. Glunz, P. W.; He, L.; Horwitz, S. B.; Chakravarty, S.; Ojima, I.; Chou, T. C.;
Danishefsky, S. J. Tetrahedron Lett. 1999, 40, 6895.