Outer-ring stereochemical modulation of cytotoxicity in cephalostatins.

Article abstract of DOI:10.1021/ol991153y

[structure: see text] 20- and 25'-epimers of cephalostatin 7, prepared by directed unsymmetrical pyrazine synthesis, address outer-ring topographical and stability questions and intimate an oxacarbenium ion rationale for the role in bioactivity of the spi

Full text of DOI:10.1021/ol991153y

ORGANIC
LETTERS
2000
Vol. 2, No. 1
33-36
Outer-Ring Stereochemical Modulation
of Cytotoxicity in Cephalostatins1
Thomas G. LaCour,*,†,‡ Chuangxing Guo,†,§ Michael R. Boyd,| and
,†
Philip L. Fuchs*
Department of Chemistry, Purdue UniVersity, West Lafayette, Indiana 47907, and
National Cancer Institute, Frederick Cancer Research Center,
Frederick, Maryland 21702
pfuchs@purdue.edu
Received October 16, 1999
ABSTRACT
20- and 25′-epimers of cephalostatin 7, prepared by directed unsymmetrical pyrazine synthesis, address outer-ring topographical and stability
questions and intimate an oxacarbenium ion rationale for the role in bioactivity of the spiroketal (E/F, E′/F′) rings of this class of antitumor
agents.
The cephalostatins1,2 and ritterazines3 comprise a family of
45 structurally unprecedented marine products with extreme
cytotoxicity against human tumors. Cephalostatin 1 (1) and
ritterazine B display ∼1 nM GI50s in two-day tests in the
NCI 60-line screen and 10-14 M GI50s in six-day tests in the
Purdue 6-line minipanel.4 We recently reported total syn-
theses of 15a via directed unsymmetrical pyrazine formation5b
and cephalostatin 7 (2) in biomimetic fashion.5c Nearly 50
analogues have been disclosed along with partial SAR
rationales.3-6,7 The mechanism of action of the bissteroidal
pyrazines, apparently shared by monosteroidal glycosides
such as OSW-1,8 remains unknown. Beside a steroidal
platform, critical features implicated in the pharmacophore
include a set of covalently linked polar and nonpolar domains
and the spiroketals (or C22 equivalent).2-8
Criteria for the latter seem unclear beyond their evident
necessity for high activity. Pending identification of a
mechanism or binding site, empirical studies on the type,
substituent effects, and reactivity of steroidal spiroketals may
facilitate emergence of an intelligible SAR pattern. A role
as a network of hydrogen bond donors/acceptors has been
envisioned.9 E-Ring oxacarbenium ions (E-ox.) have been
proposed as active intermediates relating the OSW and
cephalostatin classes,10 and F-ring ions (F-ox.) merit con-
sideration.7 We therefore sought to illuminate relevant
parameters for the outer rings of cephalostatin cytotoxins.
† Purdue University.
‡ lacour@purdue.edu.
§ Present address: Agouron Chemicals, La Jolla, CA 92121.
| National Cancer Institute.
(6) (a) Jautelat, R.; Mu¨ller-Fahrnow, A.; Winterfeldt, E. Chem. Eur. J.
1999, 5, 1226. (b) Dro¨gemu¨ller, M.; Flessner, T.; Jautelat, R.; Scholz, U.;
Winterfeldt, E. Eur. J. Org. Chem. 1998, 2811 and references therein. (c)
Heathcock, C. H.; Smith, S. C. J. Org. Chem. 1994, 59, 6828. (d) For an
early review, see: Ganesan, A. Angew. Chem., Int. Ed. Engl. 1996, 35 and
references therein.
(1) Cephalostatin Support Studies. 19. For parts 17 and 18: ref 15.
(2) Pettit, G. R.; Tan, R.; Xu, J.-P.; Ichihara, Y.; Williams, M. D.; Boyd,
M. R. J. Nat. Prod. 1998, 61, 953 and references therein.
(3) Fukuzawa, S.; Matsunaga, S.; Fusetani, N. J. Org. Chem. 1997, 62,
4484 and references therein.
(4) LaCour, T. G.; Guo, C.; Ma, S.; Jeong, J. U.; Boyd, M. R.; Matsunaga,
S.; Fusetani, N.; Fuchs, P. L. Bioorg. Med. Chem. Lett. 1999, 9, 2587 and
references therein.
(5) (a) LaCour, T. G.; Guo, C.; Bhandaru, S.; Boyd, M. R.; Fuchs, P. L.
J. Am. Chem. Soc. 1998, 120, 692. (b) Guo, C.; Bhandaru, S.; Fuchs, P. L.;
Boyd, M. R. J. Am. Chem. Soc. 1996, 118, 10672. (c) Jeong, J. U.; Sutton,
S.; Kim, S.; Fuchs, P. L. J. Am. Chem. Soc. 1995, 117, 10157.
(7) Some alternative oxacarbenium ions are discussed in Guo, C.; LaCour,
T. G.; Boyd, M. R.; Fuchs, P. L. Bioorg. Med. Chem. Lett. 1999, 9, 419.
(8) Mimaki, Y.; Kuroda, M.; Kameyama, A.; Sashida, Y.; Hirano, T.;
Oka, K.; Maekawa, R.; Wada, T.; Sugita, K.; Beutler, J. A. Bioorg. Med.
Chem. Lett. 1997, 7, 633.
(9) Pan, Y.; Merriman, R. L.; Tanzer, L. R.; Fuchs, P. L. Bioorg. Med.
Chem. Lett. 1992, 2, 967.
10.1021/ol991153y CCC: $19.00 © 2000 American Chemical Society
Published on Web 12/15/1999

5a-8a and azide substitution gave 2R-azido-3-ketones (6
and 8 were utilized as such). Subsequent methoxime forma-
tion and azide reduction afforded epimeric 2R-amino-3-
methoximes 5 and 7. Analogues 20-epi-cephalostatin 7 (3)
and 25′-epi-cephalostatin 7 (4) ascended from these partners
via our protocol for unsymmetrical pyrazine synthesis
(catalytic Sn2+, benzene, 80 °C) followed by routine depro-
tection.
The 20-epi-North 1 series 5a-5 proved as stable to all
manipulations as North 1 7a-7, but the 25′-epi-South 7 series
6a-6 betrayed a promiscuous bent. Bromination of 8a
effected concomitant 25′-OMTM deprotection to give 8br
with complete spiroketal fidelity, and azide substitution
proceeded in near quantitative yield.11b By contrast, wanton
6a surrendered to C20′/22′ isomerization and bred further
5,6- and 5,5-spiroketal progeny, with several likely candi-
dates lying close in energy to 6br. Similar material losses
accompanied the azidation and coupling steps.
Limited information regarding stereochemical effects on
activity appear among C22 epimeric ritterazines.3,4 To
examine the impact of E- and F-ring stereochemistry on
activity, we designed analogues 3 and 4 with modified units
20-epi-North 1 (5) and 25′-epi-South 7 (6), derived from
isomers 5a and 6a obtained during syntheses of the North 1
(7)11a and South 7 (8)11b units, respectively.
Coupling partners 5-8 (Scheme 1) were approached from
the 3-ketosteroids in the usual manner.5,11 Bromination of
Scheme 1. Synthesis of Analogues 3 and 4
Analogues 3 and 4 were comparatively evaluated in the
National Cancer Institute’s 60-line panel of human tumors
alongside parent 2 and benchmark 1. Cytotoxicity data for
2-4 is summarized in Tables 1 and 2.12 Surprisingly,
inversion at either C20 (as in the North 1 unit E-ring of 3)
or at C25′ (as in the South 7 unit F′-ring 4) similarly
diminished the activity. In addition to a 5-100-fold increase
(10) Guo, C.; Fuchs, P. L. Tetrahedron Lett. 1998, 39, 1099.
(11) (a) Kim, S.; Sutton, S.; Guo, C.; LaCour, T. G.; Fuchs, P. L. J. Am.
Chem. Soc. 1999, 121, 2056. (b) Jeong, J. U.; Guo, C.; Fuchs, P. L. J. Am.
Chem. Soc. 1999, 121, 2070. 6a was incorrectly identified as 20S,22S
(compd 90) and a correction has been submitted. See also the Supporting
Information.
(12) (a) Boyd, M. R. In Anticancer Drug DeVelopment Guide; Teicher,
B., Ed.; Humana Press: Totowas, NJ, 1997; p 23. (b) GI50s were normalized
to those of the benchmark cephalostatin 1 (1, average 1.1 nM).
34
Org. Lett., Vol. 2, No. 1, 2000

Figure 1. Overlays of 4 and 3 with 2. Top: edge view of the pyrazine core and F-rings; 4 in bold, 3/2 in shadow. Bottom: canted top view
of the pyrazine core providing a view of the F′-rings as chairs, an edge view of the D/E bicyclic moiety, and bottom view of the F-rings;
3 in bold, 2/4 in shadow.
in many GI50s relative to 2, the number and kinds of tumor
lines affected by 3 and 4 was considerably reduced, and in
strikingly similar fashion. The possibility that mere coinci-
dence explained such consequences in rather different
variants of 2 seemed remote.
Polarity matching or functionality alteration rationales do
not apply here, while topographical changes in 3 and 4
relative to 2 seem unlikely to bear responsibility for their
similar loss of activity. No conformational difference between
2 and 4 (essentially superimposable) was evident in molec-
ular models13 (Figure 1). In 3, minor changes arose in the
E/F-ring conformations and disposition of the F-ring sub-
stituents relative to 2 and 4 to accommodate an outward (not
really endo, which the 13Me prevents) orientation of the
20âMe. Little added hindrance to the F-ring oxygen and
greater access to 17OH was evident in 3, but the opposite
obtained in 4. A simple explanation based on hydrogen
bonding therefore also appears untenable.
One might assume that hampered protonation of the F′-
(F)-ring oxygens in 4(3) relative to 2 would retard formation
of an active E′(E)-ox., accounting for the loss of activity.
However, stereoelectronics suggest that the axial lone pair
of O26′ (F′-ring oxygen) in South 7 units, the one “more
hindered” in 4, should be kinetically less basic than either
the equatorial or O16′ (E′-ring) lone pairs, as it is aligned
for participation into the O16′-C22′ antibonding orbital.14
Moreover, the 25′-epi-South 7 series 6a-6 proved far more
labile than the already acid-sensitive South 75a,11b series 8a-
8. We note that 25(S) precursor 16S cyclized to 5,6- and
5,5-spiroketals 17 and 18 (1:1), while 16R gave only the
5,5-spiroketal 19, in accord with mechanics13 relative ener-
gies (Scheme 2).11b Facile diversion of 6a-6 during bromi-
nation (0 °C, HBr coproduct) to spiroketals analogous to 19,
with further losses during azidation (25 °C, weak acid, polar
solvent) and coupling (Sn2+, molecular sieves, higher tem-
perature), are telling. A kinetic protonation argument thus
seems inadequate.
Table 1. Cytotoxicity (GI50 NM)12 of 2-4 vs Representative
Human Tumors (NCI-60) and Activity Factora
2
3 (20-epi-2)
4 (25′-epi-2)
The sapogenin 5,6-spiroketal is destabilized by 20âMe,
22âO, or 25âMe arrangements, which usually facilitate ring
opening reactions and equilibration to a more stable form.15
Related 20âMe and 22âO consequences apply to 5,5-
systems.5a Calculations predict similar destabilization of
highly oxygenated cephalostatin subunit spiroketals, with
experimental verification available in most cases.3,5,11,16
We have noted that the most active compounds feature at
least one spiroketal in a less stable form.5a Members with a
higher energy spiroketal form often display cytotoxicities
superior to otherwise identical or related members with lower
energy variants. For instance, ritterazine B 9 (22â) is
calculated (mechanics)13 to lie 2 kcal/mol above less active
leukemia: MOLT-4
lung: HOP-92
colon: HCT-116
CNS: SF-295
breast: MCF-7
skin: SK-MEL-2
ovary: IGROV1
renal: 786-0
16.5
56.1
24.5
10.7
4.8
120
468
7.9
195
204
427
251
112
177
234
380
93
98
>1000
>1000
186
>1000
>1000
120
prostate: PC-3
18.6
229
339
overall
a.f. (activity factor)a
49 ((15)
0.97
>420b
>400b
0.80
0.78
a Fraction of cell lines indicating a measurable GI50 e 10-6 M. b Lower
limit only; insufficient lines affected for a true average.
Org. Lett., Vol. 2, No. 1, 2000
35

Scheme 2. Some E-ox. and F-ox. Mediated Spiroketal
Chemistry
Figure 2. Other related epimeric or isomeric cytotoxins.
ritterazine F 10 (22R). The modified units in (similarly) less
active 3 and 4 are calculated to lie (dissimilarly) 2.7 and 1.3
kcal/mol, respectively, aboVe those of the “parent” units of
2, dispelling this simple SAR notion.
correlate with bioactivity, semiempirical calculations were
carried out on the charged species. Initial results suggest that
equipotent 3 and 4 suffer equally diminished E-ox. access
(∆∆Hf +3 kcal/mol) compared to that of 2. This relationship,
if general, should prove valuable for interpreting the SAR
of this class and help guide further work aimed at elucidating
the mechanism of action and/or binding site.
We conclude that epimeric cephalostatins 3 and 4 provide
useful new SAR clues further implicating a role for the
spiroketals. These and a number of other bissteroidal
pyrazines isomerized at outer-ring postions display intriguing
relative cytotoxicities not amenable to explanation by any
simple topography, hydrogen bonding, protonation, or spiroket-
al stability arguments. A rationale based on the currently
remaining viable postulate, namely modulated access to
oxacarbenium ions, is under active exploration to determine
its generality and potential utility.
Access to both kinds of oxacarbenium ions in South 7
type units has been indicated experimentally. Facile isomer-
izations of 17 to 1811b and of natural products ritterazine B
9 to ritterazine C 1417 likely proceed via E′-ring oxacarbe-
nium ions. Interestingly, 14 was much less potent against
P388 than 9 although they could ultimately afford the same
E-ox. Head-to-head testing of ritterazines in the NCI panel
was performed12 as for 2-4 and confirms a corresponding
large drop in potency against human cancers for “analogue”
14 vs “parent” 9 (Figure 2, Table 2). The intermediacy of
an F′-ox. or corresponding protonated 22′-ketone is consistent
with production of the even less cytotoxic translocated
spiroketals 20 and 21 upon longer exposure of 9 to the
influence of acid in polar solvent.3
Acknowledgment. We warmly thank Professors Fusetani
and Matsunaga for samples of ritterazines B and C for
comparative testing and Proctor & Gamble (T.L. fellowship)
and the NIH (CA 60548) for funding.
If spiroketal-furnished oxonium or oxacarbenium ions
contribute to the activity of these antitumor agents, we
speculated that some indication thereof might appear in their
energies. Since the relative stabilities of neutral compounds
which could eventually achieve the same ion failed to
Supporting Information Available: Experimental and
spectral details for 3-8, NMR comparisons, and bioactivity
tables for 1-15. This material is available free of charge
via the Internet at http://pubs.acs.org.
Table 2. Cytotoxicities (nM) of 2-15 vs P388 (Mouse
Leukemia, ED50)1,3 or Human (NCI-60, GI50, and Activity
Factor)12 Tumors
OL991153Y
(13) Calculations used CAChe v. 3.7.
(14) Deslongchamps, P. R. Stereoelectronics in Organic Chemistry;
Oxford: Pergamon Press: 1985; pp 53-60.
(15) (a) LaCour, T. G.; Fuchs, P. L. Tetrahedron Lett. 1999, 40, 4655.
(b) LaCour, T. G.; Tong, Z.; Fuchs, P. L. Org. Lett. 1999, 1, 1815.
(16) (a) Bhandaru, S.; Fuchs, P. L. Tetrahedron Lett. 1995, 36, 8351.
(b) Jeong, J. U.; Fuchs, P. L. Tetrahedron Lett. 1994, 35, 5385.
(17) Fukuzawa, S.; Matsunaga, S.; Fusetani, N.J. Org. Chem. 1995, 60,
608.
2
9
10
11
12
13
14
102
>46b >150b >116b 12
0.90 0.88 0.77 0.98
15
P388
NCI60 49
a.f.a
0.97
<10-4 0.17 0.81 0.81
3.2
0.98
a,b Notes to activity as in Table 1.
36
Org. Lett., Vol. 2, No. 1, 2000