Design and Synthesis of a Fluoroindolocarbazole Series as Selective Topoisomerase I Active Agents. Discovery of Water-Soluble 3,9-Difluoro-12,13-dihydro-13-[6-amino-β-D-glucopyranosyll]-5H,13H-benzo[b] -thienyl[2,3-a]pyrrolo[3,4-c]carbazole-5,7(6H)-dione (BMS-251873) with Curative Antitumor Activity against Prostate Carcinoma Xenograft Tumor Model

Article abstract of DOI:10.1021/jm034197s

A series of fluoroindolocarbazoles were studied with respect to their topoisomerase I activity, cytotoxicity, selectivity, and in vivo antitumor activity. Emerging from this series was BMS-251873, a potential clinical candidate possessing a robust pharmacological profile including curative antitumor activity against prostate carcinoma.

Full text of DOI:10.1021/jm034197s

J . Med. Chem. 2004, 47, 1609-1612
1609
Design a n d Syn th esis of a
F lu or oin d oloca r ba zole Ser ies a s
Selective Top oisom er a se I Active Agen ts.
Discover y of Wa ter -Solu ble
3,9-Diflu or o-12,13-d ih yd r o-13-[6-a m in o-â-
D-glu cop yr a n osyl]-5H,13H-ben zo[b]-
th ien yl[2,3-a ]p yr r olo[3,4-c]ca r ba zole-
5,7(6H)-d ion e (BMS-251873) w ith Cu r a tive
An titu m or Activity a ga in st P r osta te
Ca r cin om a Xen ogr a ft Tu m or Mod el§
F igu r e 1. Camptothecin (1) and rebeccamycin (2).
express low levels of the enzyme are more resistant to
these agents. Topo I levels are generally higher in
cancer cells than in normal cells, a feature that is most
likely a reflection of higher proliferation rates in trans-
formed cells. Cells that do not express detectable levels
of topo I are completely resistant to camptothecin and
its analogues.
To date, campotothecins2 are the only structural class
of antitumor agents that confer the highest degree of
topo I selectivity. Our efforts to discover non-camptoth-
ecin topo I selective agents stems from an earlier
natural products fermentation program that led to the
identification of the novel indolocarbazole1,3 class of
compounds represented by rebeccamycin (2). These are
DNA binding agents that do not share any specific topo
I targeting activity. Surprisingly, precursor feeding
experiments with fluorotryptophan resulted in the
identification of the prototypical fluoroindolocarbazole4
class of compounds (e.g., 19 and 20) endowed with
selective topo I activity. However, within this fluoroin-
dolocarbazole series, depending on the substitution of
the core nucleus, varying degrees of topo I selectivity
were observed.
Balu N. Balasubramanian,*,# Denis R. St. Laurent,*,#
Mark G. Saulnier,# Byron H. Long,† Carol Bachand,‡
Francis Beaulieu,‡ Wendy Clarke,#
Milind Deshpande,# J effrey Eummer,#
Craig R. Fairchild,† David B. Frennesson,#
Robert Kramer,† Frank Y. Lee,† Mikael Mahler,‡
Alain Martel,‡ B. Narasimhulu Naidu,#
William C. Rose,† J ohn Russell,# Edward Ruediger,‡
Carola Solomon,‡ Karen M. Stoffan,# Henry Wong,#
Kurt Zimmermann,# and Dolatrai M. Vyas#
The Bristol-Myers Squibb Pharmaceutical Research
Institute, 5 Research Parkway, Wallingford Connecticut
06492, P.O. Box 4000, Princeton, New J ersey 08543, and
Candiac, Montreal, Canada
Received September 18, 2003
Abstr a ct: A series of fluoroindolocarbazoles were studied with
respect to their topoisomerase I activity, cytotoxicity, selectiv-
ity, and in vivo antitumor activity. Emerging from this series
was BMS-251873, a potential clinical candidate possessing a
robust pharmacological profile including curative antitumor
activity against prostate carcinoma.
In this communication, we describe the syntheses, in
vitro, and in vivo antitumor activity of prototypical
fluoroindolocarbazole analogues 19, 20 (BMS-210287),
and their core modified and aminosugar analogues.
Several of these possess improved topo I dependent
selectivity, aqueous solubility, and in vivo activity
against human xenograft antitumor animal models. One
such analogue, BMS-251873 (36), was identified as a
potential clinical candidate.
In tr od u ction . Topoisomerase I1 (topo I) normally
functions to form transient single-strand breaks in
genomic DNA to relieve the torsional strain that devel-
ops during DNA replication or transcription. These
transient breaks are formed by a transesterification
process involving the hydroxyl group of an active site
tyrosine and the 5′-hydroxyl group of deoxyribose in-
volved in the phosphodiester backbone of the DNA
molecule.
Resu lts an d Discu ssion . The indolocarbazole nucleus
and its corresponding thio and oxo analogues were
efficiently assembled via the chemistry described in
Schemes 1 and 2. Stepwise addition of an appropriately
substituted (e.g., 19-26, 31-33, Table 1) indole-based
Grignard reagent to a dihalomaleimide followed by
oxidative cyclization furnished the prerequisite fluor-
oindolocarbazole cores3a,5 in modest overall yield. In
contrast, limited access to the benzofurans and ben-
zothiophenes with the desired substitution pattern
warranted the Fischer indole process6 (Scheme 2) as an
attractive alternative (e.g., 27-30, 34, Table 1). Ini-
tially, for the core glycosylation, we explored the epoxide
opening protocol described by Danishefsky.7 The gen-
eration of dissacharide in this reaction warranted an
alternative approach. Thus, the majority of the ex-
amples described herein were synthesized by alkylation8
or Mitsunobu9 protocol (Scheme 3).
Camptothecin2 (CPT, 1, Figure 1) was identified in
1985 as the first member of a novel class of cytotoxic
compounds that acted upon eukaryote topo I to ef-
fectively stabilize these transient intermediates. The
presence of these persisting DNA breaks triggers the
apoptotic cascade in cancer cells expressing wild-type
p53. The cytotoxic effects of topo I active agents are
directly proportional to the number of DNA breaks they
form. Therefore, cells that express high levels of topo I
are sensitive to these agents and, conversely, cells that
§ Dedicated to the memory of Dr. Monroe Wall, who made significant
contributions to natural product based cytotoxic agents.
* To whom correspondence should be addressed. For B.N.B.: phone,
203-677-6693; fax, 203-677-7884; e-mail, balu@bms.com. For D.R.S.L.
phone, 203-677-6972; fax, 203-677-7702; e-mail, stlaured@bms.com.
#
The Bristol-Myers Squibb Pharmaceutical Research Institute, CT.
† The Bristol-Myers Squibb Pharmaceutical Research Institute, NJ .
‡ The Bristol-Myers Squibb Pharmaceutical Research Institute,
Canada.
10.1021/jm034197s CCC: $27.50 © 2004 American Chemical Society
Published on Web 02/20/2004

1610 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 7
Letters
Ta ble 1. Structure-in Vitro Activity Relationships for Some
Indolocarbazole Analogues against Human Topoisomerase I
and Murine P388 Leukemia Cells
Sch em e 1. Grignard Construction of Indolocarbazole
Corea
compd
substitution
X
R5
topoa
P388b
R/Sc
2
rebeccamycin
2,10-diF
3,9-diF
2,3,9,10-tetraF
2,3,9,10-tetraF
3-F
2-F
10-F
9-F
3-F
NH
NH
NH
NH
NMe
NH
NH
NH
NH
S
S
O
O
N
H
>500
2.0
0.22
0.69
>600
6.6
3.1
>200
1.7
0.54
0.26
1.26
8.7
182.7
67.1
>1.0
>9.5
4.6
11.4
31.7
>51.3
232.5
13.6
63.7
196.9
19.6
>10.8
14.1
28
19
20
21
22
23
24
25
26d
27
28
29
30
31
32
33
34
35e
36e
H
H
H
H
H
H
H
H
H
H
H
H
0.018
0.007
>8.720
1.036
0.392
0.098
0.101
0.155
0.010
0.529
0.114
0.020
0.035
0.862
0.236
0.326
0.068
a
Reagents and conditions: (a) 3 M EtMgBr in THF, PhH, 80
°C; (b) DDQ, TsOH, PhH, 80 °C; (c) hν, cat. I2, air, PhH, 80 °C.
Sch em e 2. Fischer Indole Construction of
Indolocarbazole Corea
2.2
0.09
1.5
3,9-diF
3-F
3,9-diF
3,9-diF
3,9-diF
3,9-diF
3,9-diF
3,9-diF
3,9-diF
0.27
NH2 0.22
N
N
S
NH
S
OH
CH3
CH3
H
0.08
1.0
0.48
0.28
0.46
H
>115
a
Ratio of the median effective concentration (EC50, µM) of
a
Reagents and conditions: (a) NaOEt, EtOH, ClCH2CH(OEt)2,
compound for inducing single-strand breaks in the DNA substrate
divided by that obtained for CPT in the same experiment. CPT
mean topo I EC50 ) 160 nM. Mean cytotoxic concentration (IC50
room temp; (b) PPA, PhCl, 100 °C; (c) nBuLi, CH3CHO, THF/Et2O
(1:1), -78 to 0 °C; (d) PCC, CH2Cl2, Celite, room temp; (e) EtOH,
NaOAc, 75 °C; (f) 3 M EtMgBr in THF, THF; (g) hν, EtOH,
dioxane, 100 °C.
b
,
µM) following 3 days of continuous exposure of compound to P388
murine leukemia cells. CPT mean P388 IC50 ) 36 nM. c Ratio
resulting from the cytotoxicity IC50 value obtained for CPT-
resistant P388/CPT45 cells divided by that obtained for parental
Topo I activity (i.e., DNA strand breaks) of the
analogues depicted in Table 1 was measured using a
modified Hsiang10a protocol, and cytotoxicity was de-
termined by tetrazolium dye conversion assay.10b The
topo I selectivity for these analogues was determined
(as R/S ratio of their IC50’s) by measuring the cytotoxcity
against cell lines that are sensitive (S) and resistant (R)
to camptothecin.
P388 cells. Inseparable mixture with 23. e Both 35 and 36 are
d
6′-NH2 analogues.
Sch em e 3. Glycosylation of Indolocarbazole Corea
The resistant cell line expresses little or no functional
topo I. Within the fluoroindolocarbazole series, the 3,9-
difluoro substitution rendered the best topo I potency
and selectivity over the 2,10-difluoro, 2,3,9,10-tetraflu-
oro, 2,3,9-trifluoro, and 10-monofluoro 24, 23, 26, and
25 analogues, respectively. While N6-methylation of the
imide nitrogen (e.g., 33) resulted in reduced topo I
activity, the N6-amino and hydroxy substitution main-
tained the high topo I activity, implying the significance
of H-bonding in this structural motif with the topo-
isomerase I enzyme/DNA ternary complex. In contrast,
the conformational distortion11 imposed by the N12
-
indole methyl group appears to play a predominant role
over N6-imide H-bonding, which is reflected in the loss
of activity observed in example 22. The thio and oxo
analogues (28 and 30) were chosen to fully explore the
influence of electronic, steric, and H-bonding against the
bioactive conformation necessary for topo I activity.
The topo I active, nonselective lead 19, when evalu-
ated against a subcutaneously (sc) implanted A2780
ovarian carcinoma xenograft model, showed some hints
of distal site in vivo activity when administered ip (LCK
) 0.9 at 30 mg/kg dose).12 Interestingly, in the same
tumor model, the more selective lead analogue 20
displayed poor distal site activity whether administered
a
Reagents and conditions: (a) Method A, (Z ) OH) DIAD, Ph3P,
THF, -78 °C to room temp; Method B, (Z ) Cl) NaN(TMS)2, THF
or KOH, NaSO4, THF, room temp to 65 °C; (b) NaN(TMS)2, PhH,
80 °C; (c) 4-8 N HCl/MeOH; (d) 20% Pd(OH)2/C, 95% EtOH,
cyclohexene, reflux or 20% Pd(OH)2/C, 60 psi of H2, EtOAc/EtOH,
room temp; (e) 5 N KOH or 10% NaOH, EtOH, room temp followed
by concentrated HCl, 0 °C; (f) NH4OAc, AcOH, EtOH, 140 °C or
(TMS)2NH-MEOH, DMF, room temp or NH2OH or NH2NH2; (g)
MsCl, 4 Å molecular sieves, pyridine; (h) NaN3, DMF; (i) Ph3P,
wet THF; (j) HCl/MeOH, TBG ) 3,4,6-tribenzyl-L-glucopyranose.
intermittently over 9 days or via 24 h infusion at its
MTD (LCK ) 0.2-0.6). On closer examination of the
pharmacokinetic parameters, it became apparent that

Letters
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 7 1611
Ta ble 3. Median Cytotoxic Potencies of 36 against Cell Lines
after 72 h of Exposure and Antitumor Activities against Their
Corresponding in Vivo Tumor Models
HCT116
A2780
HCT29
M5076
PC3
a
IC50
0.010
10
1.3
0.005
60
1.7
0.065
70
1.6
- -
50
1.2
0.049
60
>2.8
dose (MTD)b
c
LCK
b
aIC50 (µM). Maximum tolerated dose expressed in mg/kg and
administered twice a day iv every other day for 5 days (2q2d × 5)
in cremophor/ethanol/water (10:10:80) except for PC3 (2q3d × 5).
c Log of cell kill.
F igu r e 2. Comparison of 20 in plasma and tumor after iv
and ip administration to mice at 24 mg/kg.
Ta ble 2. Pharmacokinetic Profile of a Select Set of Analogues
a
b
c
d
e
compd
T1/2
Cl
Vds
IC50
AUC
0.17
0.69
0.068 (6 h)
0.12 0.051 (6 h) >5.0
20 (24 mg/kg) 1.6
35 (24 mg/kg) 4.0
28 (5 mg/kg)
3.6
6.1
0.10 0.86
0.46 0.17
7.6
9.33 1.27 0.15
36 (40 mg/kg) 0.86 4.3
a
b
Half-life (h). Clearance (mL min-1 kg-1). c Volume of distri-
bution (L/kg). iv IC50 (µM) after 8 h. e Exposure (µM h mL-1).
d
F igu r e 4. Efficacy of BMS-251873 against CPT-11 at different
doses in the PC-3 xenograft model with 2q3d × 5 iv dosing
schedule (Rx).
36, however, is about 7-fold less cytotoxic for P388
murine leukemia cells than 28. The in vitro cytotoxic
potencies of 36 against a number of human cell lines
are compared in Table 3.
Prostate, colon, and ovarian cell lines are highly
sensitive; however, cell lines expressing P-glycoprotein
appear to be less sensitive (results not shown). Table 2
illustrates a robust pharmacokinetic profile shown by
36 compared to the prototypical leads. The high levels
of exposure in plasma and tumor were sufficiently
significant to warrant further exploration of this ana-
logue in a number of in vivo antitumor animal models.
Also, analogue 36 did not show any major metabolic and
CYP-450 inhibition liabilities. In a panel of in vivo
tumor models, this analogue displayed broad spectrum
activity against human colon (HCT116 and HCT29),
ovarian (A2780), prostate (PC3) carcinomas, and a
murine sarcoma (M5076) (see Table 3).12 In particular,
pronounced antitumor efficacy was observed for 36
against the slow-growing human PC3 xenograft model
that included cures and superior efficacy over CPT-11
(irinotecan)2b (see Figure 4).
F igu r e 3. Comparison of topo I mediated DNA cleavage
activities between 36 and camptothecin.
the poor plasma and tumor exposure, i.e., tumor levels
below cytotoxic IC50 levels of 20, was likely responsible
for the lack of distal site activity as reflected in Figure
2. Furthermore, these analogues displayed very low
solubility in many of the usual vehicles (such as cre-
mophor/EtOH, PEG) used in the in vivo experiments.
This observation prompted us to replace the 6′-hydroxyl
group on the sugar moiety with an amino functionality
(to render enhanced aqueous solubility), employing a
three-step synthetic sequence (OH f OMs f N3 f NH2)
Consequently, conversion of prototypical lead 20 to its
amino analogue 35 did show modest distal site in vivo
activity, albeit being much less topo I selective. The
increased solubility of 35 seemed to act favorably in the
exposure studies, as seen by sustained cytotoxic levels
for more than 6 h in mice when administered ip (Table
2). The amino counterpart 36 of the benzthiophene core
analogue 28 was also evaluated extensively. Analogue
28 appeared to have maintain the topo I selectivity ratio
(see Table 1); however, the plasma levels of 28 were
relatively low (Table 2). Remarkably, the corresponding
6′-aminobenzothiophene analogue 36 (BMS-251873) was
endowed with higher topo I mediated DNA cleavage
activity, cytotoxic potency, and topo I selectivity. Ana-
logue 36 is 2-fold more potent than CPT in vitro, and
the mechanism involving topo I is indistinguishable
from that of CPT as shown above in Figure 3. Analogue
The robust in vitro topo I mediated DNA cleavage
activity, cytotoxicity, and topo I selective cytotoxic
profile of 36 with improved solubility and pharmacoki-
netic behavior encouraged in vivo efficacy evaluations.
The above properties and the broad spectrum of in vivo
antitumor activities against a panel of human and
murine models, especially efficacy in the prostate tumor
xenograft model, mandated the selection of this ana-
logue as a potential clinical candidate.
Su p p or tin g In for m a tion Ava ila ble: Experimental de-
tails and analytical data are available for the preparation of
all compounds. This material is available free of charge via
the Internet at http://pubs.acs.org.

1612 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 7
Letters
(7) Gallant, M.; Link, J . T.; Danishefsky, S. J . A stereoselective
synthesis of indolo-â-N-glycosides: an application to the syn-
thesis of rebeccamycin. J . Org. Chem. 1993, 58, 343-349.
(8) Ohkubo, M.; Kawamoto, H.; Ohno, T.; Nakano, M.; Morishima,
H. Synthesis of NB-506, a new anticancer agent. Tetrahedron
1997, 53, 585-592.
(9) Ohkubo, M.; Nishimura, T.; J ona, H.; Honma, T.; Ito, S.;
Morishima, H. Synthesis of Dissymmetric Indolocarbazole Gly-
cosides Using the Mitsunobu Reaction at the Glycosylation Step.
Tetrahedron 1997, 53, 5937-50.
(10) (a) The procedure for assaying compound-induced, topoisomerase
I mediated single-strand breaks in DNA was essentially that
described by Hsiang et al. in J . Biol. Chem. 1985, 260, 14873-
14878. (b) For in vitro cell-based cytotoxicity activity, the
proliferation inhibition activity against murine P388 cell line
was measured as follows. Evaluation of a soluble tetrazolium/
formazan assay for cell growth and drug sensitivity in culture
using human and other tumor cell lines was done according to
the procedure described in Clin. Cancer Res. 2001, 7, 1429-
1437. The results are expressed as an IC50, which is the drug
concentration required to inhibit cell proliferation (i.e., absor-
bance at 450 nm) to 50% of that of untreated control cells.
(11) Facompre´, M.; Carrasco, C.; Vezin, H.; Chisholm, J . D.; Yoburn,
J . C.; Van Vranken, D. L.; Bailly, C. Indolocarbazole Glycosides
in Inactive Conformations. ChemBioChem 2003, 4, 386-395 and
references therein (e.g., refs 1 and 9).
Refer en ces
(1) (a) Wang, H.-K.; Morris-Natschke, S. L.; Lee, K.-H. Recent
Advances in the Discovery and Development of Topoisomerase
Inhibitors as Antitumor Agents. Med. Res. Rev. 1997, 17, 367-
425 and references therein. (b) Bailly, C. Topoisomerase I poisons
and suppressors as anticancer drugs. Cur. Med. Chem. 2000, 7,
39-58 and references therein.
(2) (a) Wall, M. E Camptothecin and taxol: discovery to clinic. Med.
Res. Rev. 1998, 18, 299-314 and references therein. (b) Bissery,
M. C.; Vrignaud, P.; Lavelle, F.; Chabot, G. G. Experimental
antitumor activity and pharmacokinetics of the camptothecin
analog irinotecan (CPT-11) in mice. Anti-Cancer Drugs 1996, 7,
437-460 and references therein.
(3) (a) Pindur, U.; Kim, Y.-S.; Mehrabani, F. Advances in indolo-
[2,3-a]carbazole chemistry: design and synthesis of protein
kinase C and topoisomerase I inhibitors. Curr. Med. Chem. 1999,
6, 29-69. (b) Long, B. H.; Balasubramanian, B. N. Non-
camptothecin topoisomerase I active compounds as potential
anticancer agents. Expert Opin. Ther. Pat. 2000, 10, 635-666.
(4) Long, B. H.; Rose, W. C.; Vyas, D. M.; Matson, J . A.; Forenza,
S. Discovery of antitumor indolocarbazoles: rebeccamycin, NSC
655649, and fluoroindolo-carbazoles. Curr. Med. Chem.: Anti-
Cancer Agents 2002, 2, 255-266.
(5) Ohkubo, M.; Nishimura, T.; J ona, H.; Honma, T.; Morishima,
H. Practical synthesis of indolopyrrolo-carbazoles. Tetrahedron
1996, 52, 8099-8112.
(6) (a) Bisagni, M.; Buu-Hoi, N. G.; Royer, R. Oxygen Heterocycles.
Part III. The Reactivity of Benzofuran and 2-Alkylbenzofurans.
J . Chem. Soc. 1955, 3688-3693. (b) Adaptation using 2-(2-
indolyl)benzofurans and benzothiophenes in conjunction with ref
5.
(12) LCK (log of cell kill) is defined as the in vivo antitumor activity
obtained against solid tumor implanted sc following iv, ip, or po
administration of the agent and is calculated according to the
guidelines described in Anti-Cancer Drugs 1996, 7, 437-460.
J M034197S