Design, synthesis, and structure-activity relationships of tetrahydroquinoline-based farnesyltransferase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2005.02.004

Tetrahydroquinoline-based small molecule inhibitors of farnesyltransferase (FT) have been identified. Lead compounds were shown to have nanomolar to sub-nanomolar activity in biochemical assays with excellent potency in a Ras-mutated cellular reversion as

Full text of DOI:10.1016/j.bmcl.2005.02.004

Bioorganic & Medicinal Chemistry Letters 15 (2005) 1895–1899
Design, synthesis, and structure–activity relationships of
tetrahydroquinoline-based farnesyltransferase inhibitors
Louis J. Lombardo,* Amy Camuso, John Clark, Krista Fager, Johnni Gullo-Brown,
John T. Hunt, Ivan Inigo, David Kan, Barry Koplowitz, Francis Lee, Kelly McGlinchey,
Ligang Qian, Carolyn Ricca, George Rovnyak, Sarah Traeger, John Tokarski,
David K. Williams, Laurence I. Wu, Yufen Zhao, Veeraswamy Manne and Rajeev S. Bhide
Bristol-Myers Squibb Pharmaceutical Research Institute, PO Box 4000, Princeton, NJ 08543-4000, United States
Received 12 January 2005; revised 31 January 2005; accepted 1 February 2005
Abstract—Tetrahydroquinoline-based small molecule inhibitors of farnesyltransferase (FT) have been identified. Lead compounds
were shown to have nanomolar to sub-nanomolar activity in biochemical assays with excellent potency in a Ras-mutated cellular
reversion assay. BMS-316810 (9e), a 0.7 nM FT inhibitor, was orally-active in a nude mouse tumor allograft efficacy study.
Ó 2005 Elsevier Ltd. All rights reserved.
Ras proteins are a family of GTP-binding proteins in-
volved in cellular signal transduction and cell prolifera-
tion. Oncogenic mutations of the Ras protein cause
disruption of its GTPase activity and prevent interaction
with GTPase activating protein, resulting in constitutive
activation of the Ras signaling cascade.1,2 Oncogenic ras
thereby promotes unregulated cell growth and is esti-
mated to occur in 30–50% of colon, 30% of lung, and
90% of pancreatic tumors in humans.3 The signaling
functions of both normal and oncogenic Ras are contin-
gent upon their association with the cell membrane,
which is accomplished by a series of post-translational
modifications of the newly synthesized protein.4 Farn-
esylation of the C-terminal cysteine residue of p21 ras
protein, which is catalyzed by the zinc metalloenzyme
farnesyltransferase (FT), is critical for its ability to asso-
ciate with the membrane. Thus, disruption of farnesyl-
transferase enzymatic activity is an attractive
pharmacological target for controlling ras dependent
tumor growth.5
leads, tetrahydroquinoline-based inhibitors were pro-
posed as a second-generation series. Herein, we describe
the synthesis, structure–activity relationships, and in
vivo activity of novel tetrahydroquinoline farnesyltrans-
ferase inhibitors in mutant Ras cancer models.
Compounds were prepared following the synthetic se-
quence illustrated in Scheme 1. Boc-protected 3-amino-
1,2,3,4-tetrahydroquinoline 1 was prepared by reaction
of the sodium salt of commercially available 3-amino-
quinoline with di-tert-butyldicarbonate, followed by cat-
alytic hydrogenation over palladium in acidic methanol.
Bromination with pyridinium tribromide proceeded
regioselectively to afford the 6-bromo analog, which
was converted to the corresponding 6-cyano derivative
2 by treatment with zinc cyanide and tetrakis(triphenyl-
phosphine)palladium in dimethylformamide. Removal
of the Boc group with trifluoroacetic acid in dichloro-
methane afforded the key intermediate 3-amino-6-cya-
no-1,2,3,4-tetrahydroquinoline 3 as a racemate in 27%
overall yield. Completion of the target molecules (5)
was accomplished following a three step sequence of sul-
fonamide formation, reductive amination, and alkyl-
ation. In general, reaction of amine 3 with a variety of
sulfonyl chlorides under standard conditions gave the
corresponding sulfonamides (4) in good to excellent
yield. Installation of the imidazole using previously
described reductive amination conditions,7 followed
by alkylation of the sulfonamide nitrogen with an
The activity of benzodiazepine-based farnesyltransferase
inhibitors has been described.6 Motivated by the potent
in vitro activity and excellent in vivo efficacy of these
Keywords: Farnesyltransferase inhibitor; BMS-316810.
*
Corresponding author. Tel.: +1 609 252 4265; fax: +1 609 252
4710; e-mail: louis.lombardo@bms.com
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.02.004

1896
L. J. Lombardo et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1895–1899
Table 1. Effect of alkyl group substitution on the C-3 exocyclic
nitrogen
H
NH2
NC
N
O
a, b
O
R2
N
N
N
H
3
NC
1
S
H
O O
N
O
N
c, d
e
O
N
N
H
NCH3
2
H
R1
NC
NH2
NC
N
f
S
Compd
R2
FTa,b Cellular reversiona
IC50, nM % @ 100 nM
O O
N
H
N
H
3
4
7a
7b
H10
CH3
0% (30% @ 10
3.5
lM)
R2
45%
NC
N
R1
g, h
S
7c
7d
7e
48
0% (90% @ 1 lM)
O O
N
H3C
0.9
4.0
95%
75%
5
N
O
NCH3
O
Scheme 1. Synthesis of tetrahydroquinoline-derived farnesyltransfer-
ase inhibitors. (a) NaHMDS, (Boc)2O, THF, 100%; (b) H2, Pd/C,
HOAC, CH3OH, 46%; (c) PBr3, THF, 75%; (d) ZnCN2, Pd(PPh3)4,
7f
3.0
85%
DMF, 81%; (e) 50% CF3CO2H , CHCl2, 100%; (f) R1SO2Cl, Et3N,
2
7g
>5000
2.5
0%
CH2Cl2, 60–95%; (g) 1-methyl-5-formylimidazole, Et3SiH, CF3CO2H,
CH2Cl2, 75–90%; (h) R2Br, base, DMF, 50–80%.
7h (racemic)
90%
S
S
S
appropriate alkyl halide under basic conditions (NaH-
MDS or Cs2CO3) afforded final compounds in high yield.
7i (S)-
7j (R)-
1.1
2.8
85%
80%
Multigram quantities of enantiomerically-pure 3 could
be obtained using the covalent resolution process out-
lined in Scheme 2. Carbodiimide coupling of racemic 3
with S-(+)-mandelic acid proceeded in quantitative yield
to afford diastereomeric amides (6), which could subse-
quently be separated by flash chromatography on silica
gel or fractional crystallization. Hydrolysis of the diaste-
reomerically-pure amides in refluxing acidic ethanol
afforded the corresponding enantiomerically-pure
amines (3) in high overall yield. Each enantiomer of 3
was converted to the target compounds as described in
Scheme 1. The absolute configuration at the C-3 posi-
tion of the tetrahydroquinoline ring was assigned by sin-
gle crystal X-ray crystallography of one diastereomer of
compound 6, which was determined to be of the (S,S)-
configuration.
a Experimental details for all in vitro assays are contained in Ref. 8.
b IC50 values are the mean of at least two determinations. Standard
deviations are typically within 30% of the IC50 value.
atized tetrahydroquinoline scaffold (Table 1). Consistent
with the defined pharmacophore for benzodiazepine FT
inhibitors, the 6-cyano and 1-(1-methylimidazolylmeth-
5-yl) groups were held constant in the tetrahydroquino-
line series of compounds.6 Generally, the R2 position
was tolerant of a broad range of substituents (com-
pounds 7b; 7d–f; 7h), with lower alkyl-, alkenyl-, carb-
oxymethyl-, and aralkyl-substituted analogues all
demonstrating excellent potency in a human recombi-
nant farnesyltransferase inhibition biochemical assay.
Good enzyme inhibitory activity was also observed with
the unsubstituted derivative 7a, but a reduction in po-
tency was measured with the bulkier cyclohexylmethyl
(7c) and homologated phenethyl (7g) analogues.
Initial structure–activity relationship (SAR) studies fo-
cused on assessing the effect of alkyl group substitution
(R2) on the C-3 nitrogen of a benzenesulfonamide-deriv-
The cellular activity of tetrahydroquinoline FT inhibi-
tors was evaluated in mutant H-ras transfected Rat-1
cells. This transformed cell line is characterized by un-
checked proliferation in culture, with the cells no longer
displaying contact-inhibited growth properties and pos-
sessing a distinctive, rounded and elongated phenotype.
The activity of compounds under evaluation is reported
as the percentage of cells no longer displaying this mu-
tant Ras phenotype at the screening concentration of
100 nM, unless otherwise noted. Activity in the cellular
reversion assay correlated well with biochemical potency
through the series, with two notable exceptions. The
OH
H
3
NC
NH2
NC
N
a, b
O
N
N
H
H
6
3
c
Scheme 2. Covalent resolution of the racemic 3-aminotetrahydroquin-
oline intermediate 3. (a) S-(+)-mandelic acid, EDAC, HOBT, DMF,
100%; (b) flash chromatography (SiO2; 2:1 EtOAc–hexanes) or
fractional crystallization (acetone–hexanes), 66–78% of theoretical;
(c) 15% H2SO4, EtOH, reflux, 72%.

L. J. Lombardo et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1895–1899
1897
unsubstituted analogue 7a was completely inactive at
the 100 nM screening concentration, as was the cyclo-
hexylmethyl compound 7c. It is proposed that poor cel-
lular permeability of 7a was responsible for its low
potency in this assay, but the weak activity of 7c is
not well understood. The reduced activity of the
methyl-substituted derivative 7b may also be due to
poor cellular permeability, presumably because of the
lower lipophilicity of this compound relative to others
in the series.
A series of simple ester analogues related to the tert-
butyl carboxymethyl compound 7e were prepared, but
their in vitro assessment was complicated by hydrolysis
of the ester under assay conditions. The corresponding
carboxylic acid was 3 times less active than compound
7e as a farnesyltransferase inhibitor, but had no activity
in cells at concentrations up to 1 lM. Amide analogues
did not suffer from chemical instability in the assay
media and possessed biochemical potencies similar to
7e, but were considerably less potent in the cellular
Table 2. Effect of sulfonamide R1 substitution and interaction with the R2 alkyl group
R2
3
NC
N
R1
S
O O
N
N
NCH3
Compd
R1
R2
FT IC50, nM
Cellular reversion
% @ 100 nM
H3C
8a
CH3
0.8
90% (IC50 = 8 nM)
8b
9a
2.8
3.1
0% (85% @ 1 lM)
0% (85% @ 1 lM)
CH3
N
H3C
9b
0.9
1.5
2.6
0.7
85% (IC50 = 60 nM)
85% (IC50 = 10 nM)
90% (IC50 < 20 nM)
95% (IC50 = 13 nM)
O
9c
O
9d
9e BMS-316810
S
10a
10b
CH3
17
0% (10% @ 1 lM)
N
H3C
1.2
70% (IC50 = 100 nM)
O
10c
10d
10e
11a
11b
3.9
3.5
0.9
0.3
0.6
30% (IC50 = 225 nM)
75% (IC50 = 70 nM)
85%
O
S
H3C
N
80% (IC50 = 50 nM)
90% (IC50 = 30 nM)
H3CN
O
O
11c
1.1
75%
11d
12a
12b
0.6
7.6
4.6
65%
S
CH3
N
H3CN
0% (25% @ 1 lM)
75%
12c
0.7
95%
S

1898
L. J. Lombardo et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1895–1899
reversion assay and were not pursued further (data not
shown).
To assess the stereochemical preference for FT inhibi-
tion, the enantiomers of compound 7h were evaluated
in the biochemical and cellular reversion assays. The
racemic 3-thienylmethyl analogue 7h was determined
to be virtually equipotent with both the (S)- and (R)-
enantiomers 7i (98.5% ee) and 7j (98% ee), respectively.
Consequently, all additional studies were performed on
racemic mixtures.
Subsequent to characterizing the impact of the R2 sub-
stituent and stereochemistry on the activity of the series,
SAR studies were directed toward the sulfonamide (R1)
and the interaction of this group with the R2 moiety
(Table 2). A variety of R1 substituents could be employed
without degrading the excellent biochemical potency of
the parent benzene sulfonamides. Thus, methyl (8a,b),
pyridyl (9a–e; 10a–e), imidazolyl (11a–d), and pyrazolyl
(12a–c) sulfonamide analogues were found to be low
nanomolar to subnanomolar farnesyltransferase inhibi-
tors, regardless of the R2 substituent appended to the
C-3 nitrogen. The imidazolylsulfonamide 11a was the
most potent analogue of the series, with an IC50 of
0.3 nM.
Figure 1. Model showing BMS-316810 (9e; carbon atoms in green)
bound to the crystal structure of farnesyltransferase. The a- and b-
subunits of farnesyltransferase are depicted in purple and blue,
respectively. Graphics were produced using ICM.11
361b, and the methylthiophene group binds in a hydro-
phobic pocket, stabilized by stacking against the FPP
farnesyl moiety and end-on-face stacking with Trp
102b and Trp 106b. With the exception of the imidaz-
oleÕs coordination with the zinc ion, the compound
forms only van der Waals contacts with the protein
and the FPP farnesyl moiety. The sulfonamide pyridine
ring is mostly solvent exposed, and stabilized primarily
by stacking against the tetrahydroquinoline ring.
Overall, compounds were highly effective in the cellular
assay, with many examples demonstrating complete
reversion of the Ras-transformed phenotype at the
100 nM screening concentration. Compounds could be
further differentiated by titration in this assay to obtain
IC50 values, which are included in the table when avail-
able. As a group, the 2-pyridylsulfonamides (9b–e) had
the best activity profile, with IC50s ranging from 60 to
10 nM, depending upon the R2 substituent. In contrast,
conservative modification to afford the regioisomeric 3-
pyridylsulfonamide analogues (10b–e) gave compounds
that were less effective in the reversion assay. Five-mem-
bered heterocyclic sulfonamides 11a–d and 12b,c were
also generally less potent in cells than the 2-pyridylsulf-
onamide analogues.
In a mouse exposure study, BMS-316810 (9e) was found
=
to produce high systemic drug levels (AUCtot
61 lM h; Cmax = 13 lM) following oral administration
of a single 200 mg/kg dose formulated as a solution in
10% aqueous ethanol. The good oral absorption proper-
ties of BMS-316810 led to the selection of this com-
pound for evaluation in in vivo anti-tumor efficacy
studies.
The in vivo anti-tumor activity of BMS-316810 (9e) was
assessed in nude mice implanted with subcutaneous San-
1 tumor cells, which bear an activating mutation in the
H-ras gene.12 Following implantation, tumors were
staged to approximately 125 mg before initiation of drug
administration. At the high dose of 200 mg/kg twice
daily for 14 days, BMS-316810 suppressed tumor growth
for the duration of treatment with no observed toxicity
(Fig. 2). Lower doses of 50 and 100 mg/kg, administered
on the same schedule, were not effective. Expressed
quantitatively, BMS-316810 delayed tumor growth to
500 mg by 15 days relative to untreated controls. Addi-
tionally, on day 21 following tumor implantation, BMS-
316810 treated animals had tumors 21% the size of
control animals. Taken together, these data confirm
the efficacy and safety of tetrahydroquinoline FT
inhibitors as anti-tumor agents.
Consistent with the benzenesulfonamide 7a, methyl sub-
stitution at R2 led to compounds with poor cellular
activity (9a; 10a; 12a) regardless of the sulfonamide R1
substituent. Less clear is the weak cellular activity of
methanesulfonamide 8b, especially in comparison to
the excellent potency of the propenyl analogue 8a.
A low-energy conformation of BMS-316810 (9e) was
manually docked into the available crystal structure of
farnesyltransferase (PDB ID 1sa5)9 in the FT peptide
binding site, which is a deep hydrophobic cleft formed
at the interface of the a- and b-subunits (Fig. 1). The
˚
inhibitor and residues of FT within 4 A were allowed
to relax during 2000 cycles of conjugate gradient energy
minimization.10 BMS-316810 most likely binds as a
ternary complex with farnesyl diphosphate (FPP) and
the catalytic zinc ion in a conformation consistent with
that reported for benzodiazepine-based inhibitors. FPP
forms part of the inhibitor binding surface. The central
tetrahydroquinoline ring stacks face-on-face with Tyr
In summary, tetrahydroquinoline-based small molecule
inhibitors of farnesyltransferase have been identified.
The series was shown to have excellent potency in bio-
chemical and cellular assays, but no stereochemical pref-
erence for activity in either assay. BMS-316810 (9e), a

L. J. Lombardo et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1895–1899
1899
2. (a) McCormick, F. Cell 1989, 56, 5; (b) McCormick, F.
Curr. Opin. Cell Biol. 1990, 2, 181.
3. Bos, J. L. Cancer Res. 1989, 49, 4682.
1000
4. (a) Casey, P. J.; Solski, P. J.; Der, C. J.; Buss, J. E. Proc.
Natl. Acad. Sci. U.S.A. 1989, 86, 8323; (b) Hancock, J. F.;
Magee, A. I.; Childs, J. E.; Marshall, C. J. Cell 1989, 57,
1167.
5. (a) Gibbs, J. B. Cell 1991, 65, 1; (b) Gibbs, J. B.; Oliff, A.
Annu. Rev. Pharmacol. Toxicol. 1997, 37, 143; (c) Rowin-
sky, E. K.; Windle, J. J.; Von Hoff, D. D. J. Clin. Oncol.
1999, 17, 3631.
6. (a) Ding, C. Z.; Batorsky, R.; Bhide, R.; Chao, H. J.; Cho,
Y.; Chong, S.; Gullo-Brown, J.; Guo, P.; Kim, S. H.;
Patel, M.; Penhallow, B. A.; Ricca, C.; Rose, W. C.;
Schmidt, R.; Slusarchyk, W. A.; Vite, G.; Yan, N.;
Manne, V.; Hunt, J. T. J. Med. Chem. 1999, 42, 5241;
(b) Hunt, J. T.; Ding, C. Z.; Batorsky, R.; Bednarz, M.;
Bhide, R.; Cho, Y.; Chong, S.; Chao, S.; Gullo-Brown, J.;
Guo, P.; Kim, S.-H.; Lee, F. Y. F.; Leftheris, K.; Miller,
A.; Mitt, T.; Patel, M.; Penhallow, B. A.; Ricca, C.; Rose,
W. C.; Schmidt, R.; Slusarchyk, W. A.; Vite, G.; Manne,
V. J. Med. Chem. 2000, 43, 3587.
500
0
control
50 mg/kg, BID x 14d
100 mg/kg, BID x 14d
200 mg/kg, BID x 14d
Rx
10
15
20
25
30
35
40
45
50
days post-tumor implant
Figure 2. In vivo anti-tumor activity of BMS-316810 (9e) versus
subcutaneous San-1 murine tumor implants in nude mice. Drug
formulated as a solution in 10% aqueous ethanol.
0.7 nM biochemical inhibitor of the FT enzyme, was
orally-active versus the San-1 murine tumor cell line in a
nude mouse tumor allograft efficacy study. The interac-
tion of BMS-316810 with the farnesyltransferase enzyme
was modeled through docking experiments with a re-
ported crystal structure, and is predicted to be consistent
with the binding of benzodiazepine-based inhibitors.
7. Chen, B.-C.; Sundeen, J. E.; Guo, P.; Bednarz, M. S.;
Zhao, R. Tetrahedron Lett. 2001, 42, 1245.
8. (a) Manne, V.; Yan, N.; Carboni, J. M., et al. Oncogene
1995, 10, 1763; (b) Carboni, J. M.; Yan, N.; Cox, A. D.,
et al. Oncogene 1995, 10, 1905.
9. Reid, T. S.; Beese, L. S. Biochemistry 2004, 22, 6877.
10. Hagler, A. T.; Ewig, C. S. Comput. Phys. Commun. 1994,
84, 131–155.
References and notes
11. Molsoft LLC, 3366 No. Torrey Pines Ct., Ste. 300,
La Jolla, CA 92037.
12. Manne, V.; Lee, F. Y. F.; Bol, D. K.; Gullo-Brown, J.;
Fairchild, C. R.; Lombardo, L. J.; Smykla, R. A.; Vite, G.
D.; Wen, M.-L.; Yu, C.; Wong, T. W.; Hunt, J. T. Cancer
Res. 2004, 64, 3974.
1. (a) Barbacid, M. Annu. Rev. Biochem. 1987, 56, 779; (b)
Kato, K.; Cox, A. D.; Hisaka, M. M.; Graham, S. M.;
Buss, J. E.; Der, C. J. Proc. Natl. Acad. Sci. U.S.A. 1992,
89, 6403; (c) Rowinsky, E. K.; Windle, J. J.; Von Hoff, D.
D. J. Clin. Oncol. 1999, 17, 3631.