Guiding farnesyltransferase inhibitors from an ECLiPS library to the catalytic zinc

Article abstract of DOI:10.1016/j.bmcl.2005.10.070

Farnesyltransferase inhibitors identified from an ECLiPS library were optimized using solution-phase synthesis. X-ray crystallography of inhibited complexes was used to identify substructures that coordinate to the active site zinc. The X-ray structures were ultimately used to guide the design of second-generation analogs with FTase IC₅₀s of less than 1.0 nM.

Full text of DOI:10.1016/j.bmcl.2005.10.070

Bioorganic & Medicinal Chemistry Letters 16 (2006) 507–511
Guiding farnesyltransferase inhibitors from an ECLiPSÒ
library to the catalytic zinc
Chia-Yu Huang,a Tara M. Stauffer,a Corey L. Strickland,b John C. Reader,a
He Huang,a Ge Li,a Alan B. Cooper,b Ronald J. Doll,b Ashit K. Ganguly,b
John J. Baldwina and Laura L. Rokosza,*
aPharmacopeia Drug Discovery, PO Box 5350, Princeton, NJ 08543-5350, USA
bSchering-Plough Research Institute, 2015 Galloping Hill Rd., Kenilworth, NJ 07033-1300, USA
Received 9 September 2005; revised 18 October 2005; accepted 19 October 2005
Available online 10 November 2005
Abstract—Farnesyltransferase inhibitors identified from an ECLiPSÒ library were optimized using solution-phase synthesis. X-ray
crystallography of inhibited complexes was used to identify substructures that coordinate to the active site zinc. The X-ray structures
were ultimately used to guide the design of second-generation analogs with FTase IC50s of less than 1.0 nM.
Ó 2005 Elsevier Ltd. All rights reserved.
Farnesyltransferase (FTase) is a compelling therapeutic
target in the field of oncology as it is specifically linked
to the aberrant behavior of tumor cells.1 The tricyclic
drug, SCH 66336 (Sarasar, Fig. 1), was the first FTase
inhibitor (FTI) to enter the clinic, resulting in favorable
outcomes in a number of solid tumor types and hemato-
logical malignancies.2 The significant synergy with tax-
anes in inhibiting tumor growth makes this chemical
class of FTIs particularly intriguing.3
Br
Cl
N
H
Br
O
H2N
N
N
O
Figure 1. SCH 66336, Sarasar, FTase IC50 = 1–2 nM.2b
A collection of FTIs that are structurally related to SCH
66336, identified from FT-1, a 11,718-member ECLiPSÒ
(Encoded Combinatorial Library on Polymeric Sup-
port) library, has been described recently.4 The most
potent hits from FT-1 contained a 3-bromo-8-chloro-
benzocyclohepta-pyridine ring attached to a piperazine
core with 3-pyridylmethylamide or alkyl amides at posi-
tion 2 and phenylpropanamide or 4-pyridylacetamide at
position 1. Examples of such compounds are represent-
ed by 1 and 2 (Fig. 2, Scheme 1). The substituents at
position 2 make these compounds unique since this is
a site that had never previously been explored. To
further define the SAR at these positions, several smaller
collections of analogs were made. These synthetic efforts
Br
Br
Cl
Cl
N
N
N
N
H
N
H
N
N
N
N
N
O
O
O
O
1
2
FTase IC50 = 30 ± 10 nM
FTase IC50 = 50 ± 0 nM
Figure 2. Resynthesized compounds from the ECLiPS library, FT-1.
also constituted an attempt to identify substituents that
could make a comparable contribution to the potency as
that afforded by the bromine at position C-10 of tricyclic
FTIs such as SCH 66336 (Fig. 1).5,6
Keywords: Anti-cancer; Catalytic zinc; Farnesyltransferase; Rational
drug design; Solution-phase synthesis; X-ray crystallography.
*
Corresponding author. Tel.: +1 609 452 3718; fax: +1 609 655
4187; e-mail: rock@pcop.com
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.10.070

508
C.-Y. Huang et al. / Bioorg. Med. Chem. Lett. 16 (2006) 507–511
Table 1. Elaboration of the N-1 position of core
pyridylmethyl)
7
(R1 = 3-
BOC
BOC
N
N
R1NH2
1. Piperidine
Compound R2
FTase IC50 COS IC50
(nM) SD (nM) SD
OH
NHR1
N
N
2. R1CO2H/PyBrop
DCC or PyBrop
O
O
Fmoc
Fmoc
8
Me(CH2)2CH2–
5.7 2.1
8.0 1.2
7.3 0.7
12 0.4
12 3.0
64
75
50
6
2
0
3
4
9
Me(CH2)3CH2–
Me2CHCH2–
Me3CCH2–
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
BOC
N
120 23
85
Br
Cl
1. TFA
Cycloheptyl–
Cyclohexylethyl–
Cyclopentyl–
9
N
N
NHR1
20
7
38 12
65 21
N
5
6
3
2. 1,2,2,6,6-pentamethylpiperidine
+
4
1
17 6.0
7.0 0.6
8.0 0.3
8.4 0.9
11 3.0
20 5.3
21 8.5
7.9 0.4
13 1.4
22 3.5
9.1 1.4
O
NHR1
R2
O
2
HO2C(CH2)2CH2–
MeC(O)(CH2)3CH2–
MeC(O)(CH2)2CH2–
MeC(O)CH2CH2–
MeOCH2CH2–
MeCH2OCH2–
MeOC(O)(CH2)3CH2–
Me2NC(O)CH2CH2–
Me3CO–
>200
49 15
N
Cl
Br
5
O
R2
O
54
47
2
2
0
N
7
Cl
6
100
160 12
Scheme 1. General synthetic procedure for synthesis of the FT-1
analogs.
62
11
2
2
>200
45
CH3(CH2)3NH–
CH3(CH2)2NH–
Me2CHCH2NH–
Me2CHNH–
The first collection, synthesized as shown in Scheme 1,
contained 3-pyridylmethylamide at the 2-position of
the piperazine core since this was a frequently seen fea-
ture of the most potent compounds retrieved from the
FT-1 screen.4 Starting from N-1-Fmoc N-4-Boc pipera-
zine-2-carboxylic acid 3, coupling of 3-pyridylmethyl-
amine using DCC or PyBrop gave amide 4. Removal
of the Fmoc group and coupling with various acids
using PyBrop resulted in compound 5. After deprotec-
tion of the Boc group with TFA, the neutralized amine
was reacted with previously reported7 chloride 6 to af-
ford compound 7. The compounds were tested in the
same scintillation proximity assay (SPA) used to screen
FT-1.4 Since the FTase concentration is 2.8 nM, the
lower limit of sensitivity for IC50 determinations is
approximately 5 nM. To further supplement our ability
to rank-order the compounds, all compounds were also
evaluated in a COS cell Ha-Ras processing assay.8
19 1
8.0 1.3
59 17
52 37
47
15
5
2
140
54
1
4
Me3CNH–
Cycloheptyl-NH–
Cyclohexyl-NH–
Cyclohexylmethyl-NH– 10 1.3
9.1 0.5
5.5 0.7
52 26
8.6 0.2
53
8
Cyclopentyl-NH–
Cyclopropyl-NH–
14
51 4.2
8
28 1.5
260 46
as but it also shows the smallest separation between the
enzyme and COS assay IC50s.
A trihalogenated analog of compound 8, containing an
additional bromine at position 10 of the tricycle, demon-
strates comparable enzyme activity (FTase IC50 = 5.6
2.4 nM) but improved COS activity (COS IC50 = 5.3
0.0 nM). X-ray crystallography of FTase inhibited with
this compound showed that the 3-pyridylmethyl side
chain interacts with the active site zinc (Fig. 3). The dis-
tance between the pyridine nitrogen and the zinc atom
Both acyclic (8–11) and cyclic amides (12–14) are well
tolerated (Table 1) with IC50s of less than 20 nM. The
cyclic amides show a smaller separation between enzyme
and cell activity suggesting better whole-cell penetration.
˚
is 2.29 A. This finding makes this series of compounds
particularly noteworthy since the tricyclic FTI, SCH
66336, does not contain substituents that are capable
of zinc coordination.10 The X-ray structure thus sug-
gested that other zinc ligands would be suitable replace-
ments for this substituent. Fixing R2 as butyl, a survey
of R1 modifications such as pyridines (34, 35), pyrazine
(36), isoxazole (37), and ethyl- or methyl-imidazoyl
(38–40) indicated that while all but the pyridine N-ox-
ide, 34, were tolerated, the imidazoylethyl, as exempli-
fied by 38, was most preferred yielding an enzyme
IC50 of 12 3.6 nM and a COS IC50 of 60 15 nM
(Table 2).
Elaboration of the acyclic amides at R2 revealed a toler-
ance for many functional groups on the distal end of this
position, such as carboxylic acid (15), ketones (16–18),
ethers (19, 20), ester (21), amide (22), and carbamate
(23). The tertiary amide, 22, is remarkable in that the en-
zyme IC50 (13 1.4 nM) and the COS IC50 (11 1.5)
are nearly identical. This suggests that the amide chem-
otype has a lower threshold for cell penetration. The
carboxylic acid, 15, and the carbamate, 23, show a much
higher threshold for cell penetration. Nonetheless, the
potency of 23 in the enzyme assay (IC50 = 22 3.5)
prompted us to explore other related functionalities,
such as urea.
Given the potent enzyme and cell-based activity of the
cyclohexylurea, 30 (Table 1), we sought to combine this
pharmacophore with those of the imidazoyls in Table 2.
As shown in Table 3, imidazoylpropyl (41, 42) or
imidazoylethyl (43, 44) containing compounds, along
with the 2-oxo-pyrrolidinylpropyl compound, 46, are
comparable in both enzyme and COS activity to that
of the 3-pyridylmethyl compound, 30. Compound 41
Acyclic (24–28) and cyclic (29–33) ureas were prepared
using a procedure similar to Scheme 1, except that isocy-
anates were used in place of carboxylic acids at R2. The
most potent compound is the cyclohexylurea 30, with an
enzyme IC50 of 5.5 0.7 nM and a COS IC50 of
8.6 0.2 nM. Not only is 30 the most potent of the ure-

C.-Y. Huang et al. / Bioorg. Med. Chem. Lett. 16 (2006) 507–511
509
Figure 3. X-Ray structure of the FTase complex with a 10-Br analog (FTase IC50 = 5.6 2.4 nM, COS IC50 = 5.3 0.0 nM) of 8.9 The zinc atom at
the active site is represented by the red ball.
Table 2. Pyridine replacements at the C-2 position of core
(R2 = butyl)
7
Table 3. Analogs of 38 (Core 7 where R2 = cyclohexyl-NH–)
Compound
R1
FTase IC50
(nM) SD
COS IC50
(nM) SD
Compound
R1
FTase IC50
(nM) SD
COS IC50
(nM) SD
41
42
4.5 2.3
13 1.7
4.2 0.2
9.7 1.4
N
N
8
5.7 2.1
64
6
N
N
N
N
34
35
36
1800 570
>200
H
O
N
43
44
7.0 0.4
7.3 2.9
18 8.7
18
2
100 30
N
H
N
N
N
N
13 17
190 14
560 20
570 180
N
N
O
45
46
26 7.7
14 1.5
91 17
20 0.3
N
N
37
38
39
40
>4500
60 15
>200
O
N
12
80
4
4
NH
N
N
47
61 20
550
H
N
S
N
48
49
490 130
760 250
>10,000
N
H
N
N
45 19
210
7
Not tested
N
N
O
(a mixture of four isomers) is the most potent in this col-
lection with IC50s in both the enzyme and COS assay of
less than 5.0 nM. The potency in the enzyme assay is
predicted to be greater as it appears to be beyond the
limit of sensitivity for the FTase SPA assay.
core11, compound 41 was prepared using (R)-pipera-
zine-2-carboxylic acid to give a diastereomeric mixture
and subsequently separated by HPLC to give 41a and
41b (Table 4). The absolute configuration at C-11 of
the tricyclic piece was not determined but the findings
of Strickland et al.5 suggest that the stereochemistry at
this position does not have a significant impact on
Since it is known that the (R) isomer is the preferred
configuration at the C-2 position of the piperazine

510
C.-Y. Huang et al. / Bioorg. Med. Chem. Lett. 16 (2006) 507–511
Table 4. Biological evaluation of the purified isomers of 41
Compounda FTase IC50 (nM)b GGTase I IC50 (nM)c COS IC50 (nM) SD Soft Agar IC50 (nM) SD Rat liver microsome stabilityd
41a
V1 = 2.0
V2 = 0.10 0.04
V1 = 1.5
22,000
1.2 0.3
5.2 1.0
>95%
41b
>10,000
0.56 0.06
5.0 0.6
>95%
a Compound 41 was prepared as a diastereomeric mixture and was subsequently resolved by HPLC. Compound 41a is the first-eluting isomer and
41b is the second-eluting isomer.
b The V1 FTase IC50 was quantified as described in Rokosz et al.4 The V2 FTase IC50 was quantified as described in Ref. 12.
c GGTase I activity was measured as described by Bishop et al.8 All other assays were conducted as described in the References and notes section.
d The microsome stability is defined as the percent of compound remaining following treatment with microsomes as described.14
Figure 4. X-ray crystal structure of 41b bound to FTase. The zinc atom at the active site is represented by the red ball.
enzyme activity. The less polar isomer, 41a, was arbi-
trarily designated as isomer 1, while the more polar iso-
mer, 41b, was designated as isomer 2. As expected, the
purified isomers show identical activities in both the en-
zyme and COS assays (IC50s 6 2.0 nM). To obtain a
more accurate enzyme IC50 compound 41a was re-tested
in a modified version of the SPA assay, with improved
sensitivity (V2).12 This assay yielded an IC50 for 41a of
0.10 0.04 nM. This potency is consistent with that
reported by Taveras et al.10 A soft agar assay was used
to determine if the compounds can reverse the adherent-
independent phenotype of the transformed cell line,
NIH-3T3.13 Both isomers produced soft agar IC50s of
5.0 nM. They are also greater than 10,000-fold selective
for FTase over geranylgeranyltransferase I (GGTase I).
In order to assess metabolic stability, the isomers were
incubated with rat liver microsomes for 1 h at 37 °C.14
Both compounds remained largely (>95%) intact under
these conditions. A mouse pharmacokinetic study was
used to show that 41b has an AUC (area under the
curve) of 1.66 lg/ml h and a Cmax of 2.55 lM (10 mpk
in 20% HPbCD, po). However, the overall bioavailabil-
ity, when compared to the plasma levels of intravenous
dosing, was just 10.6%. Metabolism is likely due to pep-
tidic cleavage at the amide bond. This reaction can be
tempered via conversion of the secondary nitrogen at
the amide bond to a tertiary nitrogen.10,15
enhancement in potency afforded through interaction
with the active site zinc is comparable to, if not greater
than, that contributed through the addition of a bro-
mine at the C-10 position of the tricycle. Further struc-
tural details of the FTase-41b complex will be reported
in due course.
Acknowledgments
We are grateful to Jagdish Desai for the synthesis of the
tri-halogenated compound shown in Figure 3 and to
Amin Nomeir for pharmacokinetic evaluation of 41b.
We would like to express our sincere appreciation to
Dr. Matthew Sills for critical review of the manuscript.
References and notes
1. (a) Doll, R. J.; Kirschmeier, P.; Bishop, W. R. Curr. Opin.
Drug Discovery Dev. 2004, 7, 478; (b) Dancey, J. E. Curr.
Pharm. Des. 2002, 8, 2259; (c) Kohl, N. E. Ann. N. Y.
Acad. Sci. 1999, 886, 91; (d) Gibbs, J. B.; Oliff, A. Ann.
Rev. Pharmacol. Toxicol. 1997, 37, 143.
2. (a) OÕRegan, R. M.; Khuri, F. R. Endocr. Relat. Cancer
2004, 11, 191; (b) Bishop, W. B.; Kirschmeier, P.; Baum,
C. Cancer Biol. Ther. 2003, 9, 73; (c) Caponigro, F.;
Casale, M.; Bryce, J. Exp. Opin. Invest. Drugs 2003, 12,
946; (d) Head, J. E.; Johnston, S. R. D. Exp. Opin. Emerg.
Drugs 2003, 8, 168; (e) Kelland, L. R. Exp. Opin. Invest.
Drugs 2003, 12, 417; (f) Haluska, G. K.; Dy, G. K.; Adjei,
A. A. Eur. J. Cancer 2002, 38, 1685.
X-ray crystal analysis of FTase inhibited with 41b
(Fig. 4) shows that the imidazole moiety interacts, as
expected, with the catalytic zinc in the active site. The
distance between the imidazole and the zinc atom is
3. (a) Nielsen, L. L.; Shi, B.; Hajian, G.; Yaremko, B.; Lipari,
P.; Farrari, E.; Gurnani, M.; Malkowski, M.; Chen, J.;
˚
2.08 A. Taken together, these data suggest that the

C.-Y. Huang et al. / Bioorg. Med. Chem. Lett. 16 (2006) 507–511
511
Bishop, W. R.; Liu, M. Cancer Res. 1999, 59, 5896; (b)
Wang, E.; Casiano, C. N.; Clement, R. P.; Johnson, W. W.
Cancer Res. 2001, 61, 7525; (c) Hoover, R. R.; Mahon, F.;
Melo, J. V.; Daley, G. Q. Blood 2002, 100, 1068; (d)
Caponigro, F. Anti-Cancer Drugs 2002, 13, 891.
Rossman, R. R. WO9631478, 1996; . Chem. Abstr. 1996,
126, 7997; (b) Cooper, A. B.; Doll, R. J.; Girijavallabhan,
V. M.; Ganguly, A.; Reader, J. C.; Baldwin, J. J.; Huang,
C.-Y. WO9857960, 1998; . Chem. Abstr. 1998, 130, 81426.
12. To enhance the sensitivity of the FTase SPA assay,
the following changes were made: Each reaction
mixture (80 ll total) contained 50 mM Tris, pH 7.5,
5 mM MgCl2, 5 lM ZnCl2, 5 mM DTT, 0.01% (v/v)
Triton X-100 (Sigma), and 176 nM [3H]farnesyl pyro-
phosphate (NEN; 20 Ci/mmol). Purified FTase enzyme
(280 pg, 30 pM final concentration) and inhibitor or
DMSO control (5% v/v, final) were pre-incubated for
15 min at room temperature prior to the addition to
the assay mix. The enzyme reaction was initiated with
20 ll of a solution containing 6.8 ng (100 nM final)
4. Rokosz, L. L.; Huang, C-Y.; Stauffer, T. M.; Reader, J.
C.; Chelsky, D.; Sigal, N. H..; Ganguly, A. K.; Baldwin, J.
J. Bioorg. Med. Chem. Lett. 2005, 15, 5537.
5. Strickland, C. L.; Weber, P. C.; Windsor, W. T.; Wu, Z.;
Le, H. V.; Albanese, M. M.; Alvarez, C. S.; Cesarz, D.; del
Rosario, J.; Deskus, J.; Mallams, A. K.; Njoroge, F. G.;
Piwinski, J. J.; Remiszewski, S.; Rossman, R. R.; Taveras,
A. G.; Vibulbhan, B.; Doll, R. J.; Girijavallabhan, V. M.;
Ganguly, A. K. J. Med. Chem. 1999, 42, 2125.
6. Njoroge, F. G.; Taveras, A. G.; Kelly, J.; Remiszewski, S.;
Mallams, A. K.; Wolin, R.; Afonso, A.; Cooper, A. B.;
Rane, D. F.; Liu, Y. T.; Wong, J.; Vibulbhan, B.; Pinto,
P.; Deskus, J.; Alvarez, C. S.; del Rosario, J.; Connolly,
M.; Wang, J.; Desai, J.; Rossman, R. R.; Bishop, W. R.;
Patton, R.; Wang, L.; Kirschmeier, P.; Ganguly, A. K. J.
Med. Chem. 1998, 41, 4890.
7. Njoroge, F. G.; Vibulbhan, B.; Pinto, P.; Strickland, C.;
Kirschmeier, P.; Bishop, W. R.; Girijavallabhan, V. Bioorg.
Med. Chem. Lett. 2004, 14, 5877, and references therein.
8. Ha-Ras processing in COS-7 African, green monkey
kidney cells was conducted as described in Bishop, W.
B.; Bond, R.; Petrin, J.; Wang, L.; Patton, R.; Doll, R.;
Njoroge, G.; Catino, J.; Schwartz, J.; Windsor, W.; Syto,
R.; Schwartz, J.; Carr, D.; James, L.; Kirschmeier, P. J.
Biol. Chem. 1995, 270, 30611, Ras proteins were visualized
using AttoPhos (JBL Scientific, Inc., San Luis Obispo,
CA) and quantified on a STORM phosphorimaging
device (Molecular Devices, Sunnyvale, CA). IC50 values
were calculated by fitting the data to the Michaelis–
Menten equation using the non-linear regression analysis
program Kaleidagraph (Synergy Software, Essex Junc-
tion, VT). All compounds were tested at least two times.
The data are reported as the average of multiple determi-
nations standard deviation (SD).
biotin-CVLS followed by
a 60 min incubation at
37 °C. Reactions were quenched with 150 ll of a cold
suspension containing 250 mM EDTA (Digene), pH
8.0, 0.5% bovine serum albumin (Sigma), and 200 lg
SPA beads.
13. Cell growth in soft agar was quantified as described in Du,
W.; Lebowitz, P. F.; Prendergast, G. C. Mol. Cell. Biol.
1999, 19, 1831. Colonies were photographed using a video
camera and tabulated using NIH Image software. IC50s
were calculated as described.8
14. To test for microsome stability, 5 ll of rat liver
microsomes, with
a cytochrome P450 content of
6.7 lM, was incubated for 60 min at 37 °C with
100 ll of 1 M KPi, pH 7.4, 100 ll freshly prepared
50 mM NADPH, and 10 lM compound (1 ll of a
10 mM solution in DMSO), all of which was adjusted
to 1.0 ml in H2O. The mixture was subsequently
extracted with 200 ll dichloromethane and 150 ll of
the organic phase was evaporated to dryness. The
residue was dissolved in 25 ll methanol and analyzed
by high performance liquid chromatography (HPLC)
using
a
Phenomenex (Torrance, CA) ODS C-18
column. Samples were eluted for 27 min at a rate of
0.5 ml/min with a gradient of 44% B to 100% B where
A consists of 0.1% TFA and B consists of 90% ACN
containing 0.085% TFA. The percent of the compound
remaining was quantified by comparison to samples
extracted after 0 min incubation.
9. Protein Data Bank (PDB) file name of FTase X-ray
structure is 105 M.
10. Taveras, A. G.; Kirchsmeier, P.; Baum, C. M. Curr. Top.
Med. Chem. 2003, 3, 1103.
11. (a) Afonso, A.; Baldwin, J. J.; Doll, R. J.; Li, G.; Mallams,
A. K.; Njoroge, F. G.; Rane, D. F.; Reader, J. C.;
15. Oliyai, R.; Stella, V. J. Ann. Rev. Pharmacol. Toxicol.
1993, 33, 521.