Evaluation of amino acid-based linkers in potent macrocyclic inhibitors of farnesyl-protein transferase

Article abstract of DOI:10.1016/S0960-894X(01)00340-7

A series of amino acid-based linkers was used to investigate the effects of various substituents upon the potency, pharmacokinetic properties, and conformation of macrocyclic farnesyl-protein transferase inhibitors (FTIs). As a result of the studies described herein, highly potent FTIs with improved pharmacokinetic profiles have been identified.

Full text of DOI:10.1016/S0960-894X(01)00340-7

Bioorganic & Medicinal Chemistry Letters 11 (2001) 1817–1821
Evaluation of Amino Acid-Based Linkers in Potent Macrocyclic
Inhibitors of Farnesyl-Protein Transferase
Douglas C. Beshore,a,* Ian M. Bell,a Christopher J. Dinsmore,a Carl F. Homnick,a
J. Christopher Culberson,d Ronald G. Robinson,b Christine Fernandes,b
Eileen S. Walsh,b Marc T. Abrams,b Hema G. Bhimnathwala,b Joseph P. Davide,b
Michelle S. Ellis-Hutchings,b Hans A. Huber,b Kenneth S. Koblan,b
Carolyn A. Buser,b Nancy E. Kohl,b Robert B. Lobell,b I-Wu Chen,c
Debra A. McLoughlin,c Timothy V. Olah,c Samuel L. Graham,a George D. Hartmana
and Theresa M. Williamsa
aDepartment of Medicinal Chemistry, Merck Research Laboratories, West Point, PA 19486, USA
bDepartment of Cancer Research, Merck Research Laboratories, West Point, PA 19486, USA
cDepartment of Drug Metabolism, Merck Research Laboratories, West Point, PA 19486, USA
dDepartment of Molecular Systems, Merck Research Laboratories, West Point, PA 19486, USA
Received 20 February 2001; accepted 25 April 2001
Abstract—A series of amino acid-based linkers was used to investigate the effects of various substituents upon the potency, phar-
macokinetic properties, and conformation of macrocyclic farnesyl-protein transferase inhibitors (FTIs). As a result of the studies
described herein, highly potent FTIs with improved pharmacokinetic profiles have been identified. # 2001 Elsevier Science Ltd. All
rights reserved.
Farnesyl-protein transferase (FPTase) is a zinc metal-
loenzyme that catalyzes the S-alkylation of a cysteine
residue in the C-terminal tetrapeptide Ca1a2X sequence
of proteins with farnesylpyrophosphate (FPP). This
post-translational modification localizes the proteins to
the cell membrane, enabling them to participate in sig-
nal transduction during cellular proliferation.1 The Ras
protein is a substrate of FPTase and has been impli-
cated in 20–30% of all human tumors. In these tumors,
mutant Ras loses its GTPase function, becomes con-
stitutively bound to GTP and, when localized to the cell
membrane, transmits growth signals independent of
extracellular growth factors.2 Strategies for controlling
such unimpeded oncogenic mutant ras-driven tumors
have focused on preventing the requisite cellular locali-
zation of Ras through the use of FPTase inhibitors
(FTIs), which have potential as chemotherapeutic
agents.3 FTIs have been identified that mimic the
protein substrate, FPP, or both, including small mole-
cule, nonpeptide, nonthiol inhibitors.4
In earlier studies, optimization of 1-aryl-2-piper-
azinones,5a in combination with data obtained from
transferred NOENMR experiments, led to the macro-
cyclic FTI 15b (IC50=0.1 nM, Fig. 1). Isomeric 25c
(IC50=3.5 nM) is over 10-fold less potent, which sug-
gests that the conformation of the amide and/or the
substitution of the amino acid backbone significantly
affects potency. Here we report the use of peptide-based
template 3 to explore the effects of substitution upon
conformation, potency, and the pharmacokinetic pro-
files of macrocyclic FTIs.
Scheme 1 depicts the synthesis of compounds 4–20. 8-
Amino-2-naphthol 21 was N-Boc protected, silylated,
*Corresponding author. Tel.:+1-215-652-7474; fax: +1-215-652-
7310; e-mail: douglas_beshore@merck.com
Figure 1. Macrocyclic inhibitors of FPTase.
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00340-7

1818
D. C. Beshore et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1817–1821
and deprotected to provide amine 22. PyBOP-mediated
acylation of various amino acids followed by removal of
protecting groups afforded amine 24, which was reduc-
tively alkylated with the previously reported aldehyde
265b to produce the corresponding acyclic phenol.
Macrocyclization to give compounds 4–16 was accom-
plished with cesium carbonate in DMF (ꢀ0.01 M to
prevent dimerization). For the series in which R1=Me,
carbamate reduction gave the N-methylnaphthalene 23,
which was converted to 25. Because forcing conditions
were required to acylate 23, partial racemization occur-
red and enantiopure 25 was obtained via chiral HPLC
separation. The amine was reductively alkylated with 26
and cyclized to provide compounds 17–20. The intro-
duction of an alkyl group to the secondary amine at the
R3 position was accomplished by reductive alkylation of
the parent macrocycle with formaldehyde or benzalde-
hyde to give compounds 11–16, 18, and 20. Homo-
logated macrocycles were synthesized in analogy with
previously reported methods5c by an EDC-mediated
coupling of the appropriately substituted amine (24 or
25) with [1-(4-cyano-3-fluorobenzyl)-1H-imidazol-5-
yl]acetic acid followed by cyclization to afford
compounds 29–32.
Scheme 2. Reagents and conditions: (a) HCl, NaNO2, KI, THF/H2O,
93%; (b) TIPSCl, imidazole, DMF, 40 ꢁC, 99%; (c) t-BuLi, CO2,
À78 ꢁC; H2O, TFA, rt, 43%; (d) Boc-diamine, PyBOP, DIEA, DCM;
(e) TBAF, THF; (f) HCl, EtOAc, 0 ꢁC; (g) 26, DIEA, AcOH,
NaCNBH3, MeOH; (h) Cs2CO3, DMF, 0.01 M, 65 ꢁC.
Table 1. FPTase and GGPTase-I in vitro inhibition and FPTase
binding in cell culture for compounds 4–20
Compd R1
R2
R3 FPTase FPTase GGPTase-I
in vitro binding in
IC50 cell culture
in vitro
IC50
(mM)d,b
(nM)a,b
IC50
(nM)c
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
H
H
H
H
H
H
H
H
H
H
H
H
H
Me
Me
Me
Me
H
Me (S)
Me (R)
E R) t (
Bn (R)
H
H
H
H
H
H
H
4.3
11
37
248
23
410
17
0.675
4.82
1.3e
1.2
0.313
0.469
0.066
0.072
0.147
2.07
0.47e
2.1
1.8e
2-ThienylCH2 (R)
3-PyridylCH2 (R)
43
1310
196
>100
184
29.6
240
574
55
H
Me 5.7
Me 7.0
Me 2.7
Bn 2.6
Bn 18
Me (R)
Bn (R)
H
Me (R)
Bn (R)
H
H
Me (R)
Me (R)
1.22
0.350
>20
>20
>20
6.72
12.8
3.16
6.76
Scheme 1. Reagents and conditions: (a) Boc2O, THF/DCM, 70 ꢁC,
86%; (b) TBDPSiCl, imidazole, DMF, 60 ꢁC, 96%; (c) HCl (g),
EtOAc, 0 ꢁC; (d) NaH, MEMCl, THF, 0 ꢁC–rt, 58%; (e) LAH, THF,
65 ꢁC, 88%; (f) Boc-AA-OH, PyBOP, DIEA, DCM; (g) TBAF, THF;
(h) 26, DIEA, AcOH, NaCNBH3, MeOH; (i) Cs2CO3, DMF, 0.01 M,
65 ꢁC; (j) aldehyde, AcOH, NaCNBH3 MeOH.
Bn 50
H
15
Me 8.4
16
Me 14
13
17
15
H
Scheme 2 depicts the synthesis of compounds 33 and 34.
Compound 21 was converted to 8-iodo-2-naphthol and
protected as the TIPS ether. Subsequent transmetala-
tion and carboxylation provided 27, which was acylated
with a mono-Boc protected diamine. Unmasking of the
terminal amino group provided 28, which was con-
verted to macrocycles 33 and 34. Compound 35 was
synthesized from 28 via the aforementioned EDC-
mediated coupling, which was subsequently cyclized.
aConcentration of compound required to reduce the human FPTase-
catalyzed incorporation of [3H]FPP into recombinant Ras-CVIM by
50% (ref 7a).
bAll values are n=2, unless otherwise noted.
cConcentration of compound required to displace 50% of a potent,
radiolabeled FTI from FPTase in cultured Ha-ras transformed RAT1
cells (ref 7b).
dConcentration of compound required to reduce the human
GGPTase-I-catalyzed incorporation of 3H GGPP (100 nM) into
a biotinylated K-Ras-derived peptide [1.6 mM; biotinylated peptide
corresponding to the C-terminus of human K-Ras (b-
GKKKKKKSKTKCVIM, Research Organics)] by 50%. Assay was
run with a 30 min preincubation of enzyme and inhibitor in the pre-
sence of 5 mM ATP (ref 7c).
Table 1 shows the effects of modifications to the amino
acid template. Compounds were assayed for in vitro
activity against FPTase resulting in IC50 values for
measured inhibition. Variation at R2 revealed a strong
preference for a benzyl substituent (8) and moderate
tolerance for differences in size and polarity (4–10).
Additionally, the chirality imposed by the addition of
an a-R2 substituent affected potency; the (R) enantio-
eEnzyme concentration 10 pM instead of 1 nM in assay, n=2.
mer was preferred over the (S) by 2- to 10-fold (cf. 5 vs
6).6 Trends associated with substitution at R3 (11–16,
18, and 20) were dependent upon substitution at R2.
While a large substituent was tolerated at R3 (14), a

D. C. Beshore et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1817–1821
1819
simultaneous increase in size of R2 led to diminished
inhibitory activity. This suggests that both substituents
are trying to access the same hydrophobic pocket in the
FPTase active site (cf. 16 and 15 vs 14). With the addi-
tion of a methyl group to the R1 position (17–20), a 2-
to 10-fold loss of in vitro FPTase inhibitory activity was
observed (cf. 19 vs 6). This is perhaps due to the con-
formational differences between N-H and N-Me aryl
amides (vide infra).8
slight loss in potency upon methylation of the N-aryl
amide (cf. 30 vs 29). Inversion of the amide connectivity
also did not significantly alter potency in vitro (33–35).
Neither additional homologation to an 18-membered
macrocycle nor the presence or absence of a carbonyl
significantly altered inhibitory activity (cf. 34 and 35 vs
33).
Activity in cell culture (FPTase binding assay) was
measured by an FTI’s ability to displace a potent
radiolabeled ligand from FPTase.7b Compounds 4–16
exhibited higher IC50 values in cell culture compared to
the in vitro assay, perhaps due to poor cellular penetra-
tion. Noteworthy was 10, which contains a side chain
with increased polarity that resulted in exceptionally
poor activity in cell culture. When the amide N-H was
replaced with N-Me (R1=Me), cell activity approached
or equaled intrinsic potency (17–20). Homologated
macrocycles 30–35 also exhibited higher IC50 values in
cell culture. However, conversion of an amide to an
amino group increased activity in cell culture (cf. 30 vs
17 and 35 vs 33).
Table 2 shows homologated macrocycles 29–35. Pre-
vious work suggested that to obtain a compound as
potent as 1, ring expansion of the macrocycle may be
necessary.5c Unfortunately, homologation to a 17-
membered ring provided no added potency (cf. 29 vs 4).
Table 2. FPTase and GGPTase-I in vitro inhibition and FPTase
binding in cell culture for 17- and 18-membered macrocycles 29–35
Compounds were also assayed for in vitro activity
against the related enzyme geranylgeranyl-protein
transferase-I (GGPTase-I), which prenylates cellular
proteins in a manner analogous to FPTase.7c Amide
methylation (R1=Me) increased the GGPTase-I IC50
by 4- to 10-fold (cf. 17 vs 4). Similar to FPTase SAR,
substitution at R2 with small alkyl groups provided no
significant change in activity (cf. 6 vs 4), but the addi-
tion of a larger hydrophobic group increased potency
nearly 10-fold (cf. 8 vs 4). The addition of a methyl
group at R3 generally resulted in a 2-fold or greater
decrease in inhibitory activity (cf. 18 vs 17). However,
the addition of a benzyl group to R3 rendered all com-
pounds inactive, regardless of the substitution at R1 or
R2 (14–16). Homologated macrocycles 29–32 exhibited
similar trends; alkylation of the aryl amide resulted in a
decrease in potency (cf. 30 vs 29), and the addition of
lipophilicity to the amino acid side chain lowered
GGPTase-I IC50 values (cf. 32 vs 29). GGPTase-I
potency was generally increased upon homologation to
a 17-membered macrocycle (33 and 29–32) relative to
compounds in Table 1 (cf. 30 vs 17). However, the
additional homologation to an 18-membered ring (34
and 35) resulted in a decrease in GGPTase-I inhibitory
activity (cf. 34 vs 33).
Compd
X
FPTase
in vitro
FPTase
binding in
GGPTase-1
in vitro
IC50 (nM)a,b cell culture IC50 (mM)d,b
IC50 (nM)c
29
30
31
32
33
34
35
3.8
6.6
3.3
3.5
4.5e
2.3
2.5
67
65
0.150
1.09
258
310
0.050
0.029
0.077
0.89
7.1
99
409
0.250
Pharmacokinetic (PK) data for selected compounds are
shown in Table 3.9 While 1 is a potent FTI, its PK
properties are suboptimal. Conversely, 2 is over 10-fold
less potent against FPTase and exhibits an increased
half-life. Overall, the peptide-based macrocycles descri-
bed herein have reduced clearances compared to 1,
resulting in increased half-lives. The 3-pyridylmethyl
substituted macrocycle 10 was cleared more rapidly,
although it had an improved volume of distribution.
aSee footnote a in Table 1.
bAll values are n=2, unless otherwise noted.
cSee footnote c in Table 1.
dSee footnote d in Table 1.
eEnzyme concentration 10 pM instead of 1 nM in assay, n=2.
While the 16-membered macrocycles in Table 1 exhib-
ited a preference for a benzyl substituent (8), the
homologated series was unaffected by amino acid side-
chain substitution in the FPTase in vitro assay (cf. 32
and 31 vs 29). Homologated compounds exhibited a
In an effort to understand the consistent reduction in
FTase inhibitory activity as a result of aryl amide N-
methylation, we examined the lowest energy conforma-
tions of the potent inhibitor 6 and its N-Me analogue 19

1820
D. C. Beshore et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1817–1821
Table 3. Pharmacokinetic profiles for selected compounds in dogsa
The studies described herein have examined the SAR of
amino acid-based linkers of 16-, 17-, and 18-membered
macrocyclic FPTase inhibitors. We have identified a
series of highly potent compounds that exhibit
improved pharmacokinetic profiles relative to 1. Con-
formational analysis has provided an enhanced under-
standing of structural preferences for the binding of
macrocyclic inhibitors that contain an amino acid-based
linker.
Compd
T1/2 (h)
Cl (mL/min/kg)
VD (l/kg)
1
2
4
6
8
9
10
29
1.0
3.5
3.0
4.8
2.2
2.4
0.8
3.4
5.1
1.1
0.44
0.30
0.55
0.37
0.45
0.26
0.98
0.35
2.36
0.95
2.23
1.65
17.3
1.49
aCompounds were administered intravenously to two dogs (1 mg/kg)
along with 11 other compounds and an internal standard. Plasma
extracts were analyzed by LC/MS/MS and reported data is the aver-
age of two dogs (see ref 9).
Acknowledgements
The authors thank Dr. A. S. Kim and Dr. B. W. Trotter
for manuscript suggestions and Dr. C. S. Hamann for
compound analysis of FPTase inhibition.
(Fig. 2, top).10 The most prominent structural difference
is the conformation about the amide N-C(O) bond. The
N-H amide in 6 adopts a cis amide conformation with
respect to the naphthyl and the carbonyl of the amide,
while the N-Me analogue 19 adopts the opposite amide
conformation. These results are consistent with pre-
viously reported data regarding the lowest energy con-
formations of N-H and N-Me anilides.8 As a result of
these differences, the trajectory of the naphthyl group
has changed by nearly 90ꢁ. It has been shown in pre-
vious studies that the overall orientation of the carbonyl
group within the active site of FPTase is an important
feature of structurally related, highly potent FTIs.5d The
lowest calculated energy conformation of 15b places the
carbonyl group in the presumed enzyme-bound con-
formation.5d Best-fit overlays of 6 and 19 onto 1 (Fig. 2,
bottom) require that they deviate from their lowest
energy conformations by 5.7 and 11.5 kcal/mol, respec-
tively. The higher energy required for the N-Me amide
to adopt this presumed FPTase-bound conformation
relative to 6 may account for its reduced potency. These
results reinforce the notion that the orientation of the
carbonyl group is important, and imply that the trajec-
tory of the naphthyl group may also play a role in
obtaining highly potent FTIs.
References and Notes
1. (a) Kato, K.; Cox, A. D.; Hisaka, M. M.; Graham, S. M.;
Buss, J. E. Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 6403. (b)
Rowinsky, E. K.; Windle, J. L.; Von Hoff, D. D. J. Clin.
Oncol. 1999, 17, 3631.
2. Rodenhuis, S. Semin. Cancer Biol. 1992, 3, 241.
3. (a) Gibbs, J. B. Cell 1991, 65, 1. (b) Gibbs, J. B.; Oliff, A. I.
Annu. Rep. Biopharmacol. Toxicol. 1997, 37, 143.
4. (a) Recent reviews: Williams, T. M. Expert Opin. Ther. Pat.
1999, 9, 1263. (b) Dinsmore, C. J. Curr. Opin. Oncol. Endocr.
Metab. Invest. Drugs 2000, 2, 26. (c) Bell, I. M. Expert Opin.
Ther. Pat. 2000, 10, 1813.
5. (a) Williams, T. M.; Bergman, J. M.; Brashear, K.; Breslin,
M. J.; Dinsmore, C. J.; Hutchinson, J. H.; MacTough, S. C.;
Stump, C. A.; Wei, D. D.; Zartman, C. B.; Bogusky, M. J.;
Culberson, J. C.; Buser-Doepner, C.; Davide, J.; Greenberg,
I. B.; Hamilton, K. A.; Koblan, K. S.; Kohl, N. E.; Liu, D.;
Lobell, R. B.; Mosser, S. D.; O’Neill, T. J.; Rands, E.; Scha-
ber, M. D.; Wilson, F.; Senderak, E.; Motzel, S. L.; Gibbs,
J. B.; Graham, S. L.; Heimbrook, D. C.; Hartman, G. D.;
Oliff, A. I.; Huff, J. R. J. Med. Chem. 1999, 42, 3779. (b)
Dinsmore, C. J.; Bogusky, M. J.; Culberson, J. C.; Bergman,
J. M.; Homnick, C. F.; Zartman, C. B.; Mosser, S. D.; Scha-
ber, M. D.; Robinson, R. G.; Koblan, K. S.; Huber, H. E.;
Graham, S. L.; Hartman, G. D.; Huff, J. R.; Williams, T. M.
J. Am. Chem. Soc. 2001, 123, 2107. (c) Bell, I. M.; Beese, L. S.;
Beshore, D. C.; Bogusky, M. J.; Buser-Doepner, C.; Culber-
son, J. C.; Gallicchio, S. N.; Gibbs, J. B.; Graham, S. L.;
Hartman, G. D.; Heimbrook, D. C.; Huber, H. E.; Kassahun,
K.; Koblan, K. S.; Kohl, N. E.; Lobell, R. B.; Rodrigues, D.
A.; Taylor, J. S.; Williams, T. M. Enzyme Mechanisms 17th
Conference, Marco Island, FL, Jan 3–7, 2001. (d) Dinsmore,
C. J.; Bergman, J. M.; Bogusky, M. J.; Culberson, J. C.;
Hamilton, K. A.; Graham, S. L. Org. Lett. 2001, 3, 865.
6. The (S) enantiomers of compounds 7, 8, 12, 13, 15, 16, 19,
and 20 fell within this range (data not shown).
7. (a) Graham, S. L.; deSolms, S. J.; Giuliani, E. A.; Kohl,
N. E.; Mosser, S. D.; Oliff, A. I.; Pompliano, D. L.; Rands, E.;
Breslin, M. J.; Deanna, A. A.; Garsky, V. M.; Scholz, T. H.;
Gibbs, J. B.; Smith, R. L. J. Med. Chem. 1994, 37, 725. (b)
Lobell, R.; Gibson, R.; Davide, J.; Kohl, N.; Burns, D.
Manuscript in preparation. To assess the potency of the FTIs
in cells, we utilized a competitive binding assay to determine
the concentration of the FTI required to displace 50% of a
radiolabeled FTI from FPTase in cultured v-Ha-ras trans-
formed RAT1 cells. This radiotracer, [125I] 4-{[5-({(2S)-4-(3-
iodophenyl)-2-[2-(methylsulfonyl)ethyl]-5-oxopiperazin-1-yl}-
methyl)-1H-imidazol-1yl]methyl} benzonitrile, has ꢂ50,000
Figure 2. Top: Overlay of calculated lowest energy conformations of 6
(pink) and 19 (green). Bottom: Overlay of 6 (pink, 5.7 kcal/mol relative
to lowest energy conformation) and 19 (green, 11.5 kcal/mol relative to
lowest energy conformation) with the lowest calculated energy con-
formation of 1 (see ref 10).

D. C. Beshore et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1817–1821
1821
high affinity binding sites per cell (apparent Kd of ꢂ 1 nM) in
the Rat1 cell line. The nonspecific binding signal, determined
by the addition of 1000-fold excess unlabeled competitor FTI,
is typically 5-fold lower than the specific binding signal. The
assay provides comparable results using a variety of cell lines,
and is performed as follows. Cells are seeded at 200,000 cells
per well in 24-well tissue culture plates and cultured for 16 h.
The radiotracer (ꢂ300–1000 Ci/mmol) is diluted into culture
media to a concentration of 1 nM, along with the desired
concentration of test FTI, and then added to the cell mono-
layers. After a 4 h incubation at 37 ꢁC, the cells are briefly
rinsed with phosphate-buffered saline, removed from the cul-
ture plate by trypsinization, and then subjected to gamma
counting using a CobraII1 gamma counter (Packard Instru-
ment Company). Dose–inhibition curves and IC50 values are
derived from a 4-parameter curve-fitting equation using Sig-
maPlot1 Software. (c) Huber, H. E.; Abrams, M.; Anthony,
N. J.; Graham, S. L.; Hartman, G. D.; Lobell, R. B.; Lumma,
W. L.; Nahas, D.; Robinson, R.; Sisko, J. C.; Heimbrook, D.
C. Proc. Am. Assoc. Cancer Res. 2000, 41, Abstract 2838.
8. Itai, A.; Saito, S.; Toriumi, Y.; Tomioka, N. J. Org. Chem.
1995, 60, 4715.
9. Olah, T. V.; McLoughlin, D. A.; Gilbert, J. D. Rapid
Commun. Mass Spectrom. 1997, 11, 17.
10. (a) Molecular modeling conformations were generated
using metric matrix distance geometry algorithm JG (S.
Kearsley, Merck & Co., Inc., unpublished). The generated
structures were energy-minimized within Macromodel (ref
10b) using MMFF force field. (b) Mohamadi, F.; Richards,
N. G. J.; Guida, W. C.; Liskamp, R.; Caufield, C.; Chang,
G.; Hendrickson, T.; Still, W. C. J. J. Comput. Chem. 1990,
11, 440.