Structure-based optimization of a potent class of arylamide FMS inhibitors

Article abstract of DOI:10.1016/j.bmcl.2008.04.059

An anti-inflammatory 1,2,4-phenylenetriamine-containing series of FMS inhibitors with a potential to form reactive metabolites was transformed into a series with equivalent potency by incorporation of carbon-based replacement groups. Structure-based modeling provided the framework to efficiently effect this transformation and restore potencies to previous levels. This optimization removed a risk factor for potential idiosyncratic drug reactions.

Full text of DOI:10.1016/j.bmcl.2008.04.059

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 3632–3637
Structure-based optimization of a potent class
of arylamide FMS inhibitors
Sanath K. Meegalla, Mark J. Wall, Jinsheng Chen, Kenneth J. Wilson,
Shelley K. Ballentine, Renee L. DesJarlais, Carsten Schubert, Carl S. Crysler,
Yanmin Chen, Christopher J. Molloy, Margery A. Chaikin, Carl L. Manthey,
Mark R. Player, Bruce E. Tomczuk and Carl R. Illig*
Johnson & Johnson Pharmaceutical Research & Development, Welsh & McKean Roads, Spring House, PA 19477, USA
Received 4 March 2008; accepted 23 April 2008
Available online 26 April 2008
Abstract—An anti-inflammatory 1,2,4-phenylenetriamine-containing series of FMS inhibitors with a potential to form reactive
metabolites was transformed into a series with equivalent potency by incorporation of carbon-based replacement groups. Struc-
ture-based modeling provided the framework to efficiently effect this transformation and restore potencies to previous levels. This
optimization removed a risk factor for potential idiosyncratic drug reactions.
Ó 2008 Elsevier Ltd. All rights reserved.
The macrophage colony-stimulating factor (CSF-1) is the
primary growth factor for the macrophage lineage and it
CN
specifically binds to FMS, a type III receptor tyrosine ki-
nase expressed by macrophages and their progenitor
cells.1 The binding of FMS by CSF-1 initiates an intracel-
lular signaling cascade resulting in survival, proliferation,
and differentiation of these cells.2 As macrophages are
known to play an important role in the inflammatory pro-
cess, the inhibition of CSF-1-dependent macrophage pro-
liferation might be of therapeutic value in intercepting an
inflammatory process.3 This hypothesis has also been val-
idated by various animal studies.4,5 CSF-1-deficient mice
are reported to be resistant to collagen-induced arthritis
(CIA). In a CIA model, CSF-1 was shown to increase
the severity of disease while a neutralizing anti-CSF-1
antibody had the opposite effect.
N
N
H
N
H
N
2
CN
1
O
N
H
4
O
O
N
N
N
N
IC = 0.001 μM
50
IC = 0.0008 μM
50
1a
1b
Figure 1. Anti-inflammatory FMS inhibitors 1a and 1b.
Despite the excellent potency and in vivo efficacy of this
series, the 1,2,4-phenylenetriamine core structure of 1a
and 1b was considered a liability due to its potential to
form reactive quinonediimine metabolites, which might
haptenize proteins and deplete glutathione, leading to
tissue damage (Fig. 2).
In recent publications,6 the discovery of the potent FMS
inhibitors 1a and 1b (Fig. 1) was described and results
demonstrating the anti-inflammatory activity of 1b in
a mouse model of collagen-induced arthritis were
presented.6b
In general, there are two types of adverse drug reactions
that can occur. Type A adverse drug reactions are pre-
dictable, dose-dependent, and related to the pharmacol-
ogy of the compound. Type B are idiosyncratic drug
reactions (IDRs) and are of low incidence, human-spe-
cific, have a delayed onset, and account for a significant
percentage of New Chemical Entity (NCE) attrition
during development. The mechanisms of idiosyncratic
toxicity may encompass one or more of the following:
the formation of reactive metabolites in certain patients
Keywords: FMS; cFMS; CSF-1R; M-CSF; Colony-stimulating factor-
1; Macrophages; Anti-inflammatory activity; Idiosyncratic drug
reactions.
*
Corresponding author. Tel.: +1 610 458 6057; fax: +1 610 458
8249; e-mail: cillig@prdus.jnj.com
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.04.059

S. K. Meegalla et al. / Bioorg. Med. Chem. Lett. 18 (2008) 3632–3637
3633
Excretion/ Elimination
Reactive Metabolite
(2b and 2c) with or without cleavage of the amide bond
that may then react with nucleophilic sites of endoge-
nous proteins. These haptenized proteins (2d and 2e)
could potentially act as antigens, thereby generating an
IDR.
Detoxification
Bioactivation
DRUG
Haptenization
GSH depletion
Oxidative Stress
Immune Response
In order to quickly assess the potential to produce pro-
tein adducts, the propensity to form glutathione conju-
gates in vitro was evaluated with representative
compounds. The test compound at 20 or 40 lM was
incubated with NADPH (1 mM), GSH (5 mM) in PBS
(0.1 mM), and 1 mg/mL of mouse liver microsomes at
37 °C for 1 h. The products were then analyzed by LC/
MS for the glutathione conjugate of the parent
(P+GSH), the amide bond cleavage product
(ACP+GSH), and their fragments. Several compounds
containing the 1,2,4-phenylenetriamine core formed
GSH conjugates under these conditions. Some examples
are shown in Table 1.
Tissue Damage
Figure 2. Idiosyncratic toxicity due to reactive metabolite formation.
due to the presence of polymorphs in drug-metabolizing
enzymes, immune-mediated responses to the drug or its
metabolites, the combination of drugs with low-level
inflammatory reactions or drug-induced mitochondrial
toxicity. Clinical manifestations include hepatotoxicity,
anaphylaxis, blood dyscrasias, and skin reactions.7
We hypothesized that the 1,2,4-phenylenetriamine core
structure of 1a and its congeners could result in reactive
metabolite formation and pose a risk factor for IDRs.
As shown in Figure 3, the parent compound 2 could
generate reactive quinonediimine-type intermediates
In light of these results, we directed our lead optimiza-
tion efforts to address this potential problem with the
aid of structure-based drug design. The primary focus
was the replacement of the nitrogen substituents at the
C-2 and C-4 positions with carbon substituents in order
to minimize the possibility of reactive intermediate for-
mation through in vivo oxidation. This was approached
in a stepwise fashion at the C-2 and then the C-4 posi-
tions of the benzenoid core.
N
H
Parent (P)
CN
N
O
O
R2N
2
The synthesis of C-2 carbon-substituted, C-4 piperazinyl
arylamides is shown in Scheme 1. The N-methylpipera-
zine at C-4 was first introduced by aromatic nucleophilic
substitution of 4-chloronitrobenzene and the product 9
was reduced to the corresponding aniline 10. The
electrophilic bromination of 10 installed the bromine
ortho to the amino function to yield product 11. The
ortho-bromoaniline 11 was then employed in a Suzuki
coupling reaction to introduce the C-2 carbocycle. The
Amide Bond
Cleavage
O
+
N
N
CN
NH2
N
O
O
R2N
R2N
2a Amide Cleavage Product
(ACP)
2b
Reactive Intermediate
O
HX-Protein
Table 1. In vitro GSH conjugate formation of 1,2,4-phenylenetri-
amine-containing arylamides
+
R1
N
N
H
N
NH
CN
O
N
H
O
R2N
R2N
X-Protein
N
R2
2c
2d
Reactive Intermediate
P + Protein
O
N
(most likely adduct)
X
HX-Protein
Compound
P+GSH
ACP+GSH
Yes
3
R1 = H, X = N-CH3
Yes
Yes
Yes
Yes
Yes
N
R2 = 5-cyanofuranoyl
R1 = CH3, X = N-CH3
R2 = 5-cyanofuranoyl
R1 = H, X = N-CH3
R2 = 4-cyanopyrrolyl
R1 = CH3, X = N-CH3
R2 = 4-cyanopyrrolyl
R1 = CH3, X@CH2
Potential Immune
Response
(Idiosyncratic Toxicity)
NH2
4
5
6
7
Yes
Yes
Yes
Yes
R2N
X-Protein
ACP + Protein
(most likely adduct)
2e
Figure 3. Potential mechanisms of idiosyncratic toxicity of a 1,2,4-
phenylenetriamine-containing arylamide scaffold.
R2 = 4-cyanopyrrolyl

3634
S. K. Meegalla et al. / Bioorg. Med. Chem. Lett. 18 (2008) 3632–3637
NH2
NO2
NO2
c
a
b
N
Cl
N
N
N
8
9
10
Br
A
A
H
N
CN
NH2
NH2
d
e
O
O
N
N
N
N
N
N
11
13
12
Scheme 1. Synthesis of C-2 carbon-substituted derivatives. Reagents: (a) 4-methylpiperazine; (b) Pd/H2; (c) NBS/CH2Cl2; (d) A-B(OH)2 or boronate
esters (see text)/Pd(Ph3P)4; (e) 5-cyanofurancarbonyl chloride/Et3N.
amino function of the resulting product 12 was then uti-
lized for amide bond formation with 5-cyanofuranoyl
chloride to obtain 13.
conformation toward that of the piperidine. Neverthe-
less, the selected ortho-substituted aromatic ring substit-
uents at C-2 drastically reduced the FMS enzyme
inhibitory activity (13a–d). Cyclohexyl was previously
shown to be an inadequate replacement for piperidine
(data not shown) since the ring cannot adopt the par-
tially planar nature imposed by the piperidine nitrogen.
However, both the predicted partial planarity and con-
formation of a cycloalkene ring appeared promising by
modeling. Although the cycloalkene substituents (com-
pounds 13e–g) were also less potent than 1, the loss of
potency was much less profound. The 1-cyclohexenyl
substituted analogue 13f exhibited the highest FMS en-
zyme potency while the methyl substituent on the ‘pseu-
do ortho’ position of the cyclohexene (compound 13g)
further reduced the inhibitory potency, presumably ow-
ing to undesirable conformational changes in the cyclo-
hexene ring system. Unfortunately, the 1-cyclohexenyl
substituent (13f) was still 18-fold less potent than the
parent compound 1a.
FMS IC50 values of the C-2 piperidine substituent
replacements are shown in Table 2. These replacement
substituents were chosen based on their ability to mimic
the conformation of the piperidine ring of the parent
compound. Although related in size to piperidine, an
unsubstituted phenyl ring at C-2 was predicted by mod-
eling to adopt a conformation nearly coplanar with the
central ring and quite different from the nearly orthogo-
nal conformation of the piperidine, while addition of an
ortho group on the phenyl should favorably bias the
Table 2. FMS potency of C-4 methylpiperazinyl, C-2 carbon-substi-
tuted analogues
A
H
CN
N
O
O
The general synthetic sequence for the synthesis of C-4
carbon-substituted, C-2 piperidinyl arylamides is shown
in Scheme 2. 4-Methylpiperidine was first introduced by
a selective nucleophilic substitution reaction on 4-bro-
mo-2-fluoronitrobenzene to furnish intermediate 15.
This product was then subjected to a metal-catalyzed
coupling reaction to introduce the C-4 carbon substitu-
ent to yield 16. The intermediates for compounds 18a
and 18b were made by Suzuki and Heck coupling reac-
tions of intermediate 15 with 4-pyridylboronic acid and
tert-butyl acrylate, respectively. Compound 18c was
made by acid-catalyzed hydrolysis of the tert-butyl ester
of compound 18b. To introduce acetylenic substituents,
the intermediate 15 was subjected to Sonogoshira cou-
pling reaction with propargyl alcohol and the resulting
intermediate was converted to the corresponding mesy-
late. This mesylate was then reacted with dimethylamine
and morpholine to obtain intermediates for the synthesis
of compounds 18d and 18f, respectively, and these inter-
mediates were then employed in the reduction step (c)
using Fe/NH4Cl as the reducing agent. Compound 18e
was made in an analogous manner using methylamine
as the nucleophile followed by methanesulfonylation
prior to the Fe/NH4Cl reduction of the nitro group.
The intermediate 16 for compound 18g was made by a
Suzuki coupling reaction of intermediate 15 with N-
Boc-protected 1,2,3,6-tetrahydropyridine-4-pinacolbor-
N
N
a
Compound
A
IC50 (lM)
H3C
13a
0.63
F
13b
13c
13d
13e
13f
13g
0.89
S
0.47
H3C
S
0.94
CH3
0.060
0.018
0.054
O
a Reported IC50 values are the means of three experiments. Inter-assay
variance was <20%. For assay details see Ref. 6a.

S. K. Meegalla et al. / Bioorg. Med. Chem. Lett. 18 (2008) 3632–3637
3635
diate for the synthesis of compound 18h. The amine
intermediate 17 was acylated with 5-cyanofuranoyl chlo-
ride, which provided the product 18. Compound 18i was
F
N
N
NO2
NO2
NO2
a
b
c
Br
Br
B
14
15
16
N
N
H
CN
N
NH2
O
d
O
B
B
17
18
Scheme 2. Synthesis of C-4 carbon-substituted derivatives. Reagents:
(a) 4-methylpiperidine; (b) BL/Pd(Ph3P)4; (c) [H]; (d) 5-cyanofuran-
carbonyl chloride/Et3N.
onate. Pd/H2 reduction of this intermediate provided the
intermediate necessary for the synthesis of compound
18h and Fe/NH4Cl reduction gave the required interme-
Table 3. FMS potency of C-2 piperidinyl, C-4 carbon-substituted
analogues
N
H
CN
N
O
O
B
a
Compound
B
IC50 (lM)
18a
0.01
N
O
18b
18c
18d
18e
18f
18g
18h
18i
0.038
0.002
0.015
0.016
0.005
0.0008
0.0003
0.0008
O
O
O
H
N
N
SO2Me
O
N
NH
NH
Figure 4. (A) Crystal structure of FMS, cyan carbons, with furan-
containing compound, magenta carbons (PDB ID 2i0y)8; (B) model of
FMS, cyan carbons, with pyrrole-containing compound, salmon
carbons; (C) model of FMS, cyan carbons, with imidazole-containing
compound, orange carbons. Hydrogen bonds between the inhibitor
and FMS as well as inhibitor intramolecular hydrogen bonds are
shown as green dashed lines. Figures were prepared with PyMol.9
NAc
a Reported IC50 values are the means of three experiments. Inter-assay
variance was <20%. For assay details see Ref. 6b.

3636
S. K. Meegalla et al. / Bioorg. Med. Chem. Lett. 18 (2008) 3632–3637
NC
NC
lized by an internal hydrogen bond between the furan
NC
N
N
N
N
N
SEM
a
b
c
Br
oxygen and the amide NH. This interaction cannot be
made by the 4-cyanopyrrole. However, a model of an
inhibitor containing cyanopyrrole bound to FMS indi-
cated the presence of an additional intermolecular
hydrogen bond between the pyrrole nitrogen and the
carbonyl of Glu 663 (Fig. 4B) and this alternate inter-
molecular interaction likely compensated for the ab-
sence of the intramolecular hydrogen bond. Based on
these observations, we reasoned that an inhibitor that
could make both of these interactions should be more
potent and this notion thus led to the design of 4-cyano-
imidazole-containing inhibitors (Fig. 4C).
N
H
19
SEM
20
21
NC
NC
N
N
CO Et
CO K
2
2
d
N
N
SEM
22
SEM
23
Scheme 3. Synthesis of cyanoimidazole carboxylic acid potassium salt,
23. Reagents and conditions: (a) SEMCl/K2CO3/acetone; (b) NBS/
AIBN, CCl4/60 °C; (c) i-PrMgCl/CNCO2Et; (d) KOH/EtOH.
synthesized by the acetylation of 18h. With compounds
18g and 18h, the Boc groups were then removed with
TFA to obtain the final products.
The required 4-cyanoimidazole carboxylic acid was syn-
thesized in SEM-protected form as its potassium salt 23
in four steps from 4-cyanoimidazole as shown in Scheme
3.
The FMS inhibitory activity of C-4 carbon-substituted
analogues shown in Table 3 suggests that the C-4 posi-
tion of the 2-piperidinyl aryl amide scaffolds has toler-
ance for a wide range of substituents. The C-4, 4-
pyridyl analogue 18a was 10-fold less active than the
parent 1a. The acyclic substituents such as substituted
propargyl, acrylate (18b, d–f) and other aliphatic amines
(not shown) were also less potent than the parent com-
pound, while 18c was the notable exception. However,
the replacement of the C-4 piperazine ring structure with
piperidine or tetrahydropyridine (18g–i) maintained the
extreme potency of the parent compound.
The route to the synthesis of this series of compounds
(Scheme 4) incorporates the Suzuki coupling of 4-anilin-
opinacolboronate with the enoltriflate of N-Boc-piperid-
inone to obtain intermediate 26 followed by
hydrogenation to obtain 27. Ortho-bromination with
NBS followed by a second Suzuki reaction with a cyclo-
alkenyl boronic acid or boronate introduced the C-2
cycloalkene. The resulting compound 29 was coupled
to the SEM-protected cyanoimidazole potassium salt
23 using PyBroP as the coupling agent. The final prod-
uct 31 was obtained by acid-induced deprotection of
intermediate 30. The selective acylation of the final
product produced the acylated piperidines 31e–g.
Although these efforts led to the discovery of potent
inhibitors containing suitable replacement C-4 substitu-
ents, the most active analogues still retained the C-2
piperidine ring structure. The best C-2 carbon substitu-
ent, 1-cyclohexenyl, was nearly 20-fold less active than
the parent compound.
The inhibitory activities of this final series of compounds
are shown in Table 4. As the modeling predicted, the 4-
cyanoimidazole replacement of 5-cyanofuran/4-cyano-
pyrrole enhanced the enzyme potency of C-2 carbon-
substituted derivatives by 10-fold (e.g., 31e vs 13f) thus
bringing the potency back into the range of the original
leads. Although the cyclopentene 31c was eightfold less
active than the 6-membered ring 31a, 7-membered cyclo-
alkenes (data not shown) displayed similar activity to
31a. The presence of small substituents such as mono-
The discovery that 5-cyanofuranoyl and 4-cyanopyrrol-
yl analogues 1a and 1b were equipotent6b led to a close
examination of the co-crystal structure of FMS kinase
with a 5-cyanofuran8 (Fig. 4A). The structure indicated
that the active conformation of the inhibitor was stabi-
OTf
N
Br
25
Boc
NH
NH
2
NH
2
2
NH
2
a
c
d
b
O
B
N
N
Boc
N
Boc
Boc
O
24
26
27
28
CN
CN
N
N
H
H
N
NH
N
2
N
N
H
e
f
SEM
O
O
N
N
HN
Boc
Boc
30
29
31
Scheme 4. General synthetic route to cyanoimidazole analogues. Reagents and conditions: (a) Pd(Ph3P)4/Na2CO3/dioxane; (b) 10% Pd/H2, EtOH; (c)
NBS/CH2Cl2/0 °C; (d) cycloalkenylboronic acid or boronate/Pd(Ph3P)4/Na2CO3/dioxane; (e) N-SEM-4-cyanoimidazolecarboxylic acid potassium
salt (23)/PyBroP/DIEA/CH2Cl2; (f) TFA/EtOH/CH2Cl2 (1/0.1/3).

S. K. Meegalla et al. / Bioorg. Med. Chem. Lett. 18 (2008) 3632–3637
Table 4. Biological activities of C-2, C-4 dicarbon-substituted analogues
3637
A
N
H
N
CN
N
H
O
B
a
Compound
A
B
IC50 (lM)
BMDM assay IC50 (lM)
1a
1b
0.001
0.0008
0.001
0.002
0.002
0.001
31a
NH
NH
NH
NH
31b
31c
31d
31e
31f
31g
0.0004
0.008
0.06
0.0036
0.03
>1
CH3
SO2
0.002
>0.15
0.06
0.008
>1
NAc
NAc
NAc
NH
0.08
SO2
a Reported IC50 values are the means of three experiments. Inter-assay variance was <20%. For assay details see Ref. 6b.
methyl 31b at the C-40 position of cyclohexene also did
not affect the FMS enzyme potency. However, polar
moieties in the ring (31d, 31f, and 31g) decreased the en-
zyme potency as well as the cellular potency. All highly
potent compounds also showed comparable cellular
References and notes
1. Sherr, C. J. Biochim. Biophys. Acta 1988, 948, 225.
2. (a) Pixley, F. J.; Stanley, E. R. Trends Cell Biol. 2004, 14,
628; (b) Sherr, C. J.; Roussel, M. F.; Rettenmier, C. W. J.
Cell Biochem. 1988, 38, 179.
3. (a) Yanni, G.; Whelan, A.; Feighery, C.; Bresnihan, B. Ann.
Rheum. Dis. 1994, 53, 39; (b) Yang, N.; Isbel, N. M.;
Nikolic-Paterson, D. J.; Li, Y.; Ye, R.; Atkins, R. C.; Lan,
H. Y. Kidney Int. 1998, 54, 143; (c) Isbel, N. M.; Nikolic-
Patterson, D. J.; Hill, P. A.; Dowling, J.; Atkins, R. C.
Nephrol. Dial. Transplant 2001, 16, 1638.
activity in
a
bone marrow-derived macrophage
(BMDM) proliferation assay.6a
Having successfully identified carbon-based replace-
ments for the C-2 and C-4 nitrogen substituents in par-
ent compounds 1a and 1b, the more potent new
analogues 31a, 31b, and 31e were evaluated for their po-
tential to exhibit IDRs as measured in the GSH conju-
gation assay. No GSH adduct formation was detected
under the assay conditions described previously.
4. Campbell, I. K.; Rich, M. J.; Bischof, R. J.; Hamilton, J. A.
J. Leukoc. Biol. 2000, 68, 144.
5. Cook, A. D.; Braine, E. L.; Campbell, I. K.; Rich, M. J.;
Hamilton, J. A. Arthritis Res. 2001, 3, 293.
6. (a) Patch, R. J.; Brandt, B. M.; Asgari, D.; Baindur, N.;
Chadha, N. K.; Georgiadis, T.; Cheung, W. S.; Petrounia,
I. P.; Chaikin, M. A.; Player, M. R. Bioorg. Med. Chem.
Lett. 2007, 22, 6070; (b) Illig, C. R.; Chen, J.; Wall, M. J.;
Wilson, K. J.; Ballentine, S. K.; Rudolph, J. M.; DesJarlais,
R. L.; Chen, Y.; Schubert, C.; Petrounia, I. P.; Crysler, C.
S.; Molloy, C. J.; Chaikin, M. A.; Manthey, C. L.; Player,
M. R.; Tomczuk, B. E.; Meegalla, S. K. Bioorg. Med.
Chem. Lett. 2008, 18, 1642.
7. (a) Uetrecht, J. P. Curr. Drug Met. 2000,1,133;(b)Lasser,K.E.;
Allen, P. D.; Woolhandler, S. J.; Himmelstein, D. U.; Wolfe, S.
M.; Bor, D. H. JAMA 2002, 287, 2215; (c) Nasser, A.-E. F.;
Lopez-Ananya, A. Curr. Opin. Drug Disc. Dev. 2004, 7, 126.
8. Schubert, C.; Schalk-Hihi, C.; Struble, G. T.; Ma, H. C.; Petrounia,
I. P.; Brandt, B.; Deckman, I. C.; Patch, R. J.; Player, M. R.;
Spurlino, J. C.; Springer, B. A. J. Biol. Chem. 2007, 282, 4094.
9. DeLano, W. L. The PyMOL Molecular Graphics System;
DeLano Scientific, Palo Alto, CA, USA, 2002. http://
www.pymol.org.
In conclusion, with the aid of structure-based drug de-
sign, the lead arylamide series, represented by 1a and
1b, was successfully optimized and improved address-
ing a potential IDR issue to generate novel, potent
FMS inhibitors, which did not have this liability.
The replacement substitutions required at C-2 resulted
in a loss of potency that was restored by incorporat-
ing the complementary hydrogen-bonding characteris-
tics of both 1a and 1b. The resulting imidazole
analogues confirmed the hypothesis and were approx-
imately 10-fold more potent than the corresponding
analogues containing either a furan or pyrrole. The
newly optimized series, exemplified by compounds
31a, 31b, and 31e, is equipotent to the lead series in
both enzyme and cell-based assays and is being exam-
ined as potential drug candidates.