Enhancement of kinase selectivity in a potent class of arylamide FMS inhibitors

Article abstract of DOI:10.1016/j.bmcl.2013.09.061

Structure-activity relationship (SAR) studies on a highly potent series of arylamide FMS inhibitors were carried out with the aim of improving FMS kinase selectivity, particularly over KIT. Potent compound 17r (FMS IC₅₀ 0.7 nM, FMS cell IC₅₀ 6.1 nM) was discovered that had good PK properties and a greater than fivefold improvement in selectivity for FMS over KIT kinase in a cellular assay relative to the previously reported clinical candidate 4. This improved selectivity was manifested in vivo by no observed decrease in circulating reticulocytes, a measure of bone safety, at the highest studied dose. Compound 17r was highly active in a mouse pharmacodynamic model and demonstrated disease-modifying effects in a dose-dependent manner in a strep cell wall-induced arthritis model of rheumatoid arthritis in rats.

Full text of DOI:10.1016/j.bmcl.2013.09.061

Bioorganic & Medicinal Chemistry Letters 23 (2013) 6363–6369
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Enhancement of kinase selectivity in a potent class of arylamide FMS
inhibitors
Carl R. Illig, Carl L. Manthey, Sanath K. Meegalla, Mark J. Wall, Jinsheng Chen, Kenneth J. Wilson,
Renee L. DesJarlais, Shelley K. Ballentine, Carsten Schubert, Carl S. Crysler, Yanmin Chen,
Christopher J. Molloy, Margery A. Chaikin, Robert R. Donatelli, Edward Yurkow, Zhao Zhou
⇑
Mark R. Player , Bruce E. Tomczuk
Janssen Pharmaceutical Research & Development LLC, Welsh & McKean Roads, Spring House, PA 19477, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 29 May 2013
Revised 20 September 2013
Accepted 23 September 2013
Available online 1 October 2013
Structure–activity relationship (SAR) studies on a highly potent series of arylamide FMS inhibitors were
carried out with the aim of improving FMS kinase selectivity, particularly over KIT. Potent compound 17r
(FMS IC50 0.7 nM, FMS cell IC50 6.1 nM) was discovered that had good PK properties and a greater than
fivefold improvement in selectivity for FMS over KIT kinase in a cellular assay relative to the previously
reported clinical candidate 4. This improved selectivity was manifested in vivo by no observed decrease
in circulating reticulocytes, a measure of bone safety, at the highest studied dose. Compound 17r was
highly active in a mouse pharmacodynamic model and demonstrated disease-modifying effects in a
dose-dependent manner in a strep cell wall-induced arthritis model of rheumatoid arthritis in rats.
Ó 2013 Elsevier Ltd. All rights reserved.
Keywords:
Colony-stimulating factor-1 receptor
FMS
Macrophage colony-stimulating factor
KIT
Hypocellularity
Anti-inflammatory
Rheumatoid arthritis
Colony-stimulating factor-1 (CSF-1) is a key regulator of the
myeloid lineage where it promotes the proliferation, differentia-
tion, and function of macrophages and macrophage progenitors.
CSF-1 acts on monocytes and macrophages by binding with high
affinity to its receptor CSF–1R, alternatively known as FMS, the
product of the cellular proto-oncogene homologue of the feline
McDonough sarcoma virus oncogene (v-fms).1 Although macro-
phages play a vital role in sustaining important biological func-
tions, they have also been implicated in several pathological
conditions such as rheumatoid arthritis (RA) and cancer.2 As mac-
rophages are known to play an important role in the inflammatory
process, the inhibition of macrophage proliferation by CSF-1 might
be of therapeutic value in the treatment of arthritis. CSF-1-defi-
cient mice are reported to be resistant to collagen-induced arthri-
tis. In a collagen-induced arthritis mouse model, CSF-1 was shown
to increase the severity of disease while a neutralizing anti-CSF-1
antibody had the opposite effect.2c A FMS inhibitor may also find
therapeutic value in any situation where macrophages have be-
come pathogenic.
recently.4 Publications from this laboratory5 have outlined the dis-
covery and structure–activity relationship (SAR) of a potent aryla-
mide chemotype5a which also stabilized crystals of an engineered
chimeric form of the FMS kinase domain5c allowing structure-
based optimization of future analogues. These efforts initially led
to potent proof-of-principle FMS kinase inhibitors5b which pos-
sessed good bioavailability in rodents and efficacy in mouse colla-
gen-induced arthritis. Removal of a potential metabolic liability for
idiosyncratic toxicity6 then led to compounds such as 1 and 2.5c
Despite their potency against FMS kinase and in a FMS cell-based
assay (murine bone marrow-derived macrophage, BMDM) prolifer-
ation assay,7 1 and 2 had poor efficacy in a pharmacodynamic as-
say (PD) in mice.5f,7 Unfavorable pharmacokinetic (PK) properties
were determined to be largely responsible for the disappointing
in vivo activity and became the focus of further studies. In addition,
more basic (Fig. 1) derivatives, such as 1 (calculated pKa 10.2),
showed significant off-target activity in a 50-assay counterscreen
for activity against a wide variety of receptors, transporters and
ion channels. Off-target promiscuity of compounds as a function
of their basicity is a well-studied phenomenon8 and, of greatest
concern, high levels of binding at ion channels, in particular
sodium and calcium ion channels, appeared to correlate with an
increased risk of adverse cardiovascular effects.5d Reducing the
basicity to mitigate off-target ion channel effects by replacing the
The development of FMS inhibitors has been reviewed3 and a
number of potent and selective compounds have entered the clinic
⇑
Corresponding author. Tel.: +1 215 628 7860; fax: +1 215 540 4160.
E-mail address: mplayer@its.jnj.com (M.R. Player).
0960-894X/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2013.09.061

6364
C. R. Illig et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6363–6369
FMS IC50 = 1.1 nM
BMDM IC50 = 7.2 nM
tion was again directed toward the substitution at the aryl 2-
1
R = H
and 4-positions. Modeling suggested small differences between
FMS and KIT near the 2-position that may provide an opportunity
to enhance selectivity (vide infra). The difference between KIT and
FMS near the 4-position on the aryl ring is less well-defined due to
the poorer resolution of the structure for FMS in that region. How-
ever, the proximity of that area to solvent also provides the great-
est flexibility in structural variation without impacting potency for
FMS and so was also targeted for investigation.
N
N
H
H
N
CN
CN
N
N
H
N
H
O
O
N
RN
O
S
O
FMS IC50 = 2.4 nM
R = Ac BMDM IC50 = 5.6 nM
3
FMS IC50 = 1.7 nM
BMDM IC50 = 3.6 nM
2
N
During initial SAR studies,5a only minimal optimization at the
aryl 4-position was conducted due to the early discovery of the
potency of ring-containing groups at that position. Therefore, a
re-examination of substituents at the aryl 4-position was carried
out, especially in light of the fact that the effects of 4-position sub-
stituents on kinase selectivity were not a consideration at that
time. Various groups, most of them polar to aid solubility, were
explored using a direct attachment to the aryl ring as shown by
compounds 10a–k in Table 1.
4
FMS IC50 = 0.69 nM
BMDM IC50 = 2.6 nM
Mo7e IC50 = 38 nM
N
H
CN
N
N
H
O
PD assay :
88-100% inhib @ 50 mg/kg
79% inhib @ 20 mg/kg
N
Me2N
O
Figure 1. Structures of recent generations of arylamide FMS inhibitors leading to
clinical compound 4.
In most cases, the chemistry to produce these analogues fol-
lowed one of two general routes. In the first (Scheme 1), the R1
group at the 4-position was present in the starting aniline 6 (or
nitrobenzene 5 where the nitro group was catalytically reduced
by hydrogen to 6). Bromination by N-bromosuccinimide (NBS) in
CH2Cl2 (DCM) to 7 was followed by a Suzuki–Miyaura reaction
with the appropriately substituted 1-cyclohexene boronic acid or
boronate ester. Amide coupling of the resulting aniline 8 with
the SEM-protected cyanoimidazole carboxylate 1112 afforded 9
which was subsequently deprotected with TFA to afford the final
products 10a–b, g–k. For 10g and 10h, 6 (R1 = SO2NH-t-Bu) was
used and the N-t-Bu group was later removed simultaneously with
the SEM group in the final deprotection step of 9 (R1 = (CH2)2SO2-
NHt-Bu) with TFA to give the primary sulfonamides. For the two
tetrazoles 10j and 10k, cyano groups in 8 (R1 = CN) were reacted
with Me3SnN3 (PhMe, reflux)13 and the resulting trimethylstan-
nyl–tetrazolyl intermediates were coupled with 11 (with concom-
itant loss of the stannyl group) to give amides 9j, k.
piperidine ring with mildly basic or non-basic heterocycles yielded
potent compounds such as 3 with reduced ion-channel binding but
which did not improve the PK properties.5d In the end, fine tuning
this series by careful optimization of derivatives off the piperidine
ring nitrogen afforded compounds with reduced basicity relative to
1 while retaining good solubility and ADME properties. Fully opti-
mized compound JNJ-28312141 (4) possessed good PK and PD
properties and excellent efficacy in rodent models of RA and was
advanced into clinical trials.5e
In contemplating the characteristics that might comprise a
backup compound, the kinase selectivity seemed to be one area
that could improve the overall safety profile of the series. Com-
pound 4 demonstrated fair to good selectivity against 111 tyrosine,
serine–threonine, or mixed kinases. However, 4 had only modest
selectivity for FMS versus four kinases (FLT3, AXL, KIT, and TRKA)
ranging from 7- to 43-fold.5e When further tested in cell-based as-
says designed to exhibit dependence on these kinases, only the
selectivity for KIT (16-fold) appeared to remain as a potential issue.
Although inhibition of KIT can suppress functions in dendritic cells
and mast cells suggesting possibly beneficial effects toward inflam-
mation, there is also evidence that potent inhibition of KIT may
produce anemia and bone marrow suppression,9 hypospermia,10
and loss of pigment.11 Indeed, during the course of in vivo studies,
reduced reticulocyte counts and bone marrow hypocellularity
were observed for 4 and a number of structurally related ana-
logues. Since this effect was potentially attributed to an unwanted
inhibition of KIT, improvement in the selectivity of FMS over KIT
became a focus for future efforts.
In an alternate route, also shown in Scheme 1, R1 was intro-
duced in the final step. A selective Suzuki–Miyaura reaction on
bromoiodoaniline 12 provided aniline 13 which was coupled to
protected cyanoimidazole carboxylate salt 11 as before to give 14
and then deprotected with TFA as before to afford common inter-
mediate 15. After 15 was doubly deprotonated with a Grignard
reagent, it underwent metal-halogen exchange with t-BuLi at low
Table 1
Kinase and cellular potencies for FMS of initial aryl 4-position analogues
R2 R2
Stem cell factor (SCF) is a cytokine which binds to its cell surface
receptor, CD117 (KIT), and is a growth factor important for the sur-
vival, proliferation, and differentiation of hematopoietic stem cells
and other hematopoietic progenitor cells. An assay based on stem
cell factor (SCF)-dependent proliferation of an Mo7e leukemia cell
line was used to determine inhibitory potential of compounds
against KIT in a cellular context. At an early stage of our program
we observed a good correlation between activity in this assay ver-
sus unwanted reductions in reticulocytes and bone marrow hypo-
cellularity. Therefore, cellular selectivity, that is, Mo7e IC50/BMDM
IC50, is reported throughout this Letter. Selectivity in biochemical
kinase assays were also determined for a small subset of com-
pounds where indicated. Because the cellular selectivity of 4
(Mo7e IC50/BMDM IC50) was 16, it was felt, albeit arbitrarily, that
a ratio of >60 should be targeted to mitigate potential for KIT-re-
lated effects.
CN
N
H
N
N
H
10a-k
O
R1
Compd
R1
R2
FMS IC50 (nM)
BMDMa IC50 (nM)
10a
10b
10c
10d
10e
10f
10g
10h
10i
CONH2
CONH2
CONMe2
COOH
CH(OH)Me
SO2Me
H
3.6
2.2
2.8
1.5
1.4
2.3
6.9
1.9
10
34
32
Me
Me
Me
Me
Me
H
130
>300
>150
59
>150
>150
210
nd
SO2NH2
SO2NH2
Me
H
SO2-Morpholin-1-yl
1H-Tetrazol-5-yl
1H-Tetrazol-5-yl
10j
10k
H
23
Me
3.7
>300
Due to the extensive optimization previously carried out on the
a
cyanoimidazole amide,5c,5e this moiety was left intact and atten-
Cellular FMS assay in murine bone marrow-derived macrophages (BMDM).

C. R. Illig et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6363–6369
6365
neously with the SEM group in the final deprotection step of 9 (R1 -
= CH2)2SO2NHt-Bu) with TFA to give the primary sulfonamide.
Primary amides 10x–y were derived from acid 6 (R1 = (CH2)2COOH)
by esterification with MeOH (cat H2SO4, reflux) followed by treat-
ment of the resulting ester 6 (R1 = (CH2)2CO2Me, H2SO4 salt) with
NH4OH (H2O, NaCl, rt) to give 6 (R1 = (CH2)2CONH2). Also, acid
10w was obtained by hydrolysis of amide 10x with 6 M KOH
(MeOH–THF, reflux). Synthesis of the two tetrazoles 10o–p was
carried out as before where cyano groups in 8 (R1 = CH2CN) were
reacted with Me3SnN313 and the resulting trimethylstannyl–tetraz-
olyl intermediates were coupled with 11 (with loss of the stannyl
group) to give amides 9o–p.
This exercise yielded only slightly better results (Table 2) in that
only two analogues (10u and 10v) met our cellular potency crite-
rion (<10 nM). However, enhancement in the cellular selectivity
of FMS over KIT relative to clinical compound 4 was observed for
six compounds providing selectivity ratios of 40–130 (vs 15 for 4).
An effect of 4,4-dimethyl substitution on the cyclohexene on
FMS potency was again seen with this set of analogues providing
a moderate to good boost in most cases as evidenced, for example,
by the comparison of 10o to 10p, 10u to 10v, 10bb to 10cc, or 10dd
to 10ee. Moreover, an increase in selectivity for FMS over KIT in the
cellular assay (MO7e IC50/BMDM IC50) was induced by this substi-
tution with the dimethyl analogues showing improvements over
the unsubstituted cyclohexenes of ca. two to threefold. This can
be seen by comparison of 10s to 10t, and 10u to 10v. Although
the selectivity was still not as high as desired in all cases, it repre-
sented a significant step forward.
In order to explore whether there might be a structural basis for
the observed selectivity improvement, compound 10v was mod-
eled into the FMS kinase and compared with KIT. Figure 2 shows
compound 10v in FMS (cyan) and KIT (green). There are only two
differences between FMS and KIT in the binding site region,
Val784 (FMS) to Ile798 (KIT) and Gly795 (FMS) to Cys809 (KIT).
Neither of these is predicted to be in direct contact with the inhib-
itor. The 4,4-dimethyl-cyclohexene ring extends down from the
hinge region toward the region that binds the ribose of ATP. The
phenyl ring of a Phe797 in FMS is positioned away from this region
whereas the corresponding ring in KIT (Phe811) is forced into that
space due to a sequence difference two residues before the Phe,
Gly795 in FMS substituted by Cys in KIT. This alternate Phe confor-
mation places the side chain in close proximity to the two added
methyl groups. The predicted steric clash in KIT is consistent with
lower potency in the Mo7e cell assay and the resulting selectivity
ratio (MO7e IC50/BMDM IC50) enhancement observed for FMS over
KIT of 1.8-fold for 10v (120) versus 10u (67) in the cellular assay.
Despite this observation, the metabolic stability of many com-
pounds in this set, based on values of <70% of the compound
remaining after a 10-min exposure to liver microsome prepara-
tions, was poor in one or more species (human/mouse/rat) but par-
ticularly of concern for mouse which was used for the PD assay.
Based on the hypothesis that oxidative metabolism at the benzylic
methylene of these 4-position groups could be a contributing
factor toward the low stability,15 dimethyl substitution on the ben-
zylic methylene to block metabolism16 was investigated and the
resulting derivatives are shown in Table 3. With improved selectiv-
ity over KIT from incorporation of the gem dimethyl substitution
on the cyclohexene confirmed, that substitution was used exclu-
sively in the subsequent group of compounds.
R2 R2
Br
R3
NH2
b
c
R1
R1
7
5
6
R
R
3 = NO2
3 = NH2
NH2
a
R2 R2
R1
8
CN
CN
N
N
9
X = SEM
e
KO
H
N
N
SEM
N
X
11
O
10a-b, g-k
X = H
O
R1
d
R2 R2
R2 R2
I
CN
NH2
c
d
N
H
N
NH2
N
X
Br
12
13
O
Br
Br
14
X = SEM
e
f
10c-f
10 l-gg
15 X = H
Scheme 1. General synthetic routes to analogues. Reagents and conditions; (a) H2,
10% Pd(C), EtOAc, 6 h. (b) NBS, DCM or DCM–MeCN, 0 °C to rt, 1 h. (c) 1-
Cyclohexene boronic acid or 4,4-dimethyl-1-cyclohexene pinicolboronate, Pd(Ph3-
P)4, 2 M aq Na2CO3, dioxane, 80 °C, 5 h. (d) N-SEM-cyano-imidazolecarboxylic acid
potassium salt (11), PyBrop, DIEA, DCM, 3–16 h. (e) TFA, 30% EtOH in DCM. (f) (i) i-
PrMgCl (2 equiv), THF, 0 °C, 45 min, (ii) t-BuLi (2.3 equiv), À78 °C, 10 min, various
electrophiles (10c: MeNHCON(Me)OMe, 10d: EtO2CCN, 10e: CH3CHO, 10f: (MeS)2).
temperature and then was reacted with a variety of electrophiles
to give the desired final products 10c–f. It should be noted that this
‘trianion’ approach to protect the nitrile was necessary since simi-
lar attempts on SEM-protected 14 were unsuccessful. In the case of
10d (R1 = COOH), intermediate ester 10 (R1 = CO2Et) was hydro-
lyzed to the corresponding acid with 2 M KOH (EtOH, 60 °C, then
2 M aq TFA), and for 10f, intermediate 10 (R1 = SMe) was oxidized
with mCPBA (DCM, rt) to afford the methylsulfone.
As shown in Table 1, the majority of these initial 4-position ana-
logues (10a–k) were quite potent for FMS (<10 nM). In past stud-
ies,5d substitution with one or two methyl groups in the
4-position of the cyclohexene had seemed, in select cases, to im-
part a marginal increase in FMS potency, but the decrease in solu-
bility from the added lipophilicity had not always warranted its
use. However, a beneficial effect of this 4,4-dimethyl substitution
on FMS potency was quite evident in these compounds providing
a boost of about two to sixfold (comparing 10a with 10b, 10g with
10h, and 10j with 10k). Unfortunately sufficient potency in the cel-
lular BMDM proliferation assay was lacking in every case where
IC50 values below 10 nM are desired based on experience that com-
pounds with higher values often did not exhibit pharmacodynamic
activity at reasonable (<50 mg/kg) doses. The higher IC50 values for
the cell data also did not permit a useful assessment of KIT selec-
tivity for the majority of these compounds.
Although the reason for uniformly lower cellular potency in
these compounds was not clear, extension of the polar functional
groups out from the ring by a spacer of one or two carbons was also
evaluated. These compounds (10l–gg) were all produced by the
first of the two routes described above (Scheme 1). Non-commer-
cially available precursors were prepared as follows. For 10m, the
corresponding
5
(R1 = CH2-4-(2-hydroxyethyl)-piperazin-1-yl))
was prepared by treating benzyl bromide 5 (R1 = CH2Br) with 2-
(piperazin-1-yl)ethanol in EtOH (0 °C to rt). For 10r, the corre-
sponding 5 (R1 = CH2CONH(CH2)3)OH) was prepared by treating
acid chloride 5 (R1 = CH2COCl) with 3-aminopropan-1-ol (DCM,
0 °C to rt). The sulfonamides 10z–gg were prepared starting with
sulfonyl chloride 5 (R1 = (CH2)2SO2Cl)14 and reacting with the
appropriate amines in THF. In the case of 10z and 10aa, the amine
was t-BuNH2 and the N-t-Bu group was later removed simulta-
The methodology to produce these tertiary benzylic analogues
is illustrated in Scheme 2. The tertiary benzyl alcohol 16, resulting
from addition of acetone to the previously employed aryllithium
intermediate, derived from bromide 15 and t-BuLi, was activated
by treatment with SOCl2 or TFA and the resulting benzylic carbo-
cation captured with a variety of hydroxy and amino nucleophiles
(R4H) to give the corresponding tertiary ethers and amines 17a–s.

6366
C. R. Illig et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6363–6369
Table 2
Kinase and cellular potencies for FMS and KIT and metabolic stability of extended aryl 4-position analogues
R2
R2
CN
N
H
N
N
H
10l-gg (R1 = R3(CH2)n)
R3
O
( )n
Compd
n
R3
R2
H
FMS/KIT IC50
(nM)
BMDMa IC50
(nM)
Mo7eb IC50 (nM) FMS/KIT Cell selc IC50 (nM) Met stab (H/M/R)d (% rem)
4
10l
10m
0.7
3.2
2.6
20
13
41
16
45
41
87/81/77
71/5/nd
101/79/2
1
1
Morpholin-1-yl
4-(HO(CH2)2)-Piperazin-1-
yl
890
530
Me 0.8
10n
10o
10p
10q
10r
10s
10t
10u
10v
10w
10x
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
Morpholin-1-yl-SO2
1H-Tetrazol-5-yl
1H-Tetrazol-5-yl
C(Me2)OH
HO(CH2)3NHCO
HO
H
H
13
13
48
nd
nd
nd
nd
nd
nd
nd
nd
nd
40
130
67
120
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
66/61/47
nd
7/3/3
74/41/nd
46/2/nd
66/49/nd
83/nd/19
nd
nd
nd
109/nd/5
nd
nd
nd
nd
118/93/4
nd
>300
>100
32
>100
40
18
1.8
2.6
>300
>50
25
54
86
9.423
2.921
39
32
13
Me 2.0
3.7
Me 2.4
6.8
Me 4.6
1.4
Me 0.8
H
nd
H
1600
2400
120
300
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
HO
Morpholin-1-yl
Morpholin-1-yl
HOOC
NH2CO
NH2CO
NH2SO2
NH2SO2
NH2NHSO2
NH2NHSO2
NH2NSO2
H
H
H
5.9
2.7
10y
10z
Me 1.7
2.4
Me 1.6
5.6
Me 2.6
9.4
Me 2.9
3.5
Me 4.5
H
10aa
10bb
10cc
10dd
10ee
10ff
10gg
H
H
NH2NSO2
Morpholin-1-yl-SO2
Morpholin-1-yl-SO2
H
60
a
b
c
Cellular FMS assay in murine bone marrow-derived macrophages (BMDM).
Cellular KIT assay in Mo7e cells.
Cellular selectivity ratio favoring FMS over KIT (MO7e IC50/BMDM IC50).
d
In vitro metabolic stability assay of compounds in the presence of liver microsomal preparations from human,mouse, and rat (H/M/R) as percent of the compound
remaining (% rem) after 10 min.
Figure 2. Model of compound 10v (grey) in the binding site of FMS (cyan) and KIT (green). The left image illustrates the close contact of Phe811 of KIT, which is protruding
through the FMS surface (cyan), with the 4,4-dimethylcyclohexene of compound 10v (grey). The figure shows the two proteins only: FMS in cyan and KIT in green. The two
FMS binding site residues that differ with KIT are labeled, Val784 (KIT Ile798) and Gly795 (KIT Cys809), along with Phe 797 (KIT Phe811). The Gly to Cys difference forces the
conserved Phe into a conformation that clashes with the dimethyl of the modeled inhibitor.
For the one-carbon extended analogues 20a–g, ester 1817 was
reduced (LiAlH4) and the resulting alcohol protected as the t-butyl-
dimethylsilyl (TBS) ether to give 19. After reduction of the nitro
group, ortho-bromination of the aniline, and Suzuki–Miyaura
reaction to install the 4,4-dimethyl-1-cyclohexene as before, both
the TBS and SEM protecting groups were removed with fluoride
to give 20a. Oxidation of 20a with Dess–Martin reagent and reduc-
tive amination of the resulting aldehyde 21 with the appropriate
amines afforded the final products in most cases. For sulfones
17h (R4 = MeSO2(CH2)2NH), 17n (R4 = thiomorpholin-1-yl), 20d
(R4 = thiomorpholin-1-yl-CH2), and 20f (R4 = MeSO2(CH2)2NHCH2),
the intermediate sulfides 20 (R4 = MeSO2(CH2)2NH, thiomorpholin-

C. R. Illig et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6363–6369
6367
Table 3
Potency, kinase cell selectivity, and metabolic stability of benzylic dimethyl substituted aryl 4-position analogues
CN
N
H
N
N
H
R4
O
17a-s, 20a-g
Compd R4
FMS/KIT IC50 (nM) BMDMa IC50 (nM) Mo7eb IC50 (nM) FMS/KIT Cell selc IC50 (nM) Met stab (H/M/R)d (% rem)
4
0.7/5
1.9
2.1
1.2/25
1.9/57
nd
0.6/42
2.1
1.5
7.2
1.2/92
0.7
19
0.9
3.5/250
2.3
1.5/50
1.7/40
0.7
1.3/210
2.0
2.6
25
1.8
13
38
15
87/81/77
110/124/96
104/nd/39
nde
38/93/44
nd
79/50/78
11/nd/13
93/62/28
nd
82/62/90
nd
17a
17b
17c
17d
17e
17f
17g
17h
17i
HO
2500
250
380
1200
250
520
nd
1900
nd
400
nd
4000
3000
3300
510
600
nd
13
6.1
1900
nd
1800
100
310
720
360
100
140
29
52
32
83
nd
61
nd
Me2N(CH2)2O
Pyrrolidin-1-yl-(CH2)2O
Morpholin-1-yl-(CH2)2O
HO(CH2)2NH
MeO(CH2)2NH
MeO(CH2)2NMe
23
7.8
6.3
9.8
31
21
5.8
68
>200
17
8.3
5.1
6.0
13
6.1
7.7
33
MeO2S(CH2)2NH
MeO2S(CH2)2NMe
Morpholin-1-yl-(CH2)2NH
AcNH(CH2)2NH
Pyridin-2-yl-NH
Morpholin-1-yl
S,S-Dioxothiomorpholin-1-yl
4-Me-piperazin-1-yl
4-Et-piperazin-1-yl
4-Ac-piperazin-1-yl
4-(HO(CH2)2)-Piperazin-1-yl
4-(MeO(CH2)2)-Piperazin-1-yl
HOCH2
17j
17k
17l
69
nd
<20
180
400
100
100
nd
84
97
58
nd
nd
nd
17m
17n
17o
17p
17q
17r
17s
20a
20b
20c
20d
20e
20f
112/84/98
103/110/16
108/84/32
ndf
110/98/63
72/73/40
nd
nd
nd
nd
35/nd/39
nd
Morpholin-1-yl-(CH2)
4-Ac-piperazin-1-yl-CH2
8.0
5.7
56
29
19
13
18
17
62
53
24
40
S,S-Dioxothiomorpholin-1-yl-CH2 3.5
MeO(CH2)2NHCH2
MeO2S(CH2)2NHCH2
4-(HO(CH2)2)-Piperazin-1-yl-CH2
0.6/110
1.5
2.0
20g
21
nd
a
b
c
Cellular FMS assay in murine bone marrow-derived macrophages (BMDM).
Cellular KIT assay in Mo7e cells.
Cellular selectivity ratio favoring FMS over KIT (MO7e IC50/BMDM IC50).
d
In vitro metabolic stability assay of compounds in the presence of liver microsomal preparations from human, mouse, and rat (H/M/R), percent compound remaining after
10 min.
e
Nd: not determined.
Compound was too insoluble to assay.
f
1-yl, thiomorpholin-1-yl-CH2, and MeS(CH2)2NH-CH2, respec-
tively), were oxidized to the corresponding sulfones with H2O2,
(Ti(Oi-Pr)4, DCM-iPrOH (40:1), 0 °C, 30 min; then rt, 2 h). Methyla-
tion of 17h (MeI, NaHCO3, THF, rt, 6 h) provided the N-methyl ana-
logue 17i.
As before, polar groups and lower basicity amines were incorpo-
rated to aid solubility. It is interesting to note that when this dimethyl
substitution to block metabolism was applied to the benzylic meth-
ylene of 4-substituted compounds linked by a single carbon to the
heteroatom of a functional group (17a–s), numerous compounds
were potent inhibitors, both against FMS and in the cellular BMDM
assay (Table 3). Yet surprisingly, when the same dimethyl substitu-
tion was applied to the benzylic methylene of compounds linked
by two carbons to the heteroatom of a functional group (20a–g),
the resultant compounds lacked adequate cellular activity.
CN
b
a
N
H
15
R4 = OH
R4 = Nu
N
16
17a-s
N
H
R4
O
CN
NO2
d
N
H
R4
18
N
20a
21
N
R
R
4 = CH2OH
4 = CHO
e
f
H
R4
O
R
R
4 = CO2Me
20b-g
c
4 = CH2OTBS
19
Scheme 2. General synthetic routes to 4,4-dimethylbenzylic analogues. Reagents
and conditions; (a) (i) i-PrMgCl (1.1 equiv), THF, À78 °C to rt (5 min); (ii) t-BuLi
(3.0 equiv), À78 ° C, 10 min; (iii) acetone, À78 °C to rt, 30 min. (b) SOCl2
(0.52 equiv), DCM, 0 °C to rt, 1 h, then various nucleophiles R4H (see text), 0 °C,
1 h; or, R4H (20 equiv), TFA (30 equiv), DCM, 60 °C, 6 h. (c) (i) LiAlH4, THF, 0 °C, 3 h;
(ii) TBSCl, imidazole, DCM, rt, 16 h. (d) (i) H2, 10% Pd(C), EtOAc, 6 h; (ii) NBS, DCM,
0 °C to rt, 2 h; (iii) 4,4-dimethyl-1-cyclohexene pinicolboronate, Pd(Ph3P)4, 2 M aq
Na2CO3, dioxane, 80 °C, 5 h; (iv) 11, PyBrop, DIEA, DCM, 16 h; (v) (n-Bu)4NF, THF,
50 °C, 3 h. (e) Dess–Martin periodinane, NaHCO3, DCM, 0 °C, 30 min then rt, 2 h. (f)
NaBH(OAc)3, various amines R4H, 1,2-dichloroethane, rt, 3 h. (g) Additional steps, as
needed (see text).
More importantly, when the 10 compounds within 17a–s which
met the cellular potency criterion of IC50 < 10 nM in the BMDM as-
say were tested in the Mo7e assay, a high level of cellular selectiv-
ity for FMS over KIT was obtained for the majority of them (17b, f, j,
n–p, r–s). Even when the cellular potency criterion was not met,
selectivity ratios (MO7e IC50/BMDM IC50) favoring FMS of greater
than 50 were observed in most cases (e.g., 17a, h, and m). This
represented a range of cellular kinase selectivity enhancements
relative to the comparable non-benzylic substituted analogues of

6368
C. R. Illig et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6363–6369
from twofold (17r (84) vs 10m (41)) up to fourfold (17m (180) vs
10l (45)). Thus, the combination of 4,4-dimethyl substitution on
the 1-cyclohexene together with benzylic dimethyl substitution
on the 4-aryl substituent appears to be very favorable for enhanc-
ing cellular selectivity for FMS over KIT.
IC50 values were determined for a subset of compounds in a KIT
kinase assay (Table 3) and, relative to 4, improvements in FMS/KIT
selectivity were confirmed.
An examination of the metabolic stability in liver microsomes
showed (Table 3) that stability appeared to be generally improved.
Of the 10 compounds with reasonable potency in the BMDM cell
assay, 17n appeared stable enough in all three species to further
test in the PD assay in mice. In this assay, the ability of an iv injec-
tion of CSF-1 to induce c-fos mRNA in spleen is measured 6 or 8 h
after oral administration of compound (Table 4). Two other com-
pounds (17j and 17r), while not quite reaching the desired level
of stability in at least one species were also tested as were 17o
and 20e which were relatively potent in cells but had poor stability
in one or more species. Although not all of these five candidates
were tested at directly comparable doses or time points relative
to each other or to the clinical candidate 4, clearly only 17r
appeared to possess sufficient in vivo PD activity and reasonable
stability to warrant further evaluation.
The pharmacokinetic behavior of 17r as its HCl salt was then as-
sessed in rat. Following a 1 mg/kg iv dose in male Sprague–Dawley
rats (250–300 g), the compound, formulated in 20% 2-hydroxypro-
pyl-b-cyclodextrin (HPbCD), showed a high volume of distribution
(7.6 L/kg) compared to total body water (rat TBW = 0.668 L/kg) and
moderate systemic clearance (34 ml/min/kg) compared to hepatic
blood flow (rat HBF = 55 ml/min/kg). Plasma concentrations after
a 5 mg/kg oral dose reached mean Cmax values of 88 ng/ml
Figure 3. Compound 17r blocks the progressive increase in ankle and paw swelling
during the chronic phase of SCW-induced arthritis.
an acute non-erosive phase (days 3–7) that is complement and
neutrophil dependent and which resolves, and a subsequent
chronic erosive phase (day >10) that is dependent on the develop-
ment of specific T cell immunity to SCW. The chronic phase is char-
acterized by involvement of macrophages in the inflamed synovial
pannus that models the involvement of macrophages in RA. Female
Lewis rats (80–100 gm each) on day 0, were administered an ip
dose of SCW material. On day 5, rats with a distinct acute phase ar-
thritic response based on joint swelling were randomized (n = 7
per group) into three groups. Oral BID administration of vehicle
only, or 10 mg/kg, or 40 mg/kg 17r was commenced on day 10 at
the start of the chronic phase and continued for 14 days. As shown
in Figure 3, 17r inhibited the progression of ankle swelling at
40 mg/kg and partly at 10 mg/kg.
In summary, information on the structural differences between
FMS and KIT in the ribose-binding region of the ATP pocket, to-
gether with a systematic survey of functionality at the 4-position
of the aryl core, provided a series of potent arylamide FMS inhibi-
tors with significantly enhanced selectivity for FMS over KIT. For
selected compound 17r, this correlated with in vivo observations
that were consistent with the decreased KIT activity. Compound
17r demonstrated good efficacy in an animal model of RA, a nearly
sixfold enhancement in selectivity for FMS over KIT relative to clin-
ical compound 4 and, therefore, a potentially improved safety
profile.
(0.21 lM) at 2.8 h and the terminal elimination t1/2 was 6.5 h.
The compound showed good systemic exposure after oral adminis-
tration (F = 31%, AUC = 690 (ng/ml) ⁄ h and therefore had proper-
ties adequate for further in vivo characterization.
The cellular selectivity ratio for 17r of 84 represents a 5.3-fold
enhancement in selectivity over KIT relative to the ratio of 16 for
clinical candidate 4. In order to confirm that this improvement in
the cellular assay was reflected physiologically, the tolerability of
17r was determined in rat at pharmacologically relevant doses. Fe-
male Sprague–Dawley rats (150–200 g body weight) were admin-
istered vehicle (n = 4) or 17r at 10–40 mg/kg (n = 3) po bid for 4
days. Compound was delivered as a solution at 6 ml/kg in 20%
HPbCD. On the fifth day, a final dose was given and the animals
were sacrificed 2 h later. Mean plasma levels of 17r two hours fol-
lowing the last dose of 10–40 mg/kg were 280 ng/ml (0.68
lM) and
2113 ng/ml (5.16 M), respectively. The compound was well toler-
l
ated at the highest dose tested, 40 mg/kg bid  4.5 days. At this
dose, rats did not lose body weight relative to the vehicle control
animals and blood reticulocyte levels were not reduced. The latter
observation differentiates 17r from clinical compound 4 that
caused a mild reduction in reticulocytes at doses estimated to have
equivalent pharmacodynamic activity.
Compound 17r was next evaluated in a streptococcal cell wall
References and notes
(SCW)-induced arthritis in rats.18 This model is distinguished by
1. Galland, F.; Stefanova, M.; Lafage, M.; Birnbaum, D. Cytogenet. Cell Genet. 1992,
60, 114.
2. (a) Toh, M.; Bonnefoy, J.-Y.; Accart, N.; Cochin, S.; Zemmour, C.; Haegel, H., et al
Arthritis Rheum. 2011, 63(Suppl 10), 1140; (b) Hume, D. A.; MacDonald, K. P.
Blood 2012, 119, 1810; (c) Campbell, I. K.; Rich, M. J.; Bischoff, R. J.; Hamilton, J.
A. J. Leukocyte Biol. 2000, 68, 144.
3. Patel, S.; Player, M. R. Curr. Top. Med. Chem. 2009, 9, 599.
4. (a) Scott, D. A.; Bell, K. J.; Campbell, C. T.; Cook, D. J.; Dakin, L. A.; Del Valle, D. J.;
Drew, L.; Gero, T. W.; Hattersley, M. M.; Omar, C. A.; Tyurin, B.; Zheng, X. Bioorg.
Med. Chem. Lett. 2009, 19, 701; (b) Plexxikon, Inc. press release; 2010. Available
from: www.plexxikon.com/view.cfm/73/PressReleases. (Accessed May 2,
2013.).; (c) Array Biopharma, Inc. presentation. Available from:
www.arraybiopharma.com/_documents./Publication/PubAttachment461.pdf.
(Accessed May 2, 2013.).
Table 4
Pharmacodynamic efficacy of selected benzylic dimethyl substituted aryl 4-position
analogues
Parameter
Compound
4
17j
17n
17o
17r
20e
Dose (mg/kg)
Time (h)
20
8
79
30
6
14
30
6
40
6
40
6
78
20
6
0
Inhib of c-fosa (%)
8
0
a
Percent inhibition of CSF-1-induced c-fos gene expression in mouse spleen.

C. R. Illig et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6363–6369
6369
5. (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, 17, 6070; (b) Illig, C. R.; Chen, J.; Wall, M. J.; Wilson, K. J.;
Ballentine, S. K.; Rudolph, M. J.; DesJarlais, R. L.; Chen, Y.; Schubert, C.;
Petrounia, I.; 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;
(c) Meegalla, Sanath K.; Wall, Mark J.; Chen, J.; Wilson, K. J.; Ballentine, S. K.;
DesJarlais, R. L.; Schubert, C.; Crysler, C. S.; Chen, Y.; Molloy, C. J.; Chaikin, M.
A.; Manthey, C. L.; Player, M. R.; Tomczuk, B. E.; Illig, C. R. Bioorg. Med. Chem.
Lett. 2008, 18, 3632; (d) Wilson, K. J.; Illig, C. R.; Chen, J.; Wall, M. J.;
Ballentine, S. K.; DesJarlais, R. L.; Chen, Y.; Schubert, C.; Donatelli, R.;
Petrounia, I.; 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. 2010, 20, 3925;
(e) Illig, C. R.; Manthey, C. L.; Wall, M. J.; Meegalla, S. K.; Chen, J.; Wilson, K. J.;
Ballentine, S. K.; DesJarlais, R. L.; Schubert, C.; Crysler, C. S.; Chen, Y.; Molloy,
C. J.; Chaikin, M. A.; Donatelli, R.; Yurkow, E.; Zhou, Z.; Player, M. R.; Tomczuk,
B. E. J. Med. Chem. 2011, 54, 7860; (f) Manthey, C. L.; Johnson, D. L.; Illig, C. R.;
Tuman, R. W.; Zhou, Z.; Baker, J. F.; Chaikin, M. A.; Donatelli, R.; Franks, C. F.;
Zeng, L.; Crysler, C. S.; Chen, Y.; Yurkow, E.; Boczon, L.; Meegalla, S. K.;
Wilson, K. J.; Wall, M. J.; Chen, J.; Ballentine, S. K.; Ott, H.; Baumann, C.;
Lawrence, D.; Schubert, C.; Tomczuk, B. E.; Molloy, C. J. Mol. Cancer Ther. 2009,
8, 3151.
7. Huang, H.; Hutta, D. A.; Rinker, J. M.; Hu, H.; Parsons, W. H.; Schubert, C.;
DesJarlais, R. L.; Crysler, C. S.; Chaikin, M. A.; Donatelli, R. R.; Chen, Y.; Cheng,
D.; Zhou, Z.; Yurkow, E.; Manthey, C. L.; Player, M. R. J. Med. Chem. 2009, 52,
1081.
8. (a) Peters, J.-U.; Schnider, P.; Mattei, P.; Kansy, M. ChemMedChem 2009, 4, 680;
(b) Azzaoui, K.; Hamon, J.; Faller, B.; Whitebread, S.; Jacoby, E.; Bender, A.;
Jenkins, J. J.; Urban, L. ChemMedChem 2007, 2, 874.
9. Munugalavadla, V.; Kapur, R. Crit. Rev. Oncol./Hematol. 2005, 54, 63.
10. Mauduit, C.; Hamamah, S.; Benahmed, M. Hum. Reprod. Update 1999, 5, 535.
11. Moss, K. G.; Toner, R. G. C.; Cherrington, J. M.; Mendel, D. B.; Laird, A. D. J.
Pharmacol. Exp. Ther. 2003, 307, 476.
12. Wall, M. J.; Schubert, C.; Illig, C. R. Synthesis 2008, 3377.
13. Carini, D. J.; Duncia, J. J.; Wong, P. C. B.;, R.; Layek, M. M.; Bhar, C. (E. I. Du Pont
de Nemours and Co.). U. S. Patent 5138069, 1992.
14. Islam, A.; Sahadev, K.; Reddy, M. V.; Kulkarni, R.; Layek, M. M.; Bhar, C. (Dr.
Reddy’s Laboratories Ltd.). PCT patent applic. WO 2006010079, 2006.
15. Nassar, A.-E. F.; Kamel, A. M.; Clarimont, C. Drug Discovery Today 2004, 9, 1020.
16. Cote, B.; Frenette, R.; Prescott, S.; Blouin, M.; Brideau, C.; Ducharme, Y.; Friesen,
R. W.; Laliberte, F.; Masson, P.; Styhler, A.; Girard, Y. Bioorg. Med. Chem. Lett.
2003, 13, 741.
17. Liu, X.; Hartwig, J. F. J. Am. Chem. Soc. 2004, 126, 5182.
18. Cromartie, W. J.; Craddock, J. G.; Schwab, J. H.; Anderlie, S. K.; Yang, C. J. Exp.
Med. 1977, 146, 1585.
6. Ju, C.; Uetrecht, J. P. Curr. Drug Metab. 2002, 3, 367.