Design, synthesis and optimization of bis-amide derivatives as CSF1R inhibitors

Article abstract of DOI:10.1016/j.bmcl.2017.03.064

Signaling via the receptor tyrosine kinase CSF1R is thought to play an important role in recruitment and differentiation of tumor-associated macrophages (TAMs). TAMs play pro-tumorigenic roles, including the suppression of anti-tumor immune response, promotion of angiogenesis and tumor cell metastasis. Because of the role of this signaling pathway in the tumor microenvironment, several small molecule CSF1R kinase inhibitors are undergoing clinical evaluation for cancer therapy, either as a single agent or in combination with other cancer therapies, including immune checkpoint inhibitors. Herein we describe our lead optimization effort that resulted in the identification of a potent, cellular active and orally bioavailable bis-amide CSF1R inhibitor. Docking and biochemical analysis allowed the removal of a metabolically labile and poorly permeable methyl piperazine group from an early lead compound. Optimization led to improved metabolic stability and Caco2 permeability, which in turn resulted in good oral bioavailability in mice.

Full text of DOI:10.1016/j.bmcl.2017.03.064

Accepted Manuscript
Design, synthesis and optimization of bis-amide derivatives as CSF1R inhibitors
Sreekanth A. Ramachandran, Pradeep S. Jadhavar, Sandeep K. Miglani,
Manvendra P. Singh, Deepak P. Kalane, Anil K. Agarwal, Balaji D. Sathe,
Kakoli Mukherjee, Ashu Gupta, Srijan Haldar, Mohd Raja, Siddhartha Singh,
Son M. Pham, Sarvajit Chakravarty, Kevin Quinn, Sebastian Belmar, Ivan E.
Alfaro, Christopher Higgs, Sebastian Bernales, Francisco J. Herrera, Roopa Rai
PII:
S0960-894X(17)30314-1
DOI:
http://dx.doi.org/10.1016/j.bmcl.2017.03.064
BMCL 24816
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
31 December 2016
18 March 2017
22 March 2017
Please cite this article as: Ramachandran, S.A., Jadhavar, P.S., Miglani, S.K., Singh, M.P., Kalane, D.P., Agarwal,
A.K., Sathe, B.D., Mukherjee, K., Gupta, A., Haldar, S., Raja, M., Singh, S., Pham, S.M., Chakravarty, S., Quinn,
K., Belmar, S., Alfaro, I.E., Higgs, C., Bernales, S., Herrera, F.J., Rai, R., Design, synthesis and optimization of
bis-amide derivatives as CSF1R inhibitors, Bioorganic & Medicinal Chemistry Letters (2017), doi: http://dx.doi.org/
10.1016/j.bmcl.2017.03.064
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Design, synthesis and optimization of bis-amide derivatives as CSF1R
inhibitors
Sreekanth A. Ramachandrana, Pradeep S. Jadhavara, Sandeep K. Miglania, Manvendra P. Singha,
Deepak P. Kalanea, Anil K. Agarwala, Balaji D. Sathea, Kakoli Mukherjeea, Ashu Guptaa, Srijan
Haldara, Mohd Rajaa, Siddhartha Singha, Son M. Phamb, Sarvajit Chakravartyb, Kevin Quinnb,
Sebastian Belmarc, Ivan E. Alfaroc, Christopher Higgsd, Sebastian Bernalesb, Francisco J.
Herrerab and Roopa Raib*
aIntegral BioSciences Pvt. Ltd, C-64, Hosiery Complex Phase II Extension, Noida, Uttar Pradesh 201306, India
bMedivation, now Pfizer, 525 Market Street, 36th Floor, San Francisco, CA 94105, USA
cFundación Ciencia y Vida, Avenida Zañartu 1482, Ñuñoa, Santiago, Chile 778027
dSchrödinger, Inc., 120 West 45th Street, 17th Floor, New York, New York 10036
ABSTRACT
Signaling via the receptor tyrosine kinase CSF1R is thought to play an important role in recruitment and differentiation of tumor-
associated macrophages (TAMs). TAMs play pro-tumorigenic roles, including the suppression of anti-tumor immune response,
promotion of angiogenesis and tumor cell metastasis. Because of the role of this signaling pathway in the tumor microenvironment,
several small molecule CSF1R kinase inhibitors are undergoing clinical evaluation for cancer therapy, either as a single agent or in
combination with other cancer therapies, including immune checkpoint inhibitors. Herein we describe our lead optimization effort that
resulted in the identification of a potent, cellular active and orally bioavailable bis-amide CSF1R inhibitor. Docking and biochemical
analysis allowed the removal of a metabolically labile and poorly permeable methyl piperazine group from an early lead compound.
Optimization led to improved metabolic stability and Caco2 permeability, which in turn resulted in good oral bioavailability in mice.

CSF1R (also referred to as FMS) is the receptor for the colony stimulating factor (CSF1) which regulates the survival
and differentiation of macrophages.1, 2 This receptor tyrosine kinase is expressed in several tumor types where it regulates
tumor associated macrophages (TAMs). TAMs play pro-tumorigenic roles in the tumor microenvironment (TME) by
stimulating angiogenesis, promoting tumor cell invasion and inducing an immunosuppressive environment.2-4. Given the
key roles which TAMs plays in modulating the TME, altering their activity by blocking CSF1R signaling has been
indicated as a possible anti-cancer strategy, particularly in combination with immune checkpoint therapies.4-6
A survey of the literature revealed five orally active small molecule CSF1R inhibitors currently under clinical
development. These are shown in Figure 1. PLX3397, an inhibitor of CSF1R and c-Kit, is the most advanced compound,
currently undergoing Phase 3 clinical studies in patients with tenosynovial giant cell tumor (TGCT).7 BLZ945 is a
selective CSF1R kinase inhibitor currently undergoing Phase 1/2 study either as a single agent or in combination with anti-
PD-1 antibody in advanced solid tumors.5 PLX7486 (structure not disclosed), a CSF1R and Trk inhibitor, has also
advanced to Phase 1 trial in patients with advanced solid tumors.8 ARRY-382 (structure not disclosed), is also undergoing
Phase 1/2 dose escalation trial in combination with the anti-PD-1 antibody pembrolizumab (Keytruda®), in patients with
advanced solid tumors.9 JNJ-40346527, a selective inhibitor of CSF1R was well tolerated, and preliminary antitumor
results from Phase 1/2 suggested limited activity in monotherapy for the treatment of relapsed or refractory classical
Hodgkin lymphoma (cHL).10 Antibodies against CSF1 and CSF1R have also been developed, some of which are
undergoing clinical trials.4
Figure 1. Reported CSF1R inhibitors in clinical development.
Our search began with bis-amide 1 (Table 1), a compound reported in the literature to be a mixed kinase inhibitor with
activity against both c-Kit and CSF1R. c-Kit is a closely related homolog of CSF1R and plays key roles in several
physiological processes including the regulation of hematopoietic stem cells.11 To avoid potential effects of c-Kit
inhibition on hematopoiesis, we targeted compounds that would preferentially inhibit CSF1R over c-Kit. Thus, our
screening cascade involved the biochemical potency measurement of newly synthesized compounds against CSF1R and c-
Kit. The enzymatic activity was determined in vitro using the indicated enzymes incubated with [ϒ-33P]-ATP and peptide
substrate. IC50 curves were determined for each compound (see supplemental information). These biochemical assays
were followed by testing our compounds in cell viability assays using mouse bone marrow-derived macrophages (BMM),
a primary cell culture known to require CSF1 for viability.12 Viability assays using the RAW 264.7 cell line were included
as an indication for general off-target activity since this cell line does not require CSF1 for its viability.
With the aim of identifying a CSF1R inhibitor with improved selectivity over c-Kit, our search began with the
synthesis of analogs of the bis-amide 1.13 With the corresponding phenyl core replacing the pyridine (2), we made a series
of compounds shown in Table 1. In order to understand the importance of the trifluoromethyl group at R3, we synthesized
compound 3 (R3 = H) which was inactive towards CSF1R kinase when tested up to 10 µM concentration. Compound 4,
where a chlorine replaces the trifluoromethyl group resulted in moderate potency towards CSF1R. Similarly, the role of
R2 methyl group was explored by replacing it with chlorine (compound 5) wherein the CSF1R potency was retained,
indicating that the potency was not impacted by varying the electronics of the phenyl ring. The original 2-thiophene
substituent at R1, off the phenyl group, was replaced with a pyridine (compound 6) and pyrimidine (compound 7); this
resulted in a 4-8 fold CSF1R potency improvement. This biochemical potency improvement also translated to ~60%
inhibition of BMM cell viability when tested at 1 µM. All these compounds tested were not cytotoxic when tested in
RAW 264.7 cells at 1 µM.

Table 1. Initial SAR of bis-amide series.a
CSF1R
IC50
(µM)
c-Kit
IC50
(µM)
BMM
viabilityb
RAW 264.7
viabilityb
R1
R2
R3
#
2
3
Me
Me
CF3
H
0.08
>10
0.44
>10
22
11
7
8
4
5
Me
Cl
Cl
0.23
0.04
2.7
20
-7
4
8
CF3
0.19
6
7
Me
Me
CF3
CF3
0.02
0.01
0.21
0.31
61
64
6
9
aExperimental details are in the Supplemental Information. b% inhibition @ 1 µM.
Next, single-point changes to the amide linkers were made to understand the role of the 2 amide bonds in the
compounds described in Table 1. Amide bonds may provide specific hydrogen-bonding interactions with the target
protein and play a role in potency, but at the same time, they may be prone to hydrolysis in vivo due to the presence of
plasma amidases, leading to rapid clearance. Table 2 shows analogs of compound 2, including 8 and 9, where we
successively methylated the amide and noted that these compounds suffered a potency loss for the target enzyme. In
compound 10, we reversed the amide linking to the trifluoromethyl phenyl group. This change led to ~ 2-fold potency
improvement for CSF1R and a 10-fold improvement in c-Kit selectivity compared to compound 2. Next, we
consecutively converted each amide to the reduced aminomethyl analog (compounds 11 and 12). Compound 11
maintained moderate CSF1R potency and selectivity over c-Kit, but compound 12 lost CSF1R inhibition potency, while
improving c-Kit potency.

Table 2. Linker modifications.a
CSF1R
IC50
(µM)
c-Kit
IC50
(µM)
Fold
selectivity
over c-Kit
#
2
Structure
0.08
0.44
5.5
8
9
0.63
1.8
5.9
9
>10
>6
10
11
0.03
0.35
1.68
>10
56
>29
12
11.7
0.27
<1
aExperimental details are in the Supplemental Information.
As thiophenes are a known structural alert14 that are susceptible to reactive metabolite formation15-18, and have been
linked to drug-induced-liver-injury and idiosyncratic toxicities, we embarked on a more comprehensive exploration to
replace the thiophene moiety in compound 10. A variety of heteroaromatic groups were substituted; Table 3 shows chosen
compounds with improvements in selectivity over c-Kit and cell potency. Replacing the 2-thiophene with a pyrazole (14)
and a methyl pyrazole (13) resulted in potency and c-Kit selectivity gains and more importantly, BMM cell viability
improved to ~ 81% inhibition when the compounds were tested at 1 µM. While the tetrazole analog (15) had good
biochemical potency and selectivity over c-Kit, the lack of cell viability could be explained by poor cell permeability due
to the negatively charged nature of this compound at physiological pH. The 6-membered heteroaryl compounds (16, 17,
18, 19) all had potent inhibition of CSF1R. The 2-pyridyl substituted analog (compound 17) had the best c-Kit selectivity
of the series, albeit with a weaker potency in the cell viability (31% inhibition at 1 µM).

Table 3. Analogs of compound 10. SAR of reverse
amide series.a
CSF1R
IC50
(µM)
c-Kit
IC50
(µM)
Fold
selectivity
over c-Kit
BMM
viabilityb, c
#
R1
13
14
15
0.001
<0.0005
0.02
0.74
0.02
7.5
740
>40
375
81
80
2.4
16
17
18
19
0.004
0.012
0.01
0.90
15
225
1250
39
84
31
72
76
0.39
0.009
5.97
663
aExperimental details are in the Supplemental Information. bAll compounds had no effect on RAW 264.7 cell viability. c% inhibition @ 1 µM.
In order to assess the potential for oral delivery of these CSF1R inhibitors, we profiled compound 13 with in-vitro
ADME assays as well as in vivo in a mouse PK study (Table 4). Stability was assessed in NADPH-supplemented mouse
and human liver microsomes (MLM, HLM), where modest turnover (47% of parent remaining) was observed in human
incubates but high turnover was seen in mouse (7 % remaining). As bis-amides might be substrates for plasma amidase
activity, we incubated 13 in mouse and human plasma. Compound 13 was stable in human plasma (95% remaining)
following a 4 hour incubation, however, some degradation was observed in mouse plasma (67% remaining). Kinetic
solubility measured in isotonic phosphate buffer was low (5 µM) and suggested that formulating with co-solvents may be
necessary for preclinical oral PK studies. Although stability was low in MLM and plasma, we dosed 13 to mice to
investigate the PK and demonstrate concordance with the in vitro data. High systemic clearance following intravenous
administration was observed, which approximated three times that of liver blood flow in mice (~4.4 L/h/kg)19, suggesting
extrahepatic metabolism, most likely due to plasma instability by hydrolysis. The high systemic clearance may have
contributed to high first-pass metabolism when 13 was dosed orally and resulted in low oral bioavailability (12%).
Permeability and efflux were not tested for 13, but would be expected to contribute to the low oral bioavailability as its
physicochemical properties violate Lipinski rules (MW >500 amu and cLogP >5). Additionally, the positive charge at
physiological pH due to the N-methyl- piperazine moiety would increase the potential for P-gp efflux.

Table 4. In vitro and in vivo ADME for compound 13a
In vitro ADME
H/M
Plasma
(%R)
Kinetic solubility
(µM)
HLM/MLM
(%R)
4.7
46/7
95/67
In vivo mouse PK
Route
IV
PO
Dose level
2 mg/kg
10 mg/kg
C
(µM)
0.85
NA
0.16
0.25
0.16
max
Tmax (h)
AUC (µM•h)
0.27
last
Terminal
(h)
0.23
12.6
4.23
NA
1.32
NA
t1/2
CL (L/h/Kg)
(L/Kg)
V
NA
ss
Bioavailability (%)
11.7
H: human, M: mouse, HLM: human liver microsomes, MLM: mouse liver microsomes, %R: percent of parent remaining in incubation relative to
time zero control. Cmax: max plasma concentrations, Tmax: time of Cmax, AUClast: area-under-curve from time 0 to last quantifiable concentration,
CL; systemic clearance, Vss; volume of distribution at steady-state. Bioavailability: dose normalized ratio of the oral AUC to the IV AUC.
aExperimental details are in the Supplemental Information.
Considering the poor stability of 13 in mouse liver microsomes, we explored the biotransformation products in
NADPH-supplemented MLMs to identify metabolic soft-spots. Metabolite identification (Met-ID) analysis revealed that
demethylation at the methyl piperazine of 13 was the major metabolite, with minor hydroxylation sites also noted. The
major biotransformation product was putatively suggested to be a simple N-demethylation of the piperazine moiety. This
was confirmed by preparing an authentic sample (20), which showed the same retention time and product ion spectra as
the N-demethyl metabolite. Interestingly, when tested in our assay panel, the metabolite (20) maintained inhibitory activity
against CSF1R and improved c-Kit selectivity (Figure 2). As expected, based on Met-ID results, the liver microsome
stability improved as well. However, the loss in cell-based activity prompted us to question the necessity of the piperazine
group.
Figure 2. Results of Met-ID of compound 13
There are a number of publicly available crystal structures of CSF1R in both DFG-in and DFG-out conformations.
Given that it was unclear which conformation of CSF1R this series of compounds would bind to, we initially docked 13
into both 3LCD (DFG-in)20 and 4R7I (DFG-out bound with Imatinib)7 using Glide.21 The 3D coordinates of all ligands
were generated with LigPrep22 while the structure was prepared with the Protein Preparation Wizard in Maestro23 using the
default options. It was very clear that 13 appeared to fit much better in the DFG-out conformation of CSF1R. Poses of 13
docked to 3LCD appeared to be strained or could not make interactions with the hinge. However, when docked to 4R7I,
13 bound in a manner very similar to Imatinib. The piperazine in 13 makes key interactions with the protein through
Asp796 and His776 while the amide Glu633 can make hydrogen-bonds with Asp796 and Glu633. The central phenyl ring
can make a pi/cation interaction with Lys616. Interestingly, in this pose, the compound does not appear to make a formal
hydrogen-bond to the hinge region as is typical in many other kinase inhibitors.

A
B
Figure 3. A) Binding pose and B) Ligand interaction diagram of 13 docked into CSF1R (using 4R7I), highlighting key interactions that the
piperazine makes with the protein. Residues and interactions are colored according to the following scheme: Hydrophobic (green), polar
(cyan), negatively charged (red), positive charges (purple), glycine (beige), purple arrow (hydrogen-bond), red line (pi-cation). Grey spheres
indicate an atom as being solvent exposed.
When we removed the piperazine group, compound 21 suffered a 265x loss of potency, which was understandable
based on our docking studies (Figure 3). We reasoned that we needed to include new functional groups that would interact
with the binding pocket of the protein, to regain the lost potency. For this we have docked compound 21, followed by the
design of compounds 22 and 23 which showed us that the introduction of nitrogen atom meta to methylpyrrazole ring
picked up the hinge Cys666 interaction (Figure 4). The core of the molecule is still able to maintain the key hydrogen-
bonds with Glu633 and Asp796, as well as the pi-cation interaction with Lys616. Similar docking results were obtained
for compound 24 (Figure 5). However, replacing the phenyl methyl piperazine moiety with an amino-pyrimidine allows
the inhibitor to make two new hydrogen bonds to the backbone of Cys666, one between the pyrimidine nitrogen and the
backbone N-H, a second between the amino NH2 and backbone carbonyl of Cys666.
This hypothesis was indeed supported by testing these compounds in the biochemical assay which showed potent
CSF1R activity in comparison with compound 21 (Table 5).

a
Table 5. SAR leading to removal of the N-methyl piperazine
.
CSF1R
IC50
(µM)
c-Kit
IC50
(µM)
Fold
selectivity
over c-Kit
BMM
HLM/MLM
(% rem)
#
Structure
viabilityb, c
13
21
22
0.001
0.37
0.74
0.13
0.06
740
0.35
120
81
ND
96
46/7
68/74
37/63
0.0005
23
0.0009
0.18
200
87
60/57
24
0.032
0.82
25
86
91/65
aExperimental details are in the Supplemental Information bAll compounds had no effect on RAW 264.7 cell viability. c% inhibition @ 1 µM.
A
B
Figure 4. A) Ligand interaction diagram of 22 and B) 23 docked into CSF1R. In both docked poses, the conserved hydrogen-bond interactions
between the inhibitor and Glu633 and Asp796, as well as the pi-cation interaction with Lys616, are well maintained. In addition, there is a
possible hydrogen-bond interaction between the side chain of Thr663 and the amide NH. Replacing the phenyl ring, adjacent to the methyl
piperazine moiety, by a pyridine allows the inhibitor to make a hydrogen-bond with the backbone carbonyl of Cys666, located in the hinge
region, improving potency. Residues and interactions are colored according to the following scheme: Hydrophobic (green), polar (cyan),
negatively charged (red), positive charges (purple), glycine (beige), purple arrow (hydrogen-bond), red line (pi-cation). Grey spheres indicate an
atom as being solvent exposed.

A
B
Figure 5. A) Binding pose and B) Ligand interaction diagram of 24 docked into CSF1R (using 4R7I). Residues and interactions are colored
according to the following scheme: Hydrophobic (green), polar (cyan), negatively charged (red), positively charged (purple), glycine (beige),
purple arrow (hydrogen-bond), red line (pi-cation). Grey spheres indicate an atom as being solvent exposed.
BLZ945 is a known ATP competitive CSF1R inhibitor.24 Most known kinase inhibitors are Type I inhibitors, ATP-
competitive compounds such as staurosporine, erlotinib (Tarceva®) and dasatinib (Sprycel®), that bind to the ATP
binding site and hydrogen bond with the hinge region of the kinase. Type II inhibitors are compounds which bind partially
in the ATP binding site and extend past the gatekeeper and into an adjacent allosteric site that is present only in the
inactive kinase conformation. Imatinib is a well-known Type II inhibitor. As our docking studies indicated that 13 bound
in a manner similar to imatinib, we explored the mechanism of inhibition of a few selected molecules via a biochemical
assay. Compounds 13 and 22 which were seen to pick up key interactions with the CSF1R protein, based on docking
studies, were chosen and compared with BLZ945 binding in the presence of variable concentrations of ATP. While
BLZ945 showed a ~30-fold increase in IC50 at 500 µM ATP (40X Km ATP for CSF1R enzyme), 13 showed minimal
change in IC50 at similar ATP concentration and 22 showed only marginal (~6-fold) increase in IC50 at 500 µM ATP
(Supplemental Information). This indicated that unlike BLZ945 which behaves in a purely competitive fashion with
respect to ATP, both 13 and 22 behave differently and were more like Type II inhibitors that are mildly ATP competitive
or are non-competitive with ATP. This biochemical study validates the choice of the DFG-out configuration for the
modeling studies and supports the predicted binding mode based on docking studies.
Compounds 22, 23 and 24 were tested for their in vitro stability in NADPH-supplemented human and mouse liver
microsomes. They were found to be more stable in MLM with moderate turnover (range 57% to 65% remaining), relative
to compound 13. Additionally, we studied these new lead compounds in a Caco2 permeability assay and found that there
were a range of permeabilities and efflux values for these analogs (Table 6). Both 23 (Papp = 0.5 x 10-6 cm/s) and 24
(Papp = 0.1 x 10-6 cm/s) demonstrated low absorptive permeability (apical to basolateral direction), while 22 demonstrated
high absorptive permeability (Papp = 12.5 x 10-6 cm/s). All three compounds, however, were effluxed by Caco2
membrane transporters with secretory (basolateral to apical) to absorptive (apical to basolateral) permeability ratios values
of 2.6 (22), 40 (23) and 82 (24) (Table 6). Based on its higher absorptive permeability and lower efflux, the oral
absorption of 22 would not be expected to be impacted by slow absorption or efflux back into the lumen of the gut.

Table 6. Caco2 permeability of Compounds 22, 23 and 24a
A2B
B2A
Efflux Ratio
(A2B/B2A)
#
Mean Papp
(10-6 cms-1)
Mean Papp
(10-6 cms-1)
SD
SD
22
23
24
12.5
0.5
0.81
0.04
0.00
32.5
36.3
5.3
0.47
3.34
0.38
2.6
81.5
40.4
0.1
A2B: apical to basolateral (absorption) direction, B2A: basolateral to apical (secretory) direction, Papp, apparent permeability; Efflux ratio;
B2A/A2B. aExperimental details are in the Supplemental Information.
Compound 22 was further evaluated in a mouse PK study, where systemic clearance was substantially reduced (50% of
mouse hepatic blood flow) compared to 13 and the oral bioavailability was increased (64%) as shown in Table 7. These
data are consistent with improved metabolic stability, permeability and solubility (38 µM) compared to compound 13.
Table 7. Mouse PK of Compound 22a
In vivo mouse PK, compound 22
Route
IV
PO
Dose level
2 mg/kg
10 mg/kg
C
(µM)
3.55
NA
4.60
0. 50
6.94
max
Tmax (h)
AUC (µM.h)
2.24
last
Terminal
(h)
0.87
1.86
1.10
NA
1.80
NA
t1/2
CL (L/h/Kg)
(L/Kg)
V
NA
ss
Bioavailability (%)
64.0
Cmax: max plasma concentrations, Tmax: time of Cmax, AUClast: area-under-curve from time 0 to last quantifiable concentration, CL: systemic
clearance, Vss: volume of distribution at steady-state. Bioavailability: dose normalized ratio of the oral AUC to the IV AUC. aExperimental
details are in the Supplemental Information.
The synthesis of the compounds 2-5 described in Table 1 followed the sequence of steps described in Scheme 1.
Synthesis started with the coupling of acid I with amines II and VIII to provide the ester compounds III and IX. These
esters were then hydrolyzed using NaOH to provide the corresponding acids IV, X, which were then subjected to amide
coupling with various 3-substituted 4-((4-methylpiperazin-1-yl) methyl) anilines V, VI and VII gave the bisamides 2 – 5.

Scheme 1. Synthesis of compounds 2 – 5
The synthesis of the compounds 10, 13 – 14, 16, 18 - 19 described in Table 3 followed the sequence of steps described in
Scheme 2. The commercially available amine XXI was coupled with acid XXII to provide the bromo bisamide XXIII
which were then coupled with various heterocyclic boronates resulting in compounds 10, 13 – 14, 16, 18 - 19 (Scheme 2).
Scheme 2. Synthesis of compounds 10, 13-14, 16, 18 – 19
Compound 22 was synthesized as described in Scheme 3. The commercially available amine XXXXVI was coupled with
3-(trifluoromethyl)benzoic acid to provide the bromo bisamide XXXXVIII which underwent a Suzuki coupling with
methypyrazole-4-boronic acid resulting in compound 22.
Scheme 3. Synthesis of compound 22
Compounds 6 - 9, 11 - 12, 15, 17, 20 - 21, 23 - 24 were synthesized by alternative routes (see supporting information).
In this manuscript, we have described the synthesis and optimization of CSF1R inhibitors. Initial leads contained an N-
methyl piperazine group that was required for activity, but was metabolically labile, contributing to a poor mouse
pharmacokinetic profile. Biochemical and docking analysis allowed us to remove the N-methyl piperazine group, while
introducing other hydrogen-bonding interactions with the binding pocket of CSF1R. Removal of this group also afforded
better physicochemical properties by reducing the molecular weight, charge and number of hydrogen bonding

acceptors/donors. Through this approach, we could design and synthesize new compounds with more potent inhibition of
CSF1R and improved cellular efficacy. By virtue of better physicochemical properties, our lead compound 22, also
showed good intestinal permeability in a Caco2 assay and favorable pharmacokinetics when dosed orally to mice. This
compound appears suitable for in vivo pharmacology testing in the appropriate preclinical tumor model to demonstrate
proof of concept. Future work could be focused on optimizing this lead for cancer applications.
Funding: This research did not receive any specific grant from funding agencies in the public, commercial or not-for-profit
sectors. It was funded by Medivation, now Pfizer.
Acknowledgements: The authors thank Stacy Kanno, Teresa Kameda, Ramona Almirez, Katherine Stack, Walter Yu and
Sunny Chen for in-life and bioanalysis work and the Analytical and Purification group at Integral Bioscience for
purification of compounds and for acquiring relevant analytical data.

1. Hamilton, J.A. Nat Rev Immunol. 2008, 8, 533-544.
2. Ruffell, B.; Coussens, L.M. Cancer Cell. 2015, 27, 462-472.
3. Noy, R.; Pollard, J.W. Immunity. 2014, 41, 49-61.
4. Ries, C.H.; Hoves, S.; Cannarile, M.A.; Ruttinger, D. Curr Opin Pharmacol. 2015, 23, 45-51.
5. Pyonteck, S.M.; Akkari, L.; Schuhmacher, A.J.; Bowman, R.L.; Sevenich, L.; Quail, D.F.; Olson, O.C.;
Quick, M.L.; Huse, J.T.; Teijeiro, V.; Setty, M.; Leslie, C.S.; Oei, Y.; Pedraza, A.; Zhang, J.; Brennan, C.W.;
Sutton, J.C.; Holland, E.C.; Daniel, D.; Joyce, J.A. Nat Med. 2013, 19, 1264-1272.
6. Zhu, Y.; Knolhoff, B.L.; Meyer, M.A.; Nywening, T.M.; West, B.L.; Luo, J.; Wang-Gillam, A.;
Goedegebuure, S.P.; Linehan, D.C.; DeNardo, D.G. Cancer Res. 2014, 74, 5057-5069.
7. Tap, W.D.; Wainberg, Z.A.; Anthony, S.P.; Ibrahim, P.N.; Zhang, C.; Healey, J.H.; Chmielowski, B.;
Staddon, A.P.; Cohn, A.L.; Shapiro, G.I.; Keedy, V.L.; Singh, A.S.; Puzanov, I.; Kwak, E.L.; Wagner, A.J.; Von
Hoff, D.D.; Weiss, G.J.; Ramanathan, R.K.; Zhang, J.; Habets, G.; Zhang, Y.; Burton, E.A.; Visor, G.;
Sanftner, L.; Severson, P.; Nguyen, H.; Kim, M.J.; Marimuthu, A.; Tsang, G.; Shellooe, R.; Gee, C.; West,
B.L.; Hirth, P.; Nolop, K.; van de Rijn, M.; Hsu, H.H.; Peterfy, C.; Lin, P.S.; Tong-Starksen, S.; Bollag, G. N
Engl J Med. 2015, 373, 428-437.
8. https://www.clinicaltrials.gov/ct2/results?term=PLX7486&Search=Search
9. Company website: http://www.arraybiopharma.com/product-pipeline/other-compounds/arry-382/
10. von Tresckow, B.; Morschhauser, F.; Ribrag, V.; Topp, M.S.; Chien, C.; Seetharam, S.; Aquino, R.;
Kotoulek, S.; de Boer, C.J.; Engert, A. Clin Cancer Res. 2015, 21, 1843-1850.
11. Lennartsson, J.; Ronnstrand, L. Physiol Rev. 2012, 92, 1619-1649.
12. Hume, D.A.; MacDonald, K.P. Blood. 2012, 119, 1810-1820.
13. Choi, Y.; Syeda, F.; Walker, J.R.; Finerty, P.J., Jr.; Cuerrier, D.; Wojciechowski, A.; Liu, Q.; Dhe-
Paganon, S.; Gray, N.S. Bioorg Med Chem Lett. 2009, 19, 4467-4470.
14. Kalgutkar, A.S.; Gardner, I.; Obach, R.S.; Shaffer, C.L.; Callegari, E.; Henne, K.R.; Mutlib, A.E.; Dalvie,
D.K.; Lee, J.S.; Nakai, Y.; O'Donnell, J.P.; Boer, J.; Harriman, S.P. Curr Drug Metab. 2005, 6, 161-225.
15. Wu, G.; Vashishtha, S.C.; Erve, J.C. Chem Res Toxicol. 2010, 23, 1393-1404.
16. Dansette, P.M.; Libraire, J.; Bertho, G.; Mansuy, D. Chem Res Toxicol. 2009, 22, 369-373.
17. Ono, K.; Kurohara, K.; Yoshihara, M.; Shimamoto, Y.; Yamaguchi, M. Am J Hematol. 1991, 37, 239-
242.
18. Koenigs, L.L.; Peter, R.M.; Hunter, A.P.; Haining, R.L.; Rettie, A.E.; Friedberg, T.; Pritchard, M.P.; Shou,
M.; Rushmore, T.H.; Trager, W.F. Biochemistry. 1999, 38, 2312-2319.
19. Davies, B.; Morris, T. Pharm Res. 1993, 10, 1093-1095.
20. Meyers, M.J.; Pelc, M.; Kamtekar, S.; Day, J.; Poda, G.I.; Hall, M.K.; Michener, M.L.; Reitz, B.A.;
Mathis, K.J.; Pierce, B.S.; Parikh, M.D.; Mischke, D.A.; Long, S.A.; Parlow, J.J.; Anderson, D.R.;
Thorarensen, A. Bioorg Med Chem Lett. 2010, 20, 1543-1547.
21. Glide v7.2; Schrödinger, Inc.: Portland, OR.
22. LigPrep v3.9; Schrödinger, Inc.: Portland, OR.
23. Glide v10.7; Schrödinger, Inc.: Portland, OR.
24. Daniel, D.; Joyce, J.; Sutton, J. WO2012/151541 A1.

Graphical Abstract
To create your abstract, type over the instructions in the template box below.
Fonts or abstract dimensions should not be changed or altered.
Leave this area blank for abstract info.
Design, synthesis and optimization of bis-amide derivatives as
CSF1R inhibitors
Sreekanth A. Ramachandran, Pradeep S. Jadhavar, Sandeep K. Miglani, Manvendra P.
Singh, Deepak P. Kalane, Anil K. Agarwal, Balaji D. Sathe, Kakoli Mukherjee, Ashu
Gupta, Srijan Haldar, Mohd. Raza, Siddhartha Singh, Son M. Pham, Sarvajit Chakravarty,
Kevin Quinn, Sebastian Belmar, Ivan E. Alfaro, Christopher Higgs, Sebastian Bernales,
Francisco J. Herrera and Roopa Rai
New hinge interaction
O
O
O
O
CF3
CF3
N
N
SAR
N
N
N
H
H
H
H
N
N
13
N
N
22
N
N
Good CSF1R potency
Low metabolic stability
High potential for P-gp efflux
Poor oral bioavailability
Good CSF1R potency
Improved metabolic stability
Good permeability/low efflux
Improved oral bioavailability